B2C2 Ltd v Quoine Pte Ltd [2019] SGHC(I) 03

Introduction

This is an action for breach of contract and breach of trust. The Defendant (“Quoine”) is a Singapore-registered company which operates a currency exchange platform (the “Platform”) enabling third parties to trade virtual currencies for other virtual currencies or for fiat currencies such as Singapore or US dollars. The two virtual currencies (sometimes referred to as cryptocurrencies) involved in this action are Bitcoin (“BTC”) and Ethereum (“ETH”).

The Plaintiff (“B2C2”) is company registered in England and Wales trading inter alia as an electronic market maker. As an electronic market maker, B2C2 provides liquidity on exchange platforms by actively buying or selling at the prices it quotes for virtual currency pairs, thereby generating trading revenue.

In recent years, there has been a significant growth in virtual currencies of which Bitcoin is perhaps the best known. They are not linked to any particular country nor regulated by any central monetary authority. They are traded for other virtual currencies or traditional currencies on computer networks such as the Platform.

B2C2’s claim arises out of an incident that occurred on the Platform on 19 April 2017 when, as a result of a series of events which I shall consider in more detail below, seven trades for the sale by the Plaintiff of ETH for BTC at an exchange rate of either 9.99999 or 10 BTC for 1 ETH were effected by the Platform. This was at a rate approximately 250 times the rate of about 0.04 BTC for 1 ETH which had been the previous going rate. The proceeds of the sale (3092.517116 BTC) were automatically credited to B2C2’s account by the platform and the corresponding amount of 309.2518 ETH was automatically debited from that account.

When Mr Mario Antonio Gomez Lozada (“Mr Lozada”), the Chief Technology Officer (“CTO”) of Quoine, became aware of the trades the following morning, he considered the exchange rate to be such a highly abnormal deviation from the previously going rate that the trades should be reversed. Accordingly, the seven trades were cancelled by Quoine and the debit and credit transactions were reversed.

B2C2 contends that Quoine had no contractual right unilaterally to cancel the trades once the orders had been effected and that its action in so doing was in breach of the version of the Terms and Conditions which regulated the trading relationship between Quoine and B2C2 at the material time. Additionally, B2C2 contends that Quoine holds the proceeds of B2C2’s account on trust for B2C2 and that the unilateral withdrawal of the BTC which had been credited to the account as a result of the trades was in breach of trust.

The writ was issued in the Singapore High Court on 18 May 2017, the Statement of Claim was served on that day and the Defence on 16 June 2017. With the agreement of the parties, the action was transferred to the Singapore International Commercial Court on 24 August 2017 and came on for trial before me on 21 November 2018.

There was an application for Summary Judgment under Order 14 of the Rules of Court (Cap 322, R 5, 2014 Rev Ed). I dismissed this application in a judgment delivered on 27 December 2017: see B2C2 v Quoine Pte Ltd [2018] 4 SLR 1. Since then there has been further disclosure coupled with affidavit evidence and expert reports. As such, the actual facts of the case as detailed below are somewhat different from the facts as they appeared to be as at the time of that summary judgment application.

Currency trading

There are a large number of trading platforms around the world and some of these provide for trading in virtual currencies. At a high level of generality, they operate in the same way. There are two types of traders, market makers and investor-traders. An investor-trader, as its name suggests, is a person who purchases a given commodity with a view to holding it for a period of time in the hope that its value increases so that it can be sold at a profit.

Market makers are more regular traders who provide liquidity to a given platform by quoting orders (both buy and sell orders) and seek to generate income by capturing what is known as the bid-ask spread. B2C2 is a market maker. Market makers will trade over a number of platforms (in B2C2’s case at the relevant time it was trading on 15) and do so by using sophisticated algorithmic trading software. At any one time a market maker may be seeking to sell a given currency on one exchange and buy it on another or, indeed, on the same exchange. The software monitors the trader’s position over all the platforms to regulate the exposure to any given currency and there will therefore be times when no buy quotes for a given currency are being placed on any of the platforms because the software has determined that the trader is over-exposed in that currency and is therefore, for the time being, only going to be a seller. The reverse can obviously also be the case.

Market makers can on occasions also be investor-traders but this action is concerned with B2C2’s activity as a market maker. Of more relevance is the fact that operators of currency exchange platforms can also trade as market makers on their own platform as Quoine did in the present case.

Whilst there are a number of ways in which an order can be placed on a platform, there are two types of orders that are material in this dispute:

(a) A market order: This is an order which is to be executed immediately at the best available current market place. The buyer or seller (as the case may be) indicates what they wish to trade and the platform automatically identifies the best available trade in the opposite direction. At the time of order placement, the trader will not know precisely what the exchange rate will be, only that it will be the best available price offered on the Platform at the time by a trader seeking to trade in the opposite direction.

(b) A limit order: This gives the trader greater control over the prices of a given trade but does not enable him to know when it will be executed. A limit order sets either the maximum price at which he is willing to buy or the minimum price at which he is willing to sell. If and when that price is reached, his order will be fulfilled. Market makers trade exclusively in limit orders.

As Mr Maxime Boonen (“Mr Boonen”), a director of B2C2, explained in his affidavit of evidence-in-chief (“AEIC”), B2C2 acted as a market maker on the Platform in the following manner:

The Platform

Quoine was cofounded by Mr Lozada together with Mr Mike Kayamori in 2014. It operates platforms in Japan and in Singapore. The Platform functions in the following way:

At the relevant time, Quoine offered three types of trading: spot trading (where trades are settled instantly), margin trading (which allows trading with borrowed funds) and futures trading (which involves both the buyer and seller agreeing to sell at a future date at a given price). Quoine no longer offers futures trading.

Margin trading is central to this dispute and therefore needs to be explained more fully. A trader may trade on the Platform using borrowed funds, borrowed either from Quoine or from other users who offer their funds to be used for peer-to-peer loans. The assets in the margin trader’s account serve as collateral for the loans. In the event that the collateral in the account falls below a pre-determined percentage of the loan, the Platform’s system takes action by making a margin call and closing out all or part of the trader’s positions to prevent further loss and to seek to avoid a default.

The system constantly monitors the margin trader’s profit and loss and is designed to force-close positions if it detects that the trader cannot pay back his loans from its collateral. These force-closures are automatically executed by the system as market orders to buy or sell the relevant commodity at the best available price. The margin trader does not have to be contacted in advance and may not know that the force-closure is going to occur. It certainly will not know the precise price at which the market order will be executed.

Quoine also acts on the Platform as a market maker “to provide liquidity and depth to the order books, minimise volatility in the marketplace and ensure that there is a continuous two-sided market on the Platform”. It did so using what is known as a Quoter Program in the following manner:

The relationship between Quoine and B2C2

B2C2 was set up by Mr Boonen and a Mr Molendini. Mr Boonen was responsible for creating the vast majority of the algorithmic trading software which I accept is the life blood of B2C2’s business and therefore highly confidential. It has been necessary during the course of this action for steps to be taken to seek to ensure the maintenance of that confidentiality whilst also ensuring a fair and open trial. The way in which this was done and the reasons therefor are set out in an earlier judgment in this action: B2C2 Ltd v Quoine Pte Ltd [2018] 4 SLR 67.

B2C2 opened an account on the Platform on 28 June 2015. This was done electronically online by Mr Boonen and Mr Boonen was asked to confirm that B2C2 accepted Quoine’s Terms and Conditions which Mr Boonen did.

Mr Boonen gave evidence that initially he had some concerns about trading on the Platform but following some reassurance from Mr Lozada in June 2016 and the provision of a line of credit from Quoine to B2C2, B2C2 commenced full trading on the Platform in January 2017. A good deal of written evidence and some cross-examination was directed to the reasons underlying those concerns and how and why they were alleviated but I am satisfied that this has no bearing on the matters in dispute.

The Agreement and the Risk Disclosure Notice

The Terms and Conditions (“the Agreement”) which were on the Defendant’s website in June 2015 when B2C2 opened its account were stated to have been “Last updated April 2014”.

A great deal turns on the terms of the Agreement and its relationship with other documents that were on the website at the date of the incident. It is therefore included in full as Annex 1 to this judgment. The most important provisions are as follows:

Additionally, Quoine asserts that a Risk Disclosure Statement was uploaded onto the website on 22 March 2017. Whilst this was not admitted by B2C2, it was not seriously challenged and I accept the evidence demonstrating that it was. This document figures in one of Quoine’s proposed defences and, again, it is necessary to see the passages relied upon in context so this document forms Annex 2 to this Judgment. I shall revert to it later.

The incident on 19 April 2017

Put in very simple terms, late in the evening of 19 April 2017 (all times are GMT/UTC), two margin traders, Pulsar Trading Capital and Mr Yu Tomita (“Pulsar” and “Mr Tomita” respectively, and “the Counterparties” collectively) were trading in the ETH/BTC market using ETH borrowed from Quoine. The Platform was alerted to the fact that both Counterparties’ collateral had fallen below the maintenance margins. The Platform thus automatically placed market orders to sell the Counterparties’ assets at the best available prices to repay the ETH loans. Some of those orders were met by limit orders which were subsequently placed on the Platform by B2C2’s trading software at the exchange rate of about 10 BTC to 1 ETH. There was no human intervention in this. It was all done by computers acting in accordance with their respective programs. Further, the placing of the orders by B2C2 was not the cause of the Counterparties’ collateral falling below the maintenance margins. This had occurred before the orders were placed on the order book.

As indicated above, Quoine reversed these orders on 20 April 2017, the next day. The background to these trades and the way in which they came to be carried out is complex and has been the subject of both fact evidence and expert reports.

The Witnesses

Evidence was given by Mr Boonen for B2C2 and by Mr Lozada and Mr Tseung, a co-founder of Pulsar, for Quoine. Expert evidence was given by Mr Alexis Atkinson (“Mr Atkinson”) on behalf of B2C2 and Mr Vikram Kapoor (“Mr Kapoor”) on behalf of Quoine.

Mr Boonen received a Bachelor of Engineering degree from the University of Louvain in Belgium and then a Master’s degree in Financial Economics from the University of Oxford in England. He was first employed by Google where he learned about electronic trading and then worked in the fixed income department of Goldman Sachs in London. He left Goldman Sachs in May 2015 to found B2C2.

Mr Boonen was well-informed as to the trading environment in which B2C2 operated and was plainly a highly experienced computer programmer who had successfully designed B2C2’s trading software. In the witness box, he was both reasoned and precise and explained technical matters with clarity. It became apparent in cross-examination that an aspect of his AEIC was overstated but in the event nothing turned on that. I found him to be an honest and straightforward witness who gave reliable evidence, even if, on occasions, his desire to be precise led to overlong answers. His demeanour as a witness was not the subject of any general criticism in Quoine’s closing submissions but it was suggested that his evidence was unbelievable in certain respects. I shall assess the weight that can be placed on aspects of Mr Boonen’s evidence later.

Mr Lozada has both an undergraduate and Master’s Degree in Computer Sciences from the University of Kansas, USA. Initially he was employed in Tokyo by Fusions Systems Japan which provided financial software for banks. He then spent 10 years with Merrill Lynch ending up as CTO for the fixed income business in Japan. He then moved to Credit Suisse in 2009 or 2010 where he was both CTO for the fixed income business and Chief Information Officer for the Japan business. He left Credit Suisse in 2013 to cofound Quoine. Mr Lozada was well informed as to Quoine’s trading environment and was an experienced computer programmer having designed the Platform’s system.

In B2C2’s closing submissions it was submitted that Mr Lozada was very unconvincing and far from forthcoming so that his evidence should be viewed with extreme circumspection. A number of aspects of his evidence were brought to my attention to support this assertion. These related in part to evidence which displayed an element of uncertainty in his overall evidence and in part to evidence given at trial which was inconsistent with earlier evidence given on the application for summary judgment. I have considered carefully each of these assertions as, at the time he gave his evidence, I perceived him to be an honest and straightforward witness. I consider that, at best, it could be said that Mr Lozada did not investigate the circumstances surrounding the incident of 19 April in sufficient detail before giving the evidence he did on the summary judgment application. But I am satisfied that this was not done with any intention to mislead the Court. The facts are complex and it is not surprising that they became clearer in the course of time. Equally where there were passages in his evidence which constituted assertions as to the intentions of B2C2, these were, I consider, honestly held beliefs. The areas of uncertainty in his evidence were due to imperfect recollection rather than any intention to deceive. All in all, I see no reason to alter my original perception that he was an honest and straightforward witness who was particularly frank in identifying the areas where the Quoine system had proved to be inadequate and the steps that he had taken to rectify this.

The final witness of fact was Mr Tseung of Pulsar. Once certain passages of his evidence were ruled to be inadmissible, the Plaintiff no longer sought to cross-examine him. Notwithstanding this, in its written closing, B2C2 asserted that Mr Tseung’s evidence was self-serving. But this was a criticism directed to the evidence which had been struck out and to Mr Tseung’s statements relating to the validity of the seven trades, which is a matter for me and not him.

Mr Atkinson has a BSc (Hons) in Mathematics and Statistics from the University of Sheffield, UK, and since graduating has been involved in designing and developing financial trading systems. In particular, whilst employed by Radobank International from 2008 to 2015, he was responsible for the development, maintenance and support of that bank’s EMarkets trading architecture and features, which included algorithmic market making and trading engines. He left Radobank in May 2015 to work for Nex Markets (now part of CME Group). Nex Markets provides electronic execution platforms to the financial trading sector. He is currently Head of Order Driven Markets where he is responsible for foreign exchange platforms. Mr Atkinson had not given expert evidence before.

Mr Kapoor has a BSc in Engineering from the University of Pune, India, and an MBA from the State University of New York at Buffalo, USA. He is a senior managing director at Ankura in New York specialising in quantitative finance valuation and data analytics. His work includes analyses of algorithmic trading and trade routing by financial institutions. He has given expert evidence on a number of occasions both in the US courts and in arbitrations.

Both Mr Atkinson and Mr Kapoor are well qualified to give expert evidence on the defined issues. Both spent a considerable amount of time reviewing B2C2’s software and presented their reports on the issues in a focussed and helpful way. They were, however, very different witnesses. Mr Atkinson was significantly more diffident and less articulate than Mr Kapoor. He was more hesitant in giving his answers compared to the more polished approach of Mr Kapoor.

In its closing submissions, Quoine questioned whether Mr Atkinson acted as independently or objectively as he should have done as an expert witness. Reliance was placed on three matters. First, in his first report at paragraph 78 he gave evidence that he had been present, in his capacity as Head of Order Driven Markets at Nex, at a small number of meetings with B2C2 before agreeing to act as an expert witness in this action. In cross-examination he plainly had little or no recollection of the number or substance of such meetings. I do not consider that he was being evasive in relation to his past dealings with B2C2 and previous contact with an expert witness does not prevent him giving evidence on that person’s behalf. It is a matter that should be taken into account in assessing whether the witness has lost a sense of perspective in preparing his expert report or in giving oral evidence. I do not consider that Mr Atkinson erred in this respect.

Secondly I was invited to place adverse weight on Mr Atkinson’s evidence because he had failed to sign a copy of a confidentiality undertaking supplied to him by B2C2. Plainly he should have done that and B2C2 should have chased him for a signature. But I do not see how this failure can be said to impact upon the evidence that he did give.

Finally, in the course of trial, when Mr Atkinson was in the witness box, he sought to introduce two further documents prepared by him. This was improper but understandable as he had had to deal with Mr Kapoor’s report in a short period of time. He explained that he was seeking to supplement his second report with these documents. I refused to allow one of them to be adduced in evidence but allowed the other, on the basis that, if necessary, Mr Atkinson could be recalled to be cross-examined on matters to which this document related later in the proceedings. In the event no application was made for Mr Atkinson to be recalled. I consider that this incident was but part of the normal vicissitudes of litigation and does not reflect adversely on Mr Atkinson’s evidence.

B2C2 levelled a number of criticisms at Mr Kapoor. Mr Kapoor was responding to Mr Atkinson’s first report which, in very simple terms, supported the assertion that B2C2’s software had been designed for wholly defensive purposes. It was plain that Mr Kapoor had been instructed to focus particularly on aspects of the software that might serve to identify the mindset of the writer as not being limited to defensive purposes. This was, as will become plain, a fundamental part of Quoine’s case. He cannot be criticised for doing this nor for highlighting aspects of the software which appeared to be indicative to him of an intention on B2C2’s part to program its software to take advantage of errors of others and, in the extreme, to manipulate the market. He raised technical matters which it was necessary for B2C2 to respond to and which needed to be resolved.

However, in two respects he strayed beyond permissible limits. First, he sought to draw his own conclusions as to the intentions of the writer of the software which is a matter for me and not for him and formed part of the evidence which I was asked to strike out. Conclusory assertions of this sort are not unusual in expert reports, particularly from experts more familiar in giving evidence in US courts and tribunals. It is no doubt what he was instructed to do and doing so does not add to or subtract from the technical evidence that he gave on the issues he was asked to address.

Secondly, he gave evidence on matters that were not within the scope of the issues on which expert evidence was to be adduced. Again, this was the subject of the application to strike out and I do not doubt that this is what he was asked to do. There is, to my mind, nothing untoward in this and it does not detract from the technical evidence that Mr Kapoor gave on the relevant issues.

All in all, I am satisfied that both experts were qualified to give the technical evidence that they did. They did so in support of their respective client’s cases but in an informed and coherent manner. At all times they were doing their best to assist the court and I am grateful to them for their patience in approaching the issues and in amplifying their reasoning in their oral evidence. It will be for me to assess the weight that I place on each expert’s evidence on these technical matters but I do not accept that either witness’ evidence should be discounted on the basis of the matters considered above.

Inadmissible evidence

Prior to trial, both parties served notices identifying passages in the AEICs and expert reports of the other party which were contended to be inadmissible. Sensibly, counsel agreed that the question of admissibility should not be considered before the witnesses gave evidence but should be left to closing submissions.

However, in advance of the trial, B2C2 did issue a summons to strike out certain matters arising in the affidavits of Mr Lozada and Mr Tseung and in Mr Kapoor’s expert report. This was initially heard in chambers but by agreement of the parties was adjourned into court to be heard at a convenient moment during the trial.

The Court directed that so far as the application sought to strike out parts of the affidavits of the witnesses of fact, this should be determined at the outset of the trial with the objections to Mr Kapoor’s report being determined at a later time.

I therefore heard argument on the fact witnesses. I indicated at the conclusion of that argument that three paragraphs would be struck out and that I would give my reasons for so doing as part of the Judgment after trial. These are my reasons.

The paragraphs in question were paragraph 28 of Mr Lozada’s first AEIC and paragraphs 9 and 10 of Mr Tseung’s AEIC. These state:

It is convenient to deal with paragraph 28 of Mr Lozada’s first AEIC and paragraph 9 of Mr Tseung’s AEIC together.

It will be seen that these paragraphs relate to an alleged common practice for trading exchanges to cancel clearly erroneous transactions. Mr Danny Ong for B2C2 contended that an investigation into how other exchanges approached the question of dealing with erroneous transactions was irrelevant to the issues arising on the pleadings in this action. He drew my attention to the fact that there is no express term in the Agreement providing for such cancellation and that the plea of an implied term to this effect contained in paragraph 19(d) of the Defence (Amendment No 4) (“the Defence”) was solely on the basis of giving business efficacy to the agreement and to give effect to the intentions of the parties. This did not, he submitted, provide for the implication of such a term on the basis that there was a common practice in the trade for this to be done. Further, he claimed that in any event the evidence constituted impermissible opinion evidence of what the witnesses perceived to be common practice. If the evidence was allowed to stand, it would be necessary for B2C2 to adduce further evidence to rebut the contention.

Mr Paul Ong for Quoine submitted, first, that the correct course was to deal with the question of admissibility by way of closing submissions as was being done in relation to the other disputed evidence. Secondly, whilst he accepted that there was no plea of common practice, he said that the evidence was relevant to demonstrate that this was a term to which the parties would have agreed had it been posed by the notional hypothetical onlooker at the time the Agreement was made. He said that if this was the policy of other trading platforms this would be an indication that it was more likely that the parties would have agreed. Thirdly, he contended that this was not impermissible opinion evidence.

Whilst I accept the third of these submissions, the agreements that are said to evidence such a common practice contain an express term to that effect; there is therefore no need for any implied term. There is no such express term in the Agreement. The pleading contends that one should be implied to give business efficacy to the Agreement. The mere fact that the parties might have agreed such a term had it been proposed is, of itself, insufficient for the term to be implied. If the needs of business efficacy do not require the implication of a term, it will not be implied even if, had it been proposed, it might have been accepted. I am therefore satisfied that the evidence contained in paragraph 28 of Mr Lozada’s first AEIC and paragraph 9 of Mr Tseung’s AEIC is irrelevant to any pleaded issue in this case.

Whilst there is often merit in adopting the course of leaving questions of admissibility to be discussed during closing submissions, equally there is merit in a clear case of striking out inadmissible evidence at an early stage if this is going to avoid the possibility of further evidence and/or cross-examination on the irrelevant material. I consider this to be a clear case.

Paragraph 10 of Mr Tseung’s AEIC falls into a different category. It is plainly speculation on the part of Mr Tseung as to why B2C2 programmed its trading systems in the way it did. Speculation by a witness of fact as to why another entity might have done something is of no assistance to the Court and time should not be spent at trial cross-examining Mr Tseung on the basis for his speculation.

Accordingly, I ordered that the three paragraphs should be struck out.

At a later stage in the trial before the expert witnesses were to be cross-examined, I heard submissions on the admissibility of the passages in Mr Kapoor’s expert report to which objection was taken. There was an earlier hearing in this case on 20 February 2018 at which the question of expert evidence was discussed. I made it clear that in a case of this nature, expert evidence would only be allowed in relation to specific issues which were to be determined by the parties and approved by the Court. The parties proposed a number of possible avenues of expert evidence. In particular, the Defendant proposed that there should be expert evidence on market manipulation. Having heard submissions, I directed that expert reports should be directed to (and limited to) certain specific issues concerned with the working of the B2C2’s trading algorithm and that if the parties wished at a later date to adduce expert evidence on different issues, they should make an application for permission in advance of doing so.

The specific issues were drafted by the parties and settled by the Court. They were as follows:

It was directed that the expert reports should be prepared sequentially which resulted in the first report of Mr Atkinson being sworn on 28 August 2018 and Mr Kapoor’s in reply being sworn on 5 November 2018. Mr Atkinson’s report addresses the issues but Mr Kapoor’s report went somewhat further and sought to give expert evidence on two other matters. First, he concluded on the basis of his review of the B2C2’s computer programs that the “deep quotes” set by the algorithm (ie, the price of 10 BTC to 1 ETH) were not merely a defensive measure (of which more below) but constituted what are known as “stub quotes”, a practice which Mr Kapoor opined constituted impermissible market manipulation which was contrary to some regulatory requirements. Secondly, he reviewed the Platform’s software and the Quoter Program and gave evidence on how the Platform calculated the position of margin traders and expressed the view that if the Quoter Program had been working properly there would not have been a margin call.

As regards the first aspect (the stub quotes issue), Mr Danny Ong contended that the evidence should be struck out primarily on the basis that it was not directed to the questions that the experts had been asked to address and that no application had been made for different issues to be the subject of expert evidence. Secondly, he said that Mr Kapoor was not qualified to give evidence on what was essentially a regulatory matter. Thirdly, the evidence went to an unpleaded issue of stub quotes. Lastly, the issue of market manipulation was a matter for the Court and not for a witness. He indicated that if the evidence was admitted it would be necessary for the Plaintiff to instruct a witness to respond to Mr Kapoor’s assertion.

Mr Paul Ong contended that Mr Kapoor’s qualifications (or lack of them) were not a ground for striking out the evidence. His degree of competence to give the evidence went to weight, not to admissibility. To a certain extent I agree with this unless it is apparent from the witness’ CV that he has no relevant qualifications. Had this been the only objection I would have allowed him to be cross-examined on his credentials and then ruled on admissibility.

He went on to submit that the evidence was relevant to understanding the nature of the strategy underlying the trading software, to demonstrate that it was not merely a defensive measure and had, as he put it, the more sinister objective of market manipulation. This had been pleaded in general terms and, in particular, was relevant to the issue of unconscionability under unilateral mistake in equity.

On the second aspect (the Platform and Quoter Program issue), the primary argument related to the fact that no issue concerning those programs had been included in the list of issues and that the code for the programs had not been disclosed or considered by the Plaintiff. It was therefore wrong at this late stage for the issue to be canvassed any further than had been done in the AEICs of the witnesses of fact.

The starting point on the stub quotes issue is that I was satisfied that justice could be done in this case without having expert evidence. The two experts have put in evidence saying how the B2C2 system works and why it works. Having heard all the evidence, including the factual evidence, it will then for me to draw the conclusion as to what were the motives behind B2C2's writing of the program in the way it was written. On the basis of that conclusion, I shall have to consider the question of unconscionability under Singapore law. In order to do this, it is not also necessary to reach a conclusion as to whether the writing of the program was or was not in breach of some particular regulatory requirements and my assessment on those motives cannot be delegated to an expert witness. Accordingly, I did not consider that I needed the assistance of experts in regulatory matters.

However, even if that was not the case, I concluded that it would be quite wrong to admit expert evidence on a new issue at this late stage. As I made clear at the earlier hearing, I consider that, in cases such as this, the expert evidence ought to be directed and only directed to specific defined issues. This enables the reports to be filed without unforeseen expert evidence being adduced which is likely to lead to delay and expense in dealing with that evidence in subsequent reports.

This is precisely what would happen in this case. Mr Paul Ong suggested that the stub quote issue could be hived off, B2C2 could adduce expert evidence in reply to Mr Kapoor and the matter could then be addressed at a subsequent hearing. I accept that Mr Kapoor had to prepare his report under certain time pressures and that his report contains material which he relies on as justifying the stub quote assertion, but this does not excuse the fact that Quoine could have sought leave to adduce expert evidence on the issue of market manipulation in advance of the service of his report.

For both these reasons, I ordered that the passages relating to the stub quotes issue should be struck out.

On the second issue, again I was satisfied that justice could be done without going into detail on Quoine’s programs. There is no dispute that the Platform did determine that the Counterparties were in breach of their respective margin limits nor is there a dispute as to when this happened. The reason why this happened was explained by Mr Lozada and it is not necessary to go into any greater detail as to the steps that the code goes through to reach this conclusion.

Equally, there is no dispute that the Quoter Program was inoperative after 13 April. Whether or not the consequent lack of liquidity would have occurred had it been operative is irrelevant.

Moreover, for the same reasons as are given above, I considered that it would be wrong to admit expert evidence on issues not covered in the current list of issues. Accordingly, I ordered that the passages relating to these issues should be struck out.

The facts

I propose to consider the facts under the following headings:

(a) The Quoter Program;

(b) The Platform and the Margin Calls;

(c) The B2C2 Trading Software; and

(d) The mindset of Mr Boonen.

Many of the facts are either common ground or not in dispute. In so far as any matter was the subject of dispute, I have reached an assessment on the basis of the balance of probabilities.

The Quoter Program

The causes of the incident on 19 April 2017 have their origin in some work done by Quoine on 13 April when changes were made to some login passwords for several critical systems on the Platform. This was done for security reasons. Due to an oversight, certain necessary changes to the Platform’s Quoter Program were not implemented.

The Quoter Program is responsible for retrieving external market prices from other exchanges which are then used to create new orders to be placed by Quoine on the Platform for market making purposes and to create liquidity. The Quoter Program is personal to Quoine and information generated by it is not available to other users of the Platform. The effect of the oversight was that it could not access data from those other exchanges and accordingly it became inoperative and stopped creating new ETH/BTC orders on the Platform. Unfortunately, this did not generate what is known as an exception message which would have alerted Quoine to the oversight as the notification flag for such messages had been disabled. Quoine did not become aware of the oversight until after the events of 19 April.

The Platform and the Margin Calls

At this time, Quoine was the principal market maker on the Platform. Mr Lozada estimated that it was responsible for around 98% of market making trades. When the Quoter Program became inoperative, there were a number of orders previously on the Platform placed by Quoine and others which continued to be available for trading purposes. During the period between 13 and 19 April 2017 however, the volume of trading on the ETH/BTC order book was low and it was not until 19 April 2017 that the gradual depletion of the pre-existing orders and orders placed by traders other than Quoine, including B2C2, reached a level which Mr Lozada described as being “abnormally thin” which precipitated a number of events.

So far as concerns the Platform, it had the effect of causing the mechanism for calculating the margin traders’ positions to detect that Pulsar’s and Mr Tomita’s accounts were in what is known as “Margin Sell-out Position” which served to trigger the margin calls resulting in the placement of the market orders to buy ETH at the best available market price. This occurred, in the case of Pulsar at 11.31pm and in the case of Mr Tomita at 11.53pm. This can be seen on a print out of the force-closed orders executed on 19 April 2017 (trader ID 5539 is Pulsar and trader ID 30742 is Mr Tomita). A copy of the print out forms Annex 3.

As Mr Lozada made clear in his evidence, the force-closed transactions were carried out notwithstanding that the available BTC balances in Pulsar and Tomita’s accounts were insufficient to meet B2C2’s orders at the limit price quoted by B2C2. This would not have happened if the Platform had a program to check whether there were sufficient available assets. There was however no such program at the time which Mr Lozada accepted in the course of cross-examination was the result of poor design:

Mr Tseung, who in the event was not cross-examined on his AEIC, gave evidence that at the relevant time the balance on Pulsar’s account on 19 April contained only 13.5256945100 BTC yet the Platform debited Pulsar’s account with over 3000 BTC. He stated as follows:

This evidence reflects the admission by Mr Lozada that the Platform should have had such a feature but did not. Had there been such a feature, the Platform would have traded only the 13.5256945100 BTC in Pulsar’s account, leaving any outstanding deficit on Pulsar’s account due to the lender, in this case Quoine, to be recovered outwith the operation of the Platform.

Returning to Annex 3, it can be seen that when the margin call was made against Pulsar the initial trades were effected with the lowest sell orders then on the order book. Some had been placed by Pulsar (5539) and others by Quoine (3). These were orders from Quoine placed before the Quoter Program became inoperative. The price of these trades started at 0.03949 BTC/ETH and slowly increased (with three exceptions of what were new orders placed by Pulsar at 0.03944) until the last and highest order placed by Quoine at 0.04233 was filled. This demonstrates how the trades work when a margin call is made. The Platform starts to purchase the relevant currency at the lowest available price and then works through the existing orders sequentially in increasing value terms.

By 11.31pm, all the pre-existing orders by Quoine had been used up and the system had to search for the next lowest price available. As can be seen, these ranged from 0.05 to 1 BTC/ETH from traders other than Quoine including one from Pulsar itself at 0.07. It is thus apparent that the loss of liquidity was causing the trading systems of other traders to react by placing higher quotes, 1 BTC/ETH being some 25 times the 0.04 level. Exactly why this was so is unclear.

Then at 11.52pm, orders at 10 and 9.99999 from Trader 5776 (B2C2) were placed on the Platform and, since there were no lower priced orders, these were filled to complete the margin call on Pulsar. The position with regard to Mr Tomita is a little more complicated but it is not suggested that anything turns on this.

The B2C2 Trading Software

The abnormally thin order book also had an effect on the B2C2 trading software. It was this issue that was the subject of the investigations by the expert witnesses.

The B2C2 trading software (“the Software”) is largely deterministic; it produces the exact same output when provided with the same input. Mr Atkinson explained that this was a benefit as it makes testing easier and the behaviour of the system is easier to understand, reason about and validate. It was devised almost exclusively by Mr Boonen and, as might be expected, had been the subject of a number of changes during the course of time. When the software was originally written, Mr Boonen included a number of routines designed both to maintain the integrity of the system and to ensure that it continued to operate in circumstances when there was insufficient liquidity in a given market for the system to operate normally.

The way in which the software determines what are appropriate prices at which to place Bid or Ask quotes is described in detail in paragraphs 42 to 52 of Mr Atkinson’s first report which was agreed to by Mr Kapoor. It is controlled by a strategy known as the PureQuote Strategy. For present purposes the effect of this strategy can be summarised relatively briefly. The software evaluates the first 20 price levels on a given platform on both the Bid and Ask side. It then excludes from that evaluation any price levels that relate to an order of a specified low volume. It then calculates an appropriate price at which to quote on either the Bid or Ask side as the case may be. However, it is accepted that there may be occasions when either the order book is empty or where it is populated by a large number of low volume orders within the first 20 that the strategy is unable to determine a price. One way to cater for this would be to introduce a circuit breaker so that the system would “error out”, ie, stop working. But this is undesirable in a program that is designed to work continuously without human oversight so Mr Boonen built into the software the requirement that at all times there should be added two “deep prices” to both the Bid side and the Ask side of the software’s internal representation of the order book. By this means there would always be prices which the software could draw upon so that the system would not error out.

Mr Atkinson described this in paragraph 54 of his first report:

These prices have to be chosen in a manner such that on the Ask side it is sufficiently large that the prospect of a trade at that level is both unlikely and, if it occurs, will adequately protect B2C2 against the risk of any adverse consequences of such a trade. Conversely, the chosen price on the Bid side needs to be low. As at 19 April 2017, these prices were, respectively, for the Ask side 10 BTC/ETH and for the Bid side 0.00001 BTC/ETH. As Mr Boonen put it, “the prices had to be sufficiently advantageous so that B2C2 would not regret it if orders place at or near those price levels were filled”. In cross-examination, he said that it was his “role as a trader to try to find numbers that make sense in a situation of very [sic] uncertainty as to the future”.

He also indicated that the choice of price had to be at a level that was not so low or high that the trading platform might be programmed to reject it and thus act as an undesirable circuit breaker.

Hence, when the order book became very low or empty in the late evening of 19 April the two deep prices on the Ask side came into effect and were placed on the order book. It is however to be noted that this did not happen immediately once the order book became illiquid. There was a delay because at the time B2C2 had sold sufficient ETH on other platforms in the recent past such that the software was not sending sell orders for ETH to the Platform to avoid accumulating too much risk.

Mr Atkinson described the placing of these deep prices as being a defensive programming technique. Mr Boonen identified this as being the role of a programmer to think of potential, even if unlikely events, and design the code to cater for the worst case scenario that could occur . Mr Kapoor accepted that it was in part defensive but said there were aspects that were not necessary for defensive purposes and Mr Paul Ong relied on these aspects as indicating an adverse motive on the part of Mr Boonen to take advantage of a situation of low or no liquidity by triggering margin calls.

The first point made by Mr Kapoor was that he considered that the software was configured to place a deep quote only on the Ask side and not on the Bid side. He did a simulation in which the deep prices on both sides were entered into the PureQuote strategy and the result was that an Ask quote was sent to the Platform but not to the Bid side because the “Bid quantity (was) too small”. This he contends is because the software imposes a minimum order quantity equivalent to USD20 based on an internal global benchmark (being a notional USD/BTC conversion rate of USD1200/BTC) which in his simulation led to an order quantity of only USD2.43. On the other hand, Mr Kapoor identified that there was no equivalent maximum order size preventing a large-sized order being sent out. Mr Kapoor contended that this could have been done.

Mr Boonen accepted that this could have been done and that it had not been. He however drew attention to the fact that there was a provision whereby the code does not send out orders that exceed a certain USD amount. In cross-examination, he said:

It can thus be seen that the software does not take into account the actual price of the order (ie, 10 BTC for 1 ETH). What it does is to have regard only to the quantity of the order in terms of ETH and multiply that by the internal benchmark for the USD value of ETH which at that time was USD50. If the sum of those two numbers exceeded a set value, then the order would not be placed. Hence the maximum figure did not relate to the value of the order but only the quantity and thus served to ensure that an order for sale of ETH did not exceed a certain quantity. There were, Mr Boonen said, very serious risk management reasons for this, a statement which was not challenged.

The position therefore is that a Bid order below a given minimum USD value based on the notional BTC/USD conversion rate would not be placed but an Ask order would be placed unless it was for a quantity of ETH above a given threshold based on that notional ETH/USD regardless of what the BTC proceeds of such a sale were. The former is to avoid meaningless low value orders coming onto the Platform. The latter is to protect B2C2 from trading in large quantities of ETH which, when the program was written, Mr Boonen regarded as being the more volatile currency and thus was the major risk . Mr Boonen did not include a maximum order limit on an Ask order to restrict the value of BTC which B2C2 might receive from any Ask order to sell a quantity of ETH which did fall below the maximum number. In the circumstances of this case, this would have acted to protect the Counterparties to B2C2’s detriment.

The second factor Mr Kapoor drew attention to was that there had been changes in the deep prices that had been set in the software since it was first written. When Mr Boonen initially wrote the program, he configured the software to quote a deep price on the Ask side at a BTC/ETH ratio which was close to infinity and subsequently altered this to a price which was equivalent to a ratio of 50 BTC/ETHJHH. On 16 June 2016 Mr Boonen altered the deep price to the 10 BTC/ETH level. Mr Kapoor suggested that this change was indicative of an intention on the part of the programmer to increase the likelihood that such deep prices would be executed. The basis for this was Mr Kapoor’s assertion that higher prices were likely to be rejected by the Platform whereas rejection at the 10:1 level was less likely.

Mr Boonen explained the reasons behind this change in his third AEIC and was then cross-examined at some length on it. Mr Boonen explained that in June 2016 there was what he called a refactoring of the code with the intention of making it clearer and safer. Whilst the close to infinity level had previously been altered to place another somewhat smaller but still large number for the deep price, both were a “one size fits all” number in that it applied across the board to all currency pairs traded by B2C2. It so happened that at that date the effect was that the BTC/ETH rate was around 50:1.

By 2016, B2C2 were trading in an increased number of trading pairs including ETH/BTC (ETH having come into existence after the B2C2 trading software had first been written in 2014). Mr Boonen particularly drew attention to the fact that B2C2 had by then started trading in BTC/South African Rand, a significantly more volatile currency than the USD and that he had concerns that the generic deep price might not be deep enough to protect B2C2 in the event of a significant change in the value of the Rand.

He therefore concluded that the “one size fits all” number was no longer appropriate and he placed individual deep prices for each currency pair. Mr Boonen considered the 50BTC:1ETH ratio was too high and he selected the 10BTC:1ETH level. He accepted that one of the reasons for doing this was to ensure that any sell orders placed at that level would not be rejected by the Platform. He gave evidence that the price was chosen so as to be sufficiently advantageous to cover the risks to B2C2 of trading in a potentially illiquid market and again emphasised that he regarded ETH as being the greater risk at the time. He also explained that he used the same deep price to act as a circuit breaker to protect either against the risk of a significant change in the actual market price of a given currency pair or against the possibility that there might be bugs in other parts of the B2C2 trading system. He asserted that it was his role as a trader to find numbers which made sense to him in an uncertain market.

In re-examination, he amplified upon the circuit breaker aspect by giving an example of a case where with a different currency pair the actual market prices broke through the upper range and the deep price acted as a circuit breaker. He indicated that he felt that the 50:1 ratio was so high that it would be highly unlikely ever to be activated as a circuit breaker and thus would not adequately protect B2C2 and that he also chose the 10:1 ratio with this in mind. I accept this evidence.

The third factor to which Mr Kapoor drew particular attention was his assertion that certain portions of B2C2’s code were changed after the glitch occurred on 13 April 2017 which he suggested might have caused the software to quote more frequently or to quote more in total than it was quoting before the glitch. The underlying thrust behind this suggestion was that B2C2 had appreciated sometime between 13 and 19 April that something was causing illiquidity on the Platform and that the changes were implemented with a view to triggering force-closures.

As Mr Kapoor fairly acknowledges, he was unable to review the actual code for the suggested change, but he surmises that it did occur since the initial cycle frequency was set at 10 seconds and by 19 April the frequency was 5 seconds.

In his third AEIC, Mr Boonen rejected the implicit suggestion that any change to B2C2’s code configuration files was in response to or motivated by the glitch which he contended did not come to his notice until after the orders were fulfilled on 19 April. He went on to explain that the amendments made on 13 April were part of a routine change to the Inventory Strategy which adjusts prices to maintain a target balance of B2C2’s assets on the Platform.

Mr Atkinson also considered this suggestion in detail in his second report. He concluded that there was no evidence to suggest that there were any configuration changes made in April other than those which reflected configuration changes in inventory on the Platform. All the witnesses were cross-examined on this issue.

Mr Boonen was challenged on his evidence that he did not become aware of the glitch until after the orders were filled. He explained that B2C2 was connected to 15 exchanges at the time, on some of which up to ten pairs were traded. The market for BTC/ETH on the Platform was relatively small and he would not look at it on a day-to-day basis so that it was unlikely that he would have checked it on 13 April. In any event, he said, the Platform did not allow him to see the actual orders of others on it at any given time and there was no access to the number of orders at any one time. He rejected the assertion that he did notice a drop in the volume of orders on the ETH/BTC order book on 13 April. The reasons he gives for not noticing this are cogent and again I accept his evidence.

Mr Atkinson and Mr Kapoor metaphorically crossed swords on the technical aspects of this issue. Since I have accepted that Mr Boonen did not notice the drop in the volume of orders on 13 April, it is less important that I consider the technical aspects in detail. In brief, Mr Atkinson contended that Mr Kapoor’s analysis was flawed in that it sought to determine an undefined average cycle frequency but failed to distinguish between Bids and Asks and failed to consider the mode average. He pointed to the fact that that there was a large difference between the mean and the mode average which suggested that the approach of Mr Kapoor was not appropriate. He concluded that the data available to him, properly analysed, showed that the cycle frequency was set to 5 milliseconds before 13 April and remained the same through until 19 April.

In cross-examination, Mr Kapoor accepted that in the absence of full details of the changes implemented on 13 April he could not be certain that changes had been made which would reflect an intention to take advantage of the drop in orders but maintained that this was a possibility.

I consider that the data analysed by Mr Atkinson and his reasoning are sufficient to satisfy me that the changes made on 13 April were not changes designed to take advantage of any perceived loss in the order book. I do not question that Mr Kapoor genuinely held a suspicion that the changes might have been for an ulterior motive but, having read and heard all the evidence, I am unpersuaded that this was the case.

The mindset of Mr Boonen

For reasons which I shall amplify upon when considering the law of mistake, I have concluded that when the law is faced with a contention that a contract made by and between two computer systems acting as programmed but otherwise without human intervention is void or voidable for mistake, it is necessary to have regard to the mindset of the programmer when the relevant programs, or the relevant part of those programs, were written. The knowledge of the programmer in question is to be inferred by the Court from the evidence adduced and from all the surrounding circumstances.

In the case of the seven contracts made on 19 April 2017, the offer for sale at the 10 BTC/ETH price was made by the B2C2 trading system. It was accepted by the Platform in filling the market orders already placed on the Platform by the action of a separate program which had identified that the Counterparties’ collateral was insufficient. The Counterparties in their loan arrangement with Quoine had agreed that this could be done if the Platform determined that the collateral was insufficient. There was no human intervention in this process. Mr Boonen, Mr Lozada and the Counterparties were unaware what had happened until after the event.

Mr Boonen had written the trading software originally in 2014 and the 10:1 ratio had been included in June 2016. Quoine’s case on knowledge is set out in the particulars to Paragraph 14 of the Defence. It alleges knowledge in two respects. First, Particulars (a)–(e) focus on the Plaintiff’s alleged knowledge at the time of the incident, seeking to draw the inference that the prices at which the orders were fulfilled were so abnormal that a reasonable person (and hence the Plaintiff) would have known or harboured a real suspicion that the rate was inaccurate and therefore a mistake. However, it is now clear from the evidence that no one was “in the know” at the time of the incident what orders were on the order book when the margin trades were placed or what the prices of those orders were. Further it is clear that the placing of the orders at the 10:1 ratio was not itself the cause of the force-close margin orders being placed. Accordingly, this aspect of knowledge has not been demonstrated.

The second aspect is pleaded in Particular (f) which reads as follows:

This is a clear allegation that Mr Boonen included the 10:1 ratio when writing the program with the knowledge not only that it would lead to orders being placed on the Platform at that level in cases of low liquidity but also with the knowledge that there were circumstances in which such orders might be met. Quoine amplified upon this in its written opening submissions. The suggestion that the placing of the deep prices was strictly the result of an innocuous defensive programming strategy was said to be highly artificial and strained. It was also suggested that the Plaintiff was being opportunistic in seeking to take advantage of forced and/or erroneous liquidations by programming its trading software to place orders at the deep prices which were not bona fide. Mr Atkinson’s analysis, it was said, was overly simplistic and failed to take into account the overall manner in which the code had evolved over time. Reliance was placed on the evidence of Mr Kapoor to support an allegation that Mr Boonen had put thought into what the trading software should do if the ETC/BTH order book became empty and that he had placed the 10:1 ratio into the code with the intention of triggering and thereafter benefitting from such forced or erroneous liquidations.

In its Reply, the Plaintiff denied that it had any knowledge of the trades at the time they were made and asserted that it was speculative to assert that it had “any form of knowledge of any purported system failure, loss of order book, force-closure or otherwise of any alleged mistake” and denied that its trading software was programmed with the intentions and/or objectives alleged in particular (f) to paragraph 14 of the Defence.

In his written evidence, Mr Boonen said that he was not aware that the Platform was affected by any glitch, that the trading software was not programmed to detect such glitches and that the orders were placed because of the deep pricing strategy referred to in [83] above. He accepted that there was always a possibility that such orders would be filled and that, if they were, B2C2 would be bound by them. It was for this reason that he pre-programmed the code so that the prices were sufficiently advantageous to cover the risk to B2C2 of uncertain liquidity and market conditions. He frankly stated that “the price levels had to be sufficiently advantageous so that B2C2 would not regret it if orders placed at or near those levels were filled”.

Early in his cross-examination, when considering the extent to which the deep prices were present to cover the risks of illiquidity Mr Boonen said this:

In its written closing, Quoine repeatedly referred to this passage as being an indication that Mr Boonen must have known that if the very unusual or unfathomable were to occur this was likely to be the result of a serious error or glitch.

Thereafter, Mr Boonen gave evidence that he was surprised when he first saw that the orders had been filled and that he considered the possibility that there was an error on the Platform or an error in the B2C2 system. Having checked that there were no apparent errors in either system he thought “Wow, there must have been something incredible happened [sic] in the market overnight”.

Later on, he accepted that he was aware when programming the code that it was possible that such a quote could be transacted and executed when the order book was empty but that this was an unlikely possibility. However, he rejected the suggestion that he knew that putting in an order at such a high price could potentially cause traders who were short on ETH to close out their positions or for exchanges to force-close ETH short positions.

Matters came to a head in this passage of cross-examination:

Mr Boonen was affronted at the suggestion that he had designed his software with the specific purpose of taking advantage of the errors of others and, in the result, he did become over-loquacious but the thrust of his answers was always to the effect that Quoine were seeking to read far too much into the insertion of the deep prices than was warranted. His position can be illustrated by two further extracts of the oral evidence he gave on the second day of his trial:

What then was his mindset? At the time the original program was written, ETH did not exist. In June 2016 when the 10:1 ratio was added, it was a fledgling currency and I accept Mr Boonen’s evidence that he regarded it as representing the most risk. I also accept that his primary concern when writing the program was to protect the integrity of the B2C2 trading system so as to minimise the risk of any unwarranted exposure.

He was particularly careful because of the fact that the markets in virtual currencies were unregulated. He saw the need to protect B2C2 and anticipated that other traders and the platform operators would take similar steps. He appreciated that the way he had written the software, the deep prices could go on the Platform and that it was possible, although unlikely, that a trade would be executed at that level. It would only be, as he put it in the passage quoted above, when something very unusual happened or when the unfathomable actually crystallised. He frankly conceded that were this to happen he would hope to make a profit and that the price chosen was selected so as to maximise the possibility of not making a loss. His strategy was therefore not wholly defensive.

Mr Paul Ong submitted that the facts supported the conclusion that when Mr Boonen configured the code he knew of the possibility that the order book might become empty and that in that event there was a real possibility of the deep price orders being executed. He therefore designed the software to take advantage of that. It was sufficient, he said, for me to conclude that Mr Boonen programmed the code to include a deep price which was irrational and that no trader would agree to buy ETH at such a price unless by mistake.

I do not believe that Mr Boonen turned his mind in any detail to the circumstances in which such an order might be executed although plainly he had well in mind that these virtual currencies were highly volatile. Equally, at a high level of generality, he would have appreciated that the deep prices would only be likely to be executed when there was illiquidity in the order book for that was one of the reasons underlying the addition of the deep prices in the software in the first place. His reasoning was however directed to ensuring the B2C2 system remained operational in such circumstances rather than exploiting the existence of illiquidity.

He might have appreciated that one contributing factor to illiquidity could be due to an oversight or error on the part of someone but there are no grounds for concluding that he considered that these would be the only circumstances that would lead to illiquidity. His attitude was well demonstrated by his reaction on seeing the outcome of the trades, when his first concern was that there might have been an error in the B2C2 code and then that there might have been an error in the Platform.

I therefore accept that when Mr Boonen configured the code he knew of the possibility that the order book might become empty and that in that event the deep prices would be placed on the order book but I do not accept that he ever considered that there was a real possibility of the deep price orders being executed. He considered that this was unlikely and I do not believe that he turned his mind in any detail to the circumstances in which this might happen. That was not the motivation for designing the software as he did. He did not consider that the deep prices were irrational. Indeed he gave cogent reasons for the selection of the prices which were directed to protecting the integrity of his system but at a level which he felt would not be so high as to cause a circuit break on the Platform: see [86] above. This last factor is inconsistent with any assertion that Mr Boonen understood that trades would only be matched by the Platform at prices at or near the previously existing prevailing market price: see [230] below.

The evidence does not support the conclusion that when designing the code Mr Boonen perceived that no trader would ever agree to buy ETH at such a price unless he had made a mistake and included the deep prices with this in mind. More specifically, there are no grounds for concluding that he ever turned his mind to the relationship between the margin traders and those who had loaned them money, far less that he programmed the code with a view to exploiting that relationship by causing the deep prices to lead to a margin call and hence a force-close position. Equally there are no grounds for concluding that Mr Boonen wrote the code in the way that he did with an underlying intention of manipulating the market. I therefore reject Quoine’s assertions in those respects.

His mindset therefore, in summary, was that he intended the 10:1 ratio to act primarily to protect B2C2 against the risks of illiquidity but that he was aware of the unlikely possibility that trades might be executed at that level. He did not turn his mind in any detail to the circumstances that might lead to such trades being executed but he would have been aware that one possibility was that it could be the result, wholly or in part, of some error or omission on the part of someone, including himself. He intended that B2C2 should be best placed to make a profit, not a loss, if the unlikely became a reality, whatever the cause. In this respect his programming was opportunistic but it is in no respect sinister. It is not sinister to give yourself the opportunity of making a profit rather than a loss in the case of a wholly unlikely event occurring for whatever reason.

The Contractual Relationships

There is more than one contractual relationship that will exist when parties trade on a currency platform. First, all traders, whether buyers and sellers, will have a contract with the platform operator to regulate the relationship between that trader and the platform owner so that they can trade on the platform (“Platform contracts”). It is the Platform contract between B2C2 and Quoine that Quoine is alleged to have breached when it reversed the trades. Secondly, where there are margin traders, there will be a separate contract between the borrower and the lender (in this case Quoine on the one hand and the Counterparties on the other) which will regulate their lending relationship (“Margin contracts”). Finally, when a trade is executed, the buyer and seller will have some form of contractual relationship inter se (“Trading contracts”).

Quoine contended that the relationship between buyer and seller in a trading contract was a direct one. In the case of the seven trades, B2C2 had placed a limit order offering to buy BTC and sell ETH at the rate of 10 BTC for 1 ETH. The Counterparties had previously placed a market order to sell BTC and buy ETH at whatever prices were available on the order book at the time. The orders were matched by the Platform acting as programmed to perform an exchange function so as to complete the contract for those trades between the respective parties by the operator of the Platform, Quoine, which was not itself a party to the contract. The fact that neither party knew the identity of the other was irrelevant. Pursuant to the Platform contracts, all traders knew that any orders would be matched with other traders who were subject to the same Platform contract.

Mr Danny Ong for B2C2 submitted that the relationship was, as he put it, a spider’s web of contracts with Quoine in the middle. Quoine was thus a party to a contract with the buyer and a separate contract with the seller, acting as an intermediary. He contended that Mr Boonen had accepted in evidence that B2C2 always considered that the contractual counterparty to trade on the Platform was Quoine and that Quoine must necessarily act as a central counterparty to both customers since the Platform operated on an anonymous basis.

It is relevant to decide this issue as it will be necessary to identify what the mistake was, if there was one, and who the mistaken party was. It should, however, be noted that the answer to this question has no bearing on whether or not Quoine is liable to B2C2 for breach of the Platform contract.

Resolution of this issue requires the Terms and Conditions of the Agreement to be interpreted. These make it plain that Quoine as the Platform operator is providing a service to the “Users and Members” (hereinafter “Members”) using the Platform. The first paragraph of the Section headed “Trading & Order Execution” reads:

And the third paragraph reads:

The remainder of the Terms and Conditions make it plain that Quoine is solely providing the Platform and that the parties are themselves responsible for determining whether and on what terms they shall place or fill orders. The actual contract of sale is therefore a contract directly between buyers and sellers whose rights inter se are determined by the Terms and Conditions. In my judgment, therefore, Quoine’s approach is the correct one. The Members have agreed that they can use the Platform to buy and sell currencies using limit or market orders on the understanding that if the Platform matched two orders a contract would ensue between those Members. Each of the offers to buy and sell made by B2C2 and Quoine was made at large to all other users of the Platform. When the orders were matched the identities of the actual buyers and the sellers were ascertainable.

Since it is irrelevant in law who Mr Boonen considered he was contracting with, I shall not review the evidence that B2C2 drew my attention to. Suffice it to say that I do not consider that it is at all clear that Mr Boonen had the impression that he was contracting solely with Quoine.

The Issues

There are three main issues to be canvassed:

(a) First, the Plaintiff’s Claim. There are essentially two prongs to the claim: (1) Quoine’s breach of contract in reversing the trades; and (2) Quoine’s breach of trust in disposing of the Plaintiff’s assets.

(b) Second, the Defendant’s Defences. The Defendant raised a number of potential defences which, logically, I consider should be considered in the following order:

(i) It is entitled to reverse the trades on the basis of implied terms of the Agreement.

(ii) It is entitled to reverse the trades on the basis of an express term of the Agreement when read in conjunction with the Risk Disclosure Statement.

(iii) It is entitled to reverse the trades on the basis that the contacts between B2C2 and the Counterparties were void under the doctrine of unilateral mistake at common law.

(iv) It is entitled to reverse the trades on the basis that those contracts were voidable under the doctrine of unilateral mistake in equity.

(v) It is entitled to reverse the trades on the basis that those contracts were void under the doctrine of mutual mistake at common law.

(vi) It is entitled to reverse the trades on the basis of unjust enrichment.

(c) Third, the relief to be ordered if Quoine is liable. In particular, If B2C2 is entitled to relief both in law for breach of contract and in equity for breach of trust, what relief is it entitled to?

Bifurcation

I do not have to address the issue of damages or monetary compensation in this judgment because the parties agreed that the proceedings would be bifurcated into findings on liability and damages, respectively. By a consent order dated 20 February 2018 it was ordered that:

Accordingly, if B2C2’s claim succeeds in any respect, the amount of any monetary compensation will be determined at a later hearing. As indicated at the hearing, this will include resolution of an issue concerning an alleged limitation of liability pursuant to Clause (d) under Owner Responsibilities in the Agreement.

The Plaintiff’s Claim

Breach of Contract

B2C2’s case is straightforward and short. The seven trades were executed as a result of the market orders placed (albeit unknowingly) by the Counterparties which were filled by the limit orders placed by B2C2 at the deep price of 10 BTC for 1 ETH. The proceeds were debited from and credited to the relevant accounts for the amounts of ETH and BTC respectively. The parties were notified of this in the usual way. The Agreement expressly provides that “once an order is filled, you are notified via the Platform and such an action is irreversible”.

On that basis, B2C2 contends that the reversal of the orders on the following day was in breach of the express term that the filling of an order is “irreversible”. Unless Quoine succeeds on one or more of its defences, this claim will succeed.

Breach of Trust

Quoine contended that it was clear that it did not hold any assets belonging to any Member on trust for that party. There was no provision to that effect in the Agreement and paragraph 9 of the Risk Disclosure Statement, “Bankruptcy Risk” expressly alerted the customer to the fact that although it managed any assets deposited by Members separately from Quoine’s assets, those assets were held by Quoine itself and were not on deposit in a separate account with a trust bank. Accordingly, if Quoine were to go bankrupt, it would be unable to return those assets. This was inconsistent with the assets being held on trust.

B2C2 drew my attention to the answers to a Request for Further and Better Particulars of the Defence which stated that all funds deposited by all traders were stored in a single cryptocurrency wallet which was owned and maintained by Quoine. This was consistent with a provision in the Terms and Conditions under “Withdrawals” which made it clear that in order to protect Members no currencies were kept in online wallets and was also consistent with the statement in the Risk Disclosure Statement that Member’s assets would be managed separately from Quoine’s assets. This was, he said, evidence of an intention to create a trust.

The factual position is therefore that whilst separate accounts are maintained for all Members, the actual virtual proceeds are stored together off line but are maintained separately from Quoine’s own assets.

So far as concerns the law, there was a measure of agreement between the parties. Both accepted that there were three certainties that had to be present for the creation of a trust, namely, certainty of intention, certainty of subject matter and certainty of objects.

It is convenient to consider the second certainty, certainty of subject matter, first. Quoine was prepared to assume that cryptocurrencies may be treated as property that may be held on trust. I consider that it was right to do so. Cryptocurrencies are not legal tender in the sense of being a regulated currency issued by a government but do have the fundamental characteristic of intangible property as being an identifiable thing of value. Quoine drew my attention to the classic definition of a property right in the House of Lords decision of National Provincial Bank v Ainsworth [1965] 1 AC 1175 at 1248:

Cryptocurrencies meet all these requirements. Whilst there may be some academic debate as to the precise nature of the property right, in the light of the fact that Quoine does not seek to dispute that they may be treated as property in a generic sense, I need not consider the question further.

The third certainty, certainty of objects, requires that the intended beneficiaries have to be identifiable so that it is possible to ascertain those who have the standing to enforce the trustee’s duties under the trust: Guy Neale v Nine Squares Pty Ltd [2015] 1 SLR 1097 (“Guy Neale”) at [60]. This requirement is met in the present case. The beneficiaries are identifiable from the individual accounts of each of the Members.

The dispute arises on the first certainty, certainty of intention to create a trust. In ascertaining the requisite intention, it is not necessary for express words to that effect to be used. The Court must have regard to the conduct of the alleged settlor, the words used in any relevant documents and all the surrounding circumstances: Guy Neale at [52]–[58].

In the present case there are no express words in the Agreement creating a trust. The necessary intention is thus to be determined from the wording of the document as a whole and from Quoine’s conduct when handling the assets. That practice has not changed since the Platform began to operate. More specifically, its practice did not alter when the Risk Disclosure Statement was uploaded onto the website in March 2017. To my mind, the decisive factor is that the assets are held separately as Member’s assets rather than as part of Quoine’s trading assets. This is a clear indication, not surprisingly, that Quoine claims no title to those assets and acknowledges that it is holding them to the order of the Member who can demand withdrawal at any time. This is sufficiently clear evidence that Quoine intended to hold the assets on trust for the individual Member. What will be the effect of such an arrangement were Quoine to go “bankrupt” is not a matter for me to decide.

Accordingly, if Quoine was not entitled to reverse the trades, the reversal and hence the unilateral removal of the Bitcoin from B2C2’s account was in breach of trust.

The Defendant’s Defences

Defence 1: Is Quoine entitled to reverse the trades on the basis of implied terms of the Agreement?

Quoine accepted that, as a matter of law, a term could not be implied into a contract if it contradicted an express term of that contract. If it did not contradict an express term, then a term could be implied if it was necessary to give business efficacy to the contract and to give effect to the intentions of the parties.

In paragraph 152 of its written closing submissions, Quoine sought to rely on two implied terms:

(a) Quoine may reverse any trades which are executed at any abnormal rate or price as a result of any technical and/or system failure and/or error affecting the Platform;

(b) Quoine may reverse any trades resulting from orders placed in breach of the terms of the 2014 Terms & Conditions, including any trades resulting from any orders which amounted to market manipulation and/or abuse and, therefore, an “unauthorised use” of the Platform.

There was no dispute between the parties as to the legal approach. In particular, Quoine accepted that the applicable principles were laid down by the Court of Appeal in Sembcorp Marine Ltd v PPL Holdings Pte Ltd [2013] 4 SLR 193. In paragraph 153 of its written closing submissions Quoine correctly summarised the position as follows:

B2C2’s primary contention in relation to the first proposed term is that such a term would contradict the express term in the Agreement under “Trading & Order Execution” that:

This is, it is said, indicative that the parties had turned their minds to the question of the reversibility of completed orders and had concluded that this should not be permitted. There was therefore no gap to be filled by the proposed implied term permitting reversibility in particular circumstances.

Quoine contends, however, that “irreversible”, properly interpreted, does not preclude Quoine, as opposed to the contracting parties, from reversing trades. Whereas the individual parties to the transaction should not be able unilaterally to reverse a completed trade, giving the clause a purposive construction, the same was not true for Quoine. I am unable to accept this submission. The Platform contract is between Quoine and any given Member of the Platform. The immediately preceding sentence expressly limits Quoine’s liability in relation to delays in execution and the following sentence again limits its liability in terms of loss or damage incurred by a Member as a result of its “failure to understand the nature and mechanics of virtual currencies or the markets under which such virtual currencies operate”. Furthermore, the wording of the clause itself relates not only to the filling of the order but also to the notification via the Platform, which is also consistent with all parties, including the Platform, being bound by the trade.

The word “irreversible” is not qualified in any way and when it is read in the context of the passage as a whole, the proper inference to be drawn is that the provision was included to ensure certainty for all parties including Quoine as soon as the transaction was notified as being filled. To interpret “irreversible” in a manner such that orders executed at what Quoine considered to be at an abnormal rate could subsequently be reversed thereafter runs contrary to this. Accordingly, the terms sought to be implied do contradict an express clause of the Agreement and therefore cannot be implied.

Even if this were not the case, I do not see that it is necessary for the terms to be implied to give business efficacy to the Agreement. The Agreement seeks to provide clarity as to the trading relationship between Quoine and the Members and between the Members inter se and, in doing so, to apportion risk clearly between the parties. In the case of a new trading platform clarity is essential to give confidence to traders, especially market makers who are trading on a minute-by-minute basis over a number of exchanges and who regulate their exposure across those exchanges. Any uncertainty on one exchange can have a material effect on its position overall. Mr Boonen gave evidence that if he had been asked to agree a term that Quoine could unilaterally reverse a trade, he would not have done so because he was concerned about Quoine as a platform operator.

The clause which provides for irreversibility does introduce certainty and places the risk of entering any given trade on the parties to that trade. Where one knows that the filling of an order brings finality to a transaction, the parties can thereafter carry out their business in the knowledge that the past will not be undone. They must however ensure that procedures are in place to ensure that any risks involved in algorithmic trading are guarded against. Mr Boonen suggested that in this field, “caveat emptor” applied. In truth I consider that in an unregulated field, it is in fact “caveat omnes”. There is thus no need to imply the terms to give business efficacy to the Agreement.

Defence 2. Is Quoine entitled to reverse the trades on the basis of an express term of the Agreement when read in conjunction with the Risk Disclosure Statement?

There are four relevant passages in the Agreement. First, under the heading “General Terms” it is provided:

Secondly, under “Membership and Users of the Platform” it is provided:

Third, clause (b) of the “Representation and Warranties” section provides:

Fourth, clause (h) of the “Representation and Warranties” section provides:

Clause (h) is a somewhat unusual clause which permits Quoine unilaterally to alter the Agreement without notice but it unclear as to precisely how this can be done. The use of the expression “rights, obligations, privileges” in addition to the word “terms” introduces a measure of uncertainty as does the reference not merely to the terms of usage but also to the information which may be posted. The last sentence, however, refers solely to the “terms as modified”. Clause (h) encompasses the first two passages cited above and thus it is only necessary to consider the scope of that clause.

The Risk Disclosure Statement does what its name suggests. It is entitled “Risks in Virtual Currency Transactions” and the introduction states:

Ten separate risks are described, two of which are material in this case. The first, “Risks due to Stop-Out”, relates to the right of Quoine unilaterally to place a market order to force-close a margin trader’s positions in the event that the trader’s liability exceeds its collateral:

The second, “System Risks”, applies to both spot and margin traders:

Quoine contends that when the Risk Disclosure Statement was uploaded onto the website on 22 March 2017, it had the effect, inter alia, of introducing a new term into the Agreement expressly permitting the Defendant to cancel a transaction if it had taken place based at an aberrant value (the “aberrant value clause”). It accepts that no express notice was placed on the website drawing attention to the fact that the Agreement had been modified as a result of the uploading of the Risk Disclosure Statement. It contends, however, that on its true interpretation the combined effect of clause (h) with the preceding clause (b) of the Agreement does not require this to be done. The mere uploading of the Terms and Conditions was sufficient to modify the Agreement to incorporate the aberrant value clause which automatically came into effect without notice to the users of the Platform and regardless of whether they had seen it or not. The Risk Disclosure Statement constitutes “information” and/or “terms of usage” within the meaning of clause (h) because it changed the terms on which users could trade on the Platform and altered their rights. Finally, it contends that the seven trades on 19 April 2017 were based on an aberrant value so it was contractually entitled to reverse the trades.

B2C2 disputes these contentions. It asserts:

(a) that the Risk Disclosure Statement was not within the scope of clause (h); more specifically that since the Risk Disclosure Statement and the Agreement were accessible via different links on the website, the two could not be read together and there was no other indication on the website that the Risk Disclosure Statement did contain additional terms of the Agreement;

(b) that the Risk Disclosure Statement was a mere summary of risks and was not a document which a reasonable reader would expect to contain contractual terms; more specifically that the aberrant value clause was too vague and imprecise to constitute a contractual term and was inconsistent with the express irreversibility term;

(c) that clause (h) was contrary to section 3(2)(b)(i) of the Unfair Contract Terms Act (Cap 396, 1994 Rev Ed) (“UCTA”); and

(d) that under common law where a condition is particularly onerous and unusual, the party seeking to enforce it must show that the condition was brought to the notice of the other party, relying on Interfoto Picture Library Ltd v Stiletto Visual Programmes Ltd [1988] 2 WLR 615.

The following preliminary matters therefore need to be considered.

(a) First, what is the effect in law of a unilateral variation clause such as clause (h)?

(b) Second, when can an alleged term be incorporated into an existing contract in circumstances where that term is contained in a separate document and itself has no independent contractual effect?

As to the first, unilateral variation clauses are not unlawful per se, in the sense that a party can reserve to itself the right to amend a contract without first obtaining the consent of the other party, but it is an unusual power and thus there must be clear language to reserve this sort of power: see Wandsworth London Borough Council v D’Silva [1998] IRLR 193 and OCBC Capital Investment Asia Ltd v Wong Hua Choon [2012] 2 SLR 311. I accept therefore that clause (h) is a lawful clause, the meaning and scope of which is a matter of contractual interpretation. Whilst it is not necessary for the consent of the other party to be obtained to any such modification, it must be a requirement that the other party has the means of knowing that there has been a modification and what that modification is. Again the way in which this is to be done will be a matter of interpretation of the clause in question.

As to the second, it was contended that it was not legally necessary for a document, parts of which contained terms which were said to be incorporated into a contract, itself to have contractual force so long as the particular terms sought to be incorporated were capable of having contractual effect. Reliance was placed on passages in Sir Kim Lewison’s The Interpretation of Contracts (Sweet & Maxwell, 6th Ed, 2015) (“Lewison”) at pp 127–128:

I therefore accept that it is legally permissible for a term of a contract to be contained in another document which is not itself a document having contractual effect, but it must be clear that the term in question is intended to have contractual effect and the term itself must have sufficient clarity to be contractually enforceable. Again this will be a matter of interpretation.

However, in the case of the employment law authorities cited in Lewison, the contract of employment expressly referred to the fact that some additional terms and conditions were to be found in the other document and in the Hyundai case the contention was that the previous dealings between the parties had served to incorporate some additional terms from another document (the pro forma charterparty) into the alleged agreement even though it had no independent contractual effect. The parties have not referred me to any authority or work of reference which considers the question that arises here; where it is contended that a document which is not identified in the main agreement as containing additional terms of the contract or has not come by the dealings between the parties to be accepted as containing such clauses is said to contain an additional term, being a term which constitutes a fundamental variation to the existing terms of the contract.

I do not consider that it would be legally impermissible for such a term to be incorporated from a document which had not been previously accepted by the parties for one reason or another as containing additional clauses, even if those clauses were such as to introduce a fundamental change in the contractual relationship. In my judgment, clauses could be incorporated from such a document in an appropriate case but it is likely to be an exceptional case where a document which had not previously been specifically identified by the parties as likely to contain any terms would be effective to introduce a term modifying an existing term. It will be a question of fact in each case whether the document in question could properly be said to be one which the parties intended and understood to be capable of effecting such a change.

With that background I can turn to consider the interpretation of clause (h). It must be read as a whole in the context of the Agreement and in the light of the surrounding circumstances when the Agreement was made: Zurich Insurance (Singapore) Pte Ltd v B-Gold Interior Design & Construction Pte Ltd [2008] 3 SLR(R) 1029; Y.E.S. F&B Group Pte Ltd v Soup Restaurant Singapore Pte Ltd [2015] 5 SLR 1187. The first sentence reads:

This sentence is clear on its face. It gives the Company the right to alter the Agreement without the consent of the Members and it is not required to give express notice to them.

The fact that it is express notice that is being excluded is apparent from the subsequent sentences. The second sentence reads:

which makes it clear from this that any changes can be communicated indirectly by being posted on the website.

This is confirmed by the third sentence:

which serves to highlight that notice of change may be communicated directly to each Member by e-mail or indirectly by posting it on the website. The Member is then given the opportunity to terminate its membership in consequence.

This clause therefore indicates the way in which each Member is to be informed of changes. It can be done by posting the changes on the website. On its true interpretation the clause does not permit Quoine to change the terms without drawing this to the attention of the Members in some way. It is not enough merely to upload an amended agreement and expect the Members diligently to review the Terms and Conditions on the website at regular intervals and seek to discern any changes. Read as a whole, I have concluded that the use of the word “information” in the second sentence means information relating to the terms of usage, such as a notice on the website that the terms have changed.

But this does not answer the question of how the requisite notice can be given. The most natural way of doing it, other than by way of an e-mail directly to each of the Members, would be either to upload the new agreement containing the modified or additional terms together with a notice on the website drawing attention to the fact that changes had been made or, alternatively, to set out the changes in full in a notice on the website.

Quoine did not do this but instead merely uploaded the Risk Disclosure Statement onto the website. It did not otherwise draw attention to the fact that this document was intended to contain modifications to the Agreement. The following questions therefore arise on the first two contentions of B2C2 set out in [162] above.

(a) Could uploading the Risk Disclosure Statement onto the website on 22 March 2017 have served to amend the Agreement pursuant to Clause (h)?

(b) If it did, is the aberrant value clause sufficiently clear to be enforceable?

(c) Finally, do the provisions of clause (b) affect matters?

(a) Could the uploading the Risk Disclosure Statement onto the website on 22 March 2017 have served to amend the Agreement pursuant to Clause (h)?

As to the first of these, the starting point is that at the date B2C2 became a Member there was no Risk Disclosure Statement. The only document which could contain terms of the contract between the parties was the Agreement. Hence a Member reading the Agreement at the date it was signed by B2C2 would perceive that the only “terms, rights, obligations, privileges” that could be changed were those in the Agreement and would therefore naturally assume that any changes would be incorporated directly into the Agreement. There is no evidence that the Members were ever informed that changes would be implemented in any other way. It is therefore necessary to have regard to the events on 22 March 2017 when the Risk Disclosure Statement was uploaded to see whether this makes it clear that the Risk Disclosure Statement was one which the parties intended and understood to be capable of effecting such a change and which indicated with sufficient clarity that changes were to be made.

Quoine’s then webpage looked like this:

The Risk Disclosure Statement can be found by clicking on the “Risks in Virtual Currency Transactions” link seen at the bottom of the right hand column whereas the Terms and Conditions (the Agreement) are towards the top of the left hand column. There is nothing in either document which invites the reader to read the two together. On its face the Terms and Conditions plainly are intended to have legal effect. The introduction states “This agreement is entered into by and between Quoine Pte Ltd and the user of Quoine (User or Member)…”. On the other hand, the Risk Disclosure Statement purports to be a summary of risks with no indication that it is intended to have legal effect. I do not consider that even the most assiduous reader of the website would have any reason to believe when reading the Risk Disclosure Statement that hidden within it there would be amendments to the Terms and Conditions. He or she would not be reading them with this in mind. The aberrant value clause forms part of a larger paragraph dealing with lost opportunities due to system failure and the possibility that the system might wrongly produce a wrong price. Although it purports to give Quoine a discretion to reverse trades in certain circumstances, this is to my mind far too tangential a passage to constitute any form of notice that changes were to be made to the Agreement.

For all these reasons I have concluded that the uploading of the Risk Disclosure Statement onto the website on 22 March 2017 could not serve to amend the Agreement.

(b) If it did, is the aberrant value clause sufficiently clear to be enforceable?

If I were to be wrong on this, I would have concluded that the aberrant value clause was sufficiently clear to be enforceable. Evidence was adduced of not dissimilar express clauses in agreements for use of other platforms and, although the word “aberrant” begs the question “how aberrant?”, this is a matter of judgment in each case which courts and tribunals are used to exercising.

(c) Do the provisions of clause (b) affect matters?

Quoine refers to this as the “Incorporation Clause” and contends that in combination with clause (h), this indicated that the parties had contemplated and agreed that the Agreement could be supplemented by other terms and conditions which may be posted on the website; that such terms and conditions could be amended from time to time; and that when posted on the website would be incorporated as part of the terms and conditions governing use of the Platform.

I cannot accept this. It is reading far too much into clause (b) which, read as a whole, is merely requiring Members to comply with existing use policies and not to misuse the Platform. It is nothing to do with supplementing the Agreement with other terms and conditions. If clause (h) by itself cannot assist Quoine, clause (b) cannot make things any better.

Quoine was accordingly not entitled to reverse the trades on the basis of an express term of the Agreement when read in conjunction with the Risk Disclosure Statement.

It is therefore not necessary to decide on the assertions made by B2C2 in (c) and (d) at [161] above. Both arguments, that clause (h) was contrary to section 3(2)(b)(i) of the UCTA and that under common law where a condition is particularly onerous and unusual, the party seeking to enforce it must show that the condition was brought to the notice of the other party, were directed to Quoine’s assertion that no notice of any form, whether direct or indirect, was necessary. I have not accepted that this was the case on the true interpretation of clause (h).

Defence 3. Is Quoine entitled to reverse the trades on the basis that the contracts between B2C2 and the Counterparties were void under the doctrine of unilateral mistake at common law?

This contention is raised on the basis that, contrary to the previous argument, there was a breach of contract in reversing the seven trades. In these circumstances, Quoine asserts that the doctrine of unilateral mistake, either at common law or in equity, renders the Trading contracts void or voidable. If this was the case, then Quoine would not be in breach of the Platform contract when it reversed the seven trades. I have found this issue to be the most troubling and difficult in this case.

The Law on Unilateral Mistake

The law in relation to unilateral mistake has been developed over the years in common law jurisdictions and was comprehensively considered in Singapore by the Court of Appeal in Chwee Kin Keong and others v Digilandmall.com Pte Ltd [2005] 1 SLR(R) 502 (“Chwee”).

In Chwee, the defendant sold products over the internet using two different websites. A particular Hewlett Packard laser printer was advertised for sale on both websites at a price of $3,854. One afternoon, an employee of the defendant mistakenly altered the price of the printers to $66. The error was not identified until the following morning when a potential customer checked with the defendant whether the price was correct. The error was then corrected.

During the period while the mistaken price was left uncorrected, 784 people made a total of 1,008 purchases for 4,086 printers. On discovering the mistake, the defendant informed all the purchasers that it would not honour the orders because of the mistake. The plaintiffs who between them had ordered hundreds of printers brought an action to enforce the sales and were met with a defence of unilateral mistake. The defendant succeeded at trial before V K Rajah JC (as he then was). The plaintiffs appealed to the Court of Appeal.

The judgment of the Court of Appeal was delivered by Chao Hick Tin JA. The following passages should be noted:

…

Accordingly, in order to succeed in rendering a contract void under the common law doctrine of unilateral mistake, the Defendant must show, first, that there was a sufficiently important or fundamental mistake as to a term of the contract, in the sense that the offeror did not intend the terms of the offer to be that which on its face was offered and, secondly, that the Plaintiff who is seeking to enforce that contract must have actual knowledge of the mistake. Actual knowledge is to be determined by the court as a question of fact and it is sufficient if the court concludes that the person in question was probably aware of the mistake or that he or she was shutting her mind to the obvious. The principle or purpose underlying the doctrine is set out in Chwee at [31]:

The distinction between actual knowledge and constructive knowledge, which acts not to render the contract void at law but to make it voidable in equity is, as the Court of Appeal accepted, a narrow one. The distinction lies in the difference between actual knowledge, determined as set out above, and constructive knowledge in the sense that although the person in question did not have actual knowledge, he or she ought to have known about the other’s mistake.

Chao Hick Tin JA considered the equitable jurisdiction in the following passages in Chwee:

There thus must not only be constructive knowledge but, in addition, there must be an element of impropriety. Further, carelessness on the part of the mistaken party does not of itself disentitle it to relief but is a factor to be taken into account. Regard may be had to any loss or damage to innocent third parties in weighing the equitable balance.

Although Chwee involved the use of computers as the means by which the plaintiffs placed their orders at the erroneous price, it was otherwise no different from a classic case such as Hartog v Colin & Shields [1939] 3 All ER 566. In Chwee, the first plaintiff was well aware that the offer was too good to be true and alerted his friends who also bought the printers. The mistake was a mistake by a human in getting the price wrong and the knowledge was actual knowledge of the other party to the contract at the time the contract was made. The price was a fundamental term of the contract.

The only difference between the facts in Chwee and in Hartog was that in the latter there was only one contract whereas in the former, because of on-line buying, there were a number of contracts before the mistake was spotted.

On the facts in Chwee, it was a plain case. Here, computers are involved in a different way. There was no human intervention at the time the seven trades were effected. On the facts, no human was aware of the trades until after they had been happened.

The doctrine of unilateral mistake is well developed in circumstances where the error is a human error and the nature of the error and the knowledge or lack of it is directly ascertainable from the humans involved. Where computers are concerned, the law is less well developed. In his judgment at first instance in Chwee Kin Keong and others v Digilandmall.com Pte Ltd [2004] 2 SLR(R) 594, V K Rajah JC made the following observation at [102]:

Quoine relied on this passage in support of a submission that in a case such as the present, where algorithmic trading is involved, the court should adopt a pragmatic and judicious stance such that, although there may be a need for certainty that an executed trade will stand, there is a greater need to protect the market from the potentially harmful consequences of a clearly erroneous trade. Quoine went on to submit that, in consequence the court “should intervene where the trade is clearly erroneous and it would be unjust to allow the trade to stand having regard to the circumstances in which it was executed”.

I do not read the observations of Rajah JC as justifying a fundamental change in the law of mistake which as it currently exists requires a mistake in a fundamental term coupled with knowledge or constructive knowledge of that mistake. Quoine’s formulation accepts the need for a clearly mistaken trade but does not focus on the requisite knowledge. The law is clear, regardless of whether relief is sought in law or in equity, it will not be granted unless it is shown that the non-mistaken party had the requisite knowledge of the mistake. The potential magnitude of the consequences of the mistake cannot serve to change the law.

That said, it is plain that applying the law to a case where algorithmic trading is involved does raise new questions. What mistakes have been made and to what extent are they fundamental? How does one assess knowledge or intention when the whole operation is carried out by computers acting as programmed? Whose knowledge is relevant? At what date is this knowledge to be assessed?

The researches of counsel and by the court have failed to identify any authority dealing directly with these questions.

Quoine submits that the Court should consider what the parties are likely to have known and intended if, hypothetically, they had met on the “floor of the exchange” for the purpose of reaching an agreement on the trades on 19 April 2017. On this basis, as I understand it, it is contended that mistakes can be identified by comparing theoretically what would have happened in face-to-face negotiations with what actually happened at the computer interface.

Quoine further submits that the law should treat the algorithms or computers used to enter contracts as the legal agents of their human principles and relied on two learned articles in support, namely L Scholz “Algorithmic Contracts” (2017) 20 Stanford Technology Law Review 128 and S Chopra & Laurence White “Artificial Agents and the Contracting Problems: A solution via an Agency Analysis” (2009) University of Illinois Journal of Law, Technology & Policy. On the basis of this, Quoine went on to contend that in answering the hypothetical question as to knowledge or intention on the part of the non-mistaken party, the Court should consider what the programmer of the software in question would have known and intended when writing the software at the time that software was written.

B2C2 contends that there must be a mistake as to a term of the contract and that the party placing the order, albeit by a computer, must be the person mistaken. It was not sufficient if there was a mistake as to facts surrounding that term. In this respect therefore B2C2’s position was that the assessment of a relevant mistake was no different in the case of computer transactions than in face-to-face transactions.

Further, relying on Chwee, B2C2 contends that the only relevant knowledge is knowledge at the time of contracting and suggests that it is wrong to assess knowledge or constructive knowledge at the time the deep price mechanism was programmed.

So far as concerns the nature of the mistake, I cannot accept Quoine’s hypothetical meeting contention. The question is not what would have happened if the computer element was absent. The parties have chosen to use computers as the means of entering the Trading contracts. They have entered Platform contracts with Quoine so as to be able to enter Trading contracts and, in the case of margin traders, they have entered the Margin trading contracts to regulate their position vis-à-vis the lender, in this case Quoine. All parties were therefore aware that there was to be no human element in the trades and it is wholly artificial to work on the basis of what might have happened if a human element was involved. Equally whilst B2C2’s position identifies what the nature of the mistake must be, it does not assist in determining how and when to identify it.

In the circumstances of this case, I have concluded that the relevant mistake must be a mistake by the person on whose behalf the computer placed the order in question as to the terms on which the computer was programmed to form a Trading contract in relation to that order. This mistake will have to be in existence at the date of the contract in question but may have been formed at an earlier date. The existence of a relevant mistake will be a question of fact in each case.

Turning to knowledge of the mistake, the law in relation to the way in which ascertainment of knowledge in cases where computers have replaced human actions is to be determined will, no doubt, develop as legal disputes arise as a result of such actions. This will particularly be the case where the computer in question is creating artificial intelligence and could therefore be said to have a mind of its own.

As Lord Briggs observed in a recent UK Supreme Court decision, Warner-Lambert Co Ltd v Generics (UK) Ltd [2018] UKSC 56, at [165]:

So also with computers used for trading purposes. Where the law is in a formative state it is, I think, appropriate for a court (of first instance at any rate) to develop the law only so far as necessitated by the facts of the case before it. With this in mind I do not intend to express any views on the precise legal relationship between computers and those who control or program them. The algorithmic programmes in the present case are deterministic, they do and only do what they have been programmed to do. They have no mind of their own. They operate when called upon to do so in the pre-ordained manner. They do not know why they are doing something or what the external events are that cause them to operate in the way that they do.

They are, in effect, mere machines carrying out actions which in another age would have been carried out by a suitably trained human. They are no different to a robot assembling a car rather than a worker on the factory floor or a kitchen blender relieving a cook of the manual act of mixing ingredients. All of these are machines operating as they have been programmed to operate once activated.

Where it is relevant to determine what the intention or knowledge was underlying the mode of operation of a particular machine, it is logical to have regard to the knowledge or intention of the operator or controller of the machine. In the case of the kitchen blender, this will be the person who put the ingredients in and caused it to work. His or her knowledge or intention will be contemporaneous with the operation of the machine. But in the case of robots or trading software in computers this will not be the case. The knowledge or intention cannot be that of the person who turns it on, it must be that of the person who was responsible for causing it to work in the way it did, in other words, the programmer. Necessarily this will have been done at a date earlier than the date on which the computer or robot carried out the acts in question. To this extent I reject B2C2’s contention that that the only relevant knowledge is knowledge at the time of contracting. I agree with Quoine that regard should be had to the knowledge and intention of the programmer of the program in issue when that program (or the relevant part of it) was written.

Accordingly, in my judgment, in circumstances where it is necessary to assess the state of mind of a person in a case where acts of deterministic computer programs are in issue, regard should be had to the state of mind of the programmer of the software of that program at the time the relevant part of the program was written. In the present case that person is Mr Boonen.

Applying the Law to the Facts: Obvious Mistake at common law.

Stepping back, it is clear that if a number of errors or omissions had not occurred, Quoine would have, or would very likely have, avoided any liability.

(a) Quoine could have included a specific clause in the Agreement entitling them to reverse abnormal trades. Other platform operators had done so. Quoine was considering this and did so in September 2017.

(b) The failure properly to update the login credentials in Quoine’s Quoter Program.

(c) The failure to incorporate an exception message in the Quoter Program to alert Quoine to the fact that it was not working.

(d) The failure to incorporate a circuit breaker in the Platform’s software to prevent trades when the order book was empty.

(e) The failure to incorporate a circuit breaker to prevent orders at an abnormal price from being placed on the order book.

(f) The failure to ensure that, in the case of a force-closure, the forced sales were only within a given price range.

(g) The failure to ensure that, in the case of a force-closure, only assets which were actually held by a counterparty in its account at the time were the subject of a market order.

Mr Lozada very fairly conceded that many of these steps could have been taken and were not. It is with this background that I turn to consider what are said to be the mistakes which it is alleged that Mr Boonen had either actual or constructive knowledge of when he wrote the relevant part of the program. The substantive plea is in paragraph 16 of the Defence:

Particulars are then given. Particulars (b) and (c) read as follows:

These particulars were refined in paragraph 38 of the Defendant’s Closing Submissions:

As to the first of these, Quoine submits that if the Quoter Program had not malfunctioned the Platform would not have detected the Counterparties to be in breach of their margin requirements and the force-closure orders would never have been placed. On this basis it is said that if B2C2 and the Counterparties had hypothetically met on the exchange floor for the purpose of reaching agreement, they would not have agreed to enter into contracts at the deep prices. As indicated above, I do not accept that this is the correct approach. The question that must be asked is whether the Counterparties mistakenly believed that the Platform computer was so programmed as to ensure that it was necessary to close out their positions in response to the margin calls.

There is no suggestion that the Platform computer acted otherwise than in accordance with its instructions. It is not suggested that in the circumstances which had occurred, including in particular the absence of sufficient liquidity as a result of the Quoter Program not working properly, the Platform computer was in error in assessing that a margin call was necessary. The question is directed to the Counterparties’ beliefs as to what would happen in relation to margin calls in circumstances of illiquidity.

The evidence in relation to this aspect of the case is sparse. There is nothing in the Risk Disclosure Statement which supports the asserted belief. Indeed, it points in the other direction as Risks 6 and 7 demonstrate:

It is therefore necessary to have regard to Mr Tseung’s evidence to see whether, notwithstanding this, he held the asserted belief. Mr Tseung was one of the co-founders of Pulsar, the primary Counterparty. He gives evidence that when he identified that a margin call had been made on Pulsar he was both alarmed and surprised because he considered that because of the fact that Pulsar’s leveraged positions with Quoine were not substantial, there should have been no risk of Pulsar breaching Quoine’s margin requirements. He then discussed the incident with Mr Lozada and concluded that the depletion of the order book and/or the placement of B2C2’s orders had caused the Platform wrongly to evaluate Pulsar’s positions and detect that Pulsar had breached its margin requirements . Two points should be made. First the evidence shows that the margin call was not caused by B2C2’s orders and secondly, there is no evidence that the margin call was wrongly made in the circumstances that did exist due to the illiquidity in the order book.

Mr Tseung went on to give evidence that he considered that the Platform should have been programmed to check whether the available balances under Pulsar’s account were sufficient before placing any orders to sell and that, accordingly the Platform should not have placed the market orders in excess of this. As indicated above, Mr Lozada accepted that this was an error on his part but I do not consider that it is an error which is material to the Trading contracts. It may (but I make no finding of this) have given Pulsar a claim against Quoine under the Margin contract but that is all.

Mr Tseung was not cross-examined on his evidence and it therefore stands unchallenged. The belief which is asserted is that the Platform computer was so programmed as to ensure that it was necessary to close out their positions in response to the margin calls. In one respect it was necessary to close out the positions because the illiquidity had caused the computer to identify the need for a margin call. It had acted as programmed. The point that I think is being made is that Mr Tseung believed that the computer would not get it wrong and would only force-close in circumstances that would have led to a margin call if everything was operating properly. Again, therefore, it is hard to see how this could be an error which is material to the Trading contracts. Mr Tseung does not give evidence that he believed that once a margin call had been made the relevant market orders would not be placed.

Whilst I therefore accept that Mr Tseung genuinely held the belief that there would only be a margin call if everything was working properly and one was then justified, I do not accept that this was a mistaken belief which is in any way fundamental to a term of the Trading contracts.

In any event, on the basis of the findings of fact as to Mr Boonen’s knowledge in [124] above that were no grounds for concluding that he ever turned his mind to the relationship between the margin traders and those who had loaned them money, it must follow that even if the mistaken belief was relevant to the Trading contracts, the requisite actual knowledge cannot be ascribed to Mr Boonen. The Trading contracts thus cannot be avoided on the basis of this alleged belief on the grounds of a unilateral mistake at common law.

The second alleged mistaken belief is that the Counterparties entered the contracts under the mistaken belief that they were buying at prices which accurately represented or did not deviate significantly from the true market price as of 19 April 2017. This is therefore a plea that Mr Tseung believed that, whatever happened on the Platform, if there was to be a margin call, this call would only be executed at or near the prevailing “world-wide price” (ie, an average of the prices available on all BTC/ETH exchanges and not the actual price on Quone’s Platform at the time of the trades). Mr Tseung was no doubt familiar with the concept of market making and would therefore have known that the market making system on the Platform would have involved a process akin to that which should have been the case if Quoine’s Quoter Program had been working properly. On that basis he formed the belief that no trade would be carried out otherwise than at what I have called the world-wide price.

Again there is no support for this belief in the Agreement or in the Risk Disclosure Statement as the following passages demonstrate:

The expression “best possible available market price” in the Agreement must be a reference to the best price available on the Platform rather than on other platforms as is made clear by the later statement under the “The Platform Data & Content Copyright” section of the Agreement which draws a distinction between the actual price available on the Platform and “live market value data” provided by Quoine for which Quoine accept no responsibility.

Mr Tseung, however, gives evidence that because the trades were carried out at the highly abnormal price, they were clearly invalid and that Pulsar did not consider them to be valid. He went on to state “[i]t was never contemplated by Pulsar that any trades may be transacted on the Platform at prices which deviated so substantially from the actual market prices”. Although he gave no reasons for holding this belief, he was not cross-examined on this statement and I therefore accept that this was a genuinely held belief. Indeed, having heard all the evidence, I can well understand that a trader on the Platform who was not a market maker might well have formed such a belief even having read the Agreement and the Risk Disclosure Statement. It is a belief which is founded on the premise that the Platform would either always operate as intended or, alternatively, if for some reason it did not, there would be adequate error identification and protection systems in place to prevent trading continuing in such circumstances.

I therefore accept that the Counterparties held this mistaken belief and that it is a belief which is fundamental to the Trading contracts.

It is next necessary to determine whether Mr Boonen had actual knowledge of the mistaken belief at the time he inserted the deep prices and I thus return to the findings of fact in [118]–[125] above. Can it be said that Mr Boonen knew that “it was never contemplated by [any trader] that any trades would be transacted on the Platform at prices which deviated so substantially from the actual market prices”? This amounts to a belief held by Mr Boonen at that date that the price was so abnormal that no trader would trade at that price otherwise by way of a mistake.

This is the aspect of this case that I have found to be most troubling but I have concluded on the basis of those findings of fact that Mr Boonen did not insert the deep prices with that belief. He foresaw that a number of factors might arise which would cause the deep prices to be inserted and the overriding reasons for them being inserted was to protect B2C2 in the event of the unexpected happening. He did not exclude the possibility of trades at those prices being executed. Whilst he was aware that one possible cause was that it could be the result, wholly or in part, of some error or omission on the part of someone, including himself, he did not turn his mind in any detail to the circumstances that might lead to such trades being executed. He knew that the Platform was an automated system and that therefore no opportunity would arise for any particular trade to be reviewed by the parties in advance. In the circumstances of this case, in order for him to have actual knowledge that other traders believed that in no circumstances would a trade be transacted on the Platform at prices which deviated so substantially from the actual market prices, I consider that it would be necessary for it to be demonstrated that he held that belief himself, which he did not: see [123] above.

Accordingly, whilst I accept that the Counterparties held the second mistaken belief relied upon by Quoine, I do not accept that Mr Boonen had actual knowledge of that belief. The defence of unilateral mistake at common law therefore fails.

Defence 4. Is Quoine entitled to reverse the trades on the basis that those contracts were voidable under the doctrine of unilateral mistake in equity?

In order to succeed in equity, Quoine must prove first, that notwithstanding the fact Mr Boonen did not have actual knowledge, he had constructive knowledge, in the sense that he ought to have known about the Counterparties’ mistaken belief. Secondly, the requisite impropriety must be shown to exist.

For constructive knowledge to exist on the facts of this case it must be demonstrated that Mr Boonen was acting irrationally in forming the views that he did and that any reasonable person in his position would have known that no other trader would have contemplated trades being executed at those prices. However, that person has to be placed in the exact position of Mr Boonen, knowing the reasons why he acted as he did and his motivation for inserting the deep prices. It would then be necessary to conclude that such a person could not have had the mindset of Mr Boonen. Having read Mr Boonen’s written evidence and heard him give evidence in a lengthy cross-examination, I am satisfied that his thought processes were rational and that he did not turn a blind eye to that which would have been obvious to everyone else in his position. On the facts of this case therefore, once actual knowledge has been rejected there is no place for a finding of constructive knowledge.

In these circumstances, it not necessary to consider the question of impropriety. It was however argued and I shall therefore give brief reasons for concluding that the required impropriety was not present. The impropriety in question must lie in some form of unconscionable conduct or sharp practice; mere carelessness on the part of the mistaken party does not disentitle it to relief in equity but due consideration must be given to the position of innocent third parties.

On the facts of this case there are no innocent third parties. Although not a party to the Trading contract, Quoine was the only other party involved and it cannot be thought of as an innocent third party. Indeed, had Quoine acted as it accepted it could and should have acted there would have been no margin call.

On the back of its assertions that Mr Boonen deliberately included the deep prices with the intention of triggering a margin call and/or with an underlying intention of manipulating the market, Quoine submitted at various times that B2C2’s conduct was predatory, opportunistic, unethical, improper, unconscionable and tantamount to sharp practice. I have however rejected those assertions and have concluded that although opportunistic, this was not sinister (see [125] above) but was the result of a business decision to ensure that an unlikely event resulted in a profit not a loss. There is nothing unconscionable in this when all relevant factors are taken into account – particularly the number and extent of the errors and omissions that played a part in the trades being executed: see [212] above. Had I concluded that there was constructive knowledge I would nonetheless have held that there was no impropriety. The defence of unilateral mistake in equity therefore fails.

Defence 5. Is Quoine entitled to reverse the trades on the basis that those contracts were void under the doctrine of mutual mistake at common law?

The law in relation to mutual mistake was not in dispute. B2C2 referred me to the Court of Appeal decision in Olivine Capital Pte Ltd v Chia Chin Yan [2014] 2 SLR 1371 and Quoine to the recent decision of the English High Court in Triple Seven MSN 27251 Ltd v Azman Air Services Ltd [2018] EWHC 1348. It is common ground that four requirements, in particular, have to be met. First, at the time of the contract, there must be an assumption shared by both parties as to a particular state of affairs. Secondly, that assumption must be fundamental to the performance of the contract. Third, the assumption must have been wrong and fourth, performance of the contract must either be impossible or be radically different to what was contemplated. There are other aspects to the defence but in the circumstances of this case I need not consider the law further.

As with the defence of unilateral mistake, Quoine contends that the correct approach is to ask the hypothetical “meeting on the floor of the exchange” question: see [200]–[204] above. For the same reasons I reject that approach in relation to mutual mistake as well. Quoine went on to submit that at the time the trades were executed both parties wrongly assumed that (i) the Platform was working correctly and (ii) that therefore the trades were being transacted at normal market conditions, (ie, at prices which accurately reflected or did not deviate significantly from the world-wide price).

Even if Quoine held the mistaken belief as to price, on the facts as they have been found, it is plain that Mr Boonen did not. Once therefore the hypothetical meeting on the floor of the exchange is rejected, it must follow that, if the defence of unilateral mistake fails, the defence of mutual mistake must fail also.

Defence 6. Is Quoine entitled to reverse the trades on the basis of unjust enrichment?

Paragraph 18 of the Defence pleads:

The doctrine of unjust enrichment provides a cause of action for one party against another party who has received a benefit from the first party in circumstances which make it unjust for the second party to retain the benefit. There are three elements to the claim: Singapore Swimming Club v Koh Sin Chong Freddie [2016] 3 SLR 845 at [90]. First there must have been a benefit that was received by or which accrued to the second party. Second the benefit or enrichment must be at the expense of the first party. Thirdly, that enrichment must be unjust.

Had Quoine not reversed the seven trades, B2C2 would have received a substantial benefit in terms of the BTC that had been transferred into its account. This benefit would have been at the expense of the Counterparties who would have had to make up the shortfall between the BTC actually in their accounts and those that had been transferred pursuant to their margin trading contracts with Quoine. As a matter of legal formality therefore, on the facts of this case, the cause of action does not accrue until after B2C2 succeeds in this litigation and has the benefit that it should have had if Quoine had not wrongly reversed the trades. Further, technically, it is a cause of action vested in the Counterparties who are not parties to this action rather than in Quoine.

This, however, is a matter of form rather than substance. If as a result of this action B2C2 receive money or money’s worth whether as a result of damages for breach of contract or by way of specific performance (and the magnitude of that sum is not limited by the limitation of liability clause) then this would technically be at the expense of the Counterparties.

In this case, however, it has been pleaded as a defence and B2C2 has not sought to strike it out. It joined issue with it in its Reply and the case was argued on the basis that it was properly raised as a defence. It was only in B2C2’s written closing submissions that the legal position was first raised. I therefore propose to treat it as a defence so that the substance of the matter may be decided.

The first two requirements of unjust enrichment are satisfied in this case. The sole issue is whether it would be unjust in the circumstances of this case for B2C2 to retain the proceeds of the Trades. Both parties drew my attention to the decision of the Court of Appeal in Wee Chiaw Sek Anna v Ng Li-Ann Genevieve (sole executrix of the estate of Ng Hock Seng, deceased) [2013] 3 SLR 801 (“Anna Wee”) together with extracts from the well-known works of reference, Goff & Jones: The Law of Unjust Enrichment (Sweet & Maxwell, 9th ed, 2016) (“Goff & Jones”) and Andrew Burrows, The Law of Restitution (Oxford University Press, 3rd ed, 2011) (“Burrows).

It is sufficient to cite the following passage from Anna Wee at [132]–[134] where the Court of Appeal quoted from both Burrows and an earlier edition of Goff & Jones:

In this case Quoine maintains that the unjust factors are mistake and lack of consent due to ignorance. It contends that in the circumstances in which the trades were executed, they were executed as a result of the automated actions taken by Quoine’s systems which would not have occurred if the Quoter Program was operating properly. Essentially the mistakes relied upon are the same as were raised in respect of unilateral mistake. First, the mistake as to the necessity of making the margin call, which I have rejected as being a mistaken belief and therefore need not be considered further. Secondly, the mistake as to the “world-wide price” which I have held to be an honestly held but mistaken belief.

Thirdly, it is asserted that the benefit had been conferred upon B2C2 without the consent of the Counterparties because the Counterparties believed, wrongly, that the margin call would only relate to BTC which were actually held in their accounts. I have held that this is not a mistake which is relevant to the Trading contracts.

B2C2, however, contends that regardless what the mistakes or lack of consent were, if the defences of mistake fail then the case on unjust enrichment must also necessarily fail. The contract is valid and therefore the fact that one party has made a bad bargain does not prevent the other party who has made a good bargain from keeping the fruits of the contract. My attention was drawn to two passages, one from Burrows and the other from Goff & Jones:

Burrows goes on to consider exceptions to the general rule as being in cases where the prima facie injustice constituted by there being an unjust factor is not outweighed by the Defendant’s contractual entitlement to the enrichment.

The question was considered in a Privy Council decision in Fairfield Sentry Ltd (in Liquidation) v Migani [2014] UKPC 9 where at [18], Lord Sumption said this:

I therefore accept that it is only in exceptional circumstances that a claim for unjust enrichment can succeed where a contract has been held valid. I do not consider that this is such a case. B2C2 was enriched because Quoine failed to take any of the steps necessary to protect itself or the margin traders identified in [212] above and the Counterparties did not take any sufficient steps to ensure that their beliefs were correct. This is not a case of B2C2 getting an unjustified windfall, it is the inevitable result of the way the parties have chosen to trade with each other. The defence of unjust enrichment thus fails.

The action therefore succeeds both in breach of contract and in breach of trust.

Relief if Quoine is liable

The primary relief sought by B2C2 is specific performance coupled with additional damages. It however accepts that specific performance is a discretionary remedy and will only be ordered where it is just and equitable to do so. The two primary considerations are whether damages would be an adequate remedy and whether the person against whom the relief is being sought would suffer substantial hardship: see Lee Chee Wei v Tan Hor Peow Victor and others [2007] 3 SLR(R) 537 at [53], and Lim Beng Cheng v Lim Ngee Sing [2016] 1 SLR 524 at [97].

B2C2 asserts that this is a case where damages will not be an adequate remedy because cryptocurrencies in general and BTC in particular are highly volatile so that the inherent value is difficult to ascertain and will thus make damages difficult to assess. I do not accept this. Courts are accustomed to assess damages in relation to volatile assets and this case will be no different.

To my mind the compelling factor against ordering specific performance is that it would serve to require Quoine to transfer BTC to B2C2 which it would have to obtain at today’s price which is substantially higher than the price in April 2017 when the trades were executed. B2C2 are market makers, not investors. When the BTC were originally credited to its account the B2C2 software immediately began to hedge the proceeds by selling BTC. Before the trades were reversed, B2C2’s systems had sold slightly under one third of proceeds on nine different exchanges. In these circumstances granting specific performance would cause substantial hardship to Quoine which any potential difficulty in assessing damages does not outweigh.

I therefore decline to order specific performance. B2C2 will be left to a claim in damages for both breach of contract and breach of trust which, if not agreed, will be assessed at a subsequent hearing.

Conclusion

The action succeeds both in respect of breach of contract and breach of trust but B2C2 are not entitled to an order for specific performance. Its remedy lies only in damages. The parties should liaise in drafting an appropriate order which will be settled at a subsequent hearing.

I would like to express my thanks to all the representatives of both parties for the assistance I have received in this case which raised complex questions of both law and fact.

Annex 1: The Agreement

Annex 2: the Risk Disclosure Statement

Annex 3: Copy of the print out forms