Shurbanova v Forex Capital Markets Ltd

The Defendant, FXCM Ltd (“FX"), is an online foreign exchange and commodities broker. It offers a number of trading platforms for retail and professional users.

Some of its products were offered through its Dealing Desk (“DD”) which is available to retail customers only as opposed to professional traders. In August 2012, the Claimant, Mrs Daniela Shurbanova, a retired schoolteacher from Sofia, Bulgaria opened an account with FX.

On 8 November 2013 at 8:30am New York time which was 1:30 pm Bulgaria time, Mrs Shurbanova placed the following trades electronically on FX’s DD:

Very shortly thereafter, Mrs Shurbanova closed out these trades by giving instructions to buy the gold and sell the dollars. The closeout of the Trades was completed within the same minute in which they had been opened. Over that short space of time she made US$384,210 on the Gold Trades and US$79,200 on the Dollar Trades.

However, later the same day, FX revoked the Trades on the basis that they had been made in breach of its Terms of Business (“TOB”). FX said that the Trades had been opened at a price which was a “manifest error” and/or the Trades were “abusive”, within the meaning of Clauses 26 and/or 27 of the TOB. Moreover it was said that Mrs Shurbanova was really acting as a front for her husband Alexander Shurbanov (being the masculine version of the family name) who had been banned by FX on numerous occasions from undertaking particular types of trade with FX and/or for their son Martin, who sometimes acted on the behalf of his father and who had also been subject to restrictions on his trading activities.

Mrs Shurbanova denies all of these allegations and contends that FX was in breach of contract in revoking the trades. She therefore claims damages, being the profits she would have earned (and been paid) absent the revocation i.e. the amounts referred to in paragraph 4 above.

For Mrs Shurbanova I heard from her and her son Martin. I also heard from Simon Carse, an independent Compliance Consultant to the Financial Services Industry, who gave expert evidence.

For FX I heard from Paresh Patel, a Managing Director of Forex Capital Management LLC, an associated company of FX, and head of the dealing desks, along with Juan Cafe, FX’s chief Operating Officer. I also heard from Andrew Kasapis, an expert in capital markets including in particular, foreign exchange options and contracts for differences.

In general terms, I thought that both Mr Patel and Mr Cafe gave their evidence clearly and reliably and were doing their best to assist the Court. Unfortunately, the same cannot be said of Mrs Shurbanova whose evidence I considered to be inconsistent and implausible in a number of respects as set out below. Overall she was an unreliable witness. To a lesser extent, so was her son.

As for the experts, overall, Mr Kasapis gave clearer and more persuasive evidence and on the key points, his expertise was plainly superior to that of Mr Carse.

There is a significant amount of contemporaneous documents, mainly emails and computer-generated trading data, which has been of considerable assistance to me.

The form of trading engaged in by Mrs Shurbanova is known as “news trading”. Usually, this means placing trades on the basis of the predicted outcome of a particular news event whose content is capable of influencing the currency and commodities markets. One such news event is the US Non-Farm Payroll Data announcement (“NFPD”) which is made monthly. It contains a set of figures which are an important indicator of the increase (or decrease) in non-agricultural jobs in the USA. It is common ground that positive NFPD results will usually lead to a rise in the value of the US dollar and with it, a decline in the value of gold. The reason for the latter is that there is in general an inverse relation between the values of both. The changes in these values following an NFPD announcement may be short-term.

Those who trade on news events therefore wait for a particular event to take place and they may not trade very much at other times of the year.

Because the NFPD results can influence the markets, they are released at the same time all over the world subject only to differing international timelines.

FX operates its DD as a counterparty to the retail trader who is buying or selling. In other words it acts as principal. To use a gambling analogy, FX is therefore the “House”. If the trader wins, FX loses and vice versa. This DD retail platform will quote prices for the various products on offer. When the trader places an order it will be executed by FX’s electronic systems very shortly (often milliseconds) thereafter. At times of market volatility, it is possible that by the time the system executes the trade, the price will have moved slightly in which case the trade is executed at that price. This movement from the original quoted price is known as “slippage”. The trader is able to set a tolerance for slippage i.e. how much slippage the trader is willing to accept. Then, if the slippage is outside that range the trade will simply cancel.

Furthermore, not all trades will be executed either to their full extent or at all. This is because the amount of trades may be limited by the “liquidity providers” who are banks and other institutions who operate behind FX. If a trade is placed on an “IOC" basis it means that the order is “immediate or cancel”. To the extent that part of the order cannot be fulfilled, it is simply cancelled and the balance is executed. It is possible for placed orders to fail altogether.

The retail DD is designed for individual non-professional traders, as a result of which it has two particular features. First, it allows for small trades which other platforms might reject. A large number of small trades, even if placed by the same trader, may be accepted on DD but not on a professional platform. That is important because the unit price for a single very large order tends to be less favourable than that for a small order or a collection of small orders. Second, the quoted price reacts more slowly (in relative terms) to particular events including news events, than other non-retail platforms. The slowing of the response time in the quoted price is deliberate and is known in the business as “throttling”. The purpose of this is to allow individual traders more time to think about the trades they wish to make.

On the DD platform, contracts for differences (“CFDs”) can be made. This is where the trader places a “sell” and then a matching “buy” order or vice versa. No money is required upfront for example on a buy order. Instead the result after closing-out is calculated and only then will the trader have to pay the difference to FX if it loses. If it wins, FX pays the trader. The traders all have accounts with FX containing a minimum amount of money and into which or from which the relevant sums are paid.

FX has another retail platform known as the “No Dealing Desk” (“NDD”), although it can also be used by institutional/professional traders. Here, FX does not act as principal but effectively passes the trade over to the liquidity provider. FX will earn a commission from such a trade. Accordingly, such trades are risk-free from FX’s point of view. CFDs are not available on the NDD. The quoted prices on the NDD respond more quickly, relatively speaking, to events. The prices on FX’s professional-only platforms respond even more quickly.

FX’s retail customers can place trades either manually (i.e. they input the trade’s criteria and then click to execute the trade) or they can make use of FX’s Application Program Interface (“API”). Here they can connect their own trading software application to the API and trade electronically and according to pre-set criteria. So the criteria may encompass a particular trading strategy and the customer can “sit back” as it were and wait for the results to see whether a loss or profit has been made.

With such applications and in respect of “slow” platforms like DD, there is the potential for abuse. One form of abuse is called price latency abuse. Here, the trade is triggered by the outcome of a particular news event; it is not set up in advance of that event. If the results of this news event can be sent to and “machine read” by the application fast enough it can make trades on the DD platform in the light of those results before the quoted prices on the DD have had time to respond to them. The trades will close out at a point very shortly thereafter by which time the quoted price on the DD will have adjusted to the results. Such trades are regarded as abusive because they are based not upon some prior prediction (however sophisticated) of what the results will be and what the market will do, an intelligent gamble as it were, but rather on the known outcome of the news events. The profit is made because of the time lag in the quoted price on the DD. The trader thus makes the trade with actual “knowledge” (whether by the individual or by the application being used) of the news event. Sometimes this form of trading is known as “gaming".

It is common ground that FX’s Terms of Business governed the contractual relationship between Mrs. Shurbanova and FX with regard to her trading activities. The relevant provisions are as follows:

By Clause 27.1 (the clause as a whole being headed “Gaming and/or Abusive Strategies”),

“Internet, connectivity delays and errors sometimes create a situation where the price displayed on the Trading Facility does not accurately reflect the market rates. The concept of gaming and/or abusing the system cannot exist in an OTC market where the customer is buying or selling directly from the Principal. The Company does not permit the deliberate practice of gaming and/or use of abusive trading practices on the Trading Facility. Transactions that rely on price latency opportunities may be revoked, without prior notice. The Company reserves the right to make the necessary corrections or adjustments on the Account involved without prior notice. Accounts that rely on gaming and/or abusive strategies may at the Company’s sole discretion be subject to intervention by the Company and the Company’s approval of any Orders. Any dispute arising from such quoting or execution errors will be resolved by the Company in its sole and absolute discretion.”

I shall refer to Mrs Shurbanova’s husband Alexander Shurbanov as “Mr Shurbanov” and to their son Martin as “Martin Shurbanov”. Mr and Mrs Shurbanov lived together in Sofia.

By November 2013, when the instant trades had taken place Mr Shurbanov had had a very chequered history of dealings with FX. There was no real dispute at trial as to the nature and extent of this - not least because, somewhat remarkably, Mr Shurbanov himself did not attend to give evidence in support of his wife where he could have corroborated her account on a number of matters. Indeed he did not even attend the trial. I do not accept (as suggested by Mr Stone on behalf of Mrs. Shurbanova) that he would have nothing useful to say and although in evidence Mrs Shurbanova said that he had some “health issues” preventing his attendance here, nothing more specific was said.

Mr Shurbanov had been a Professor of English and American Studies at the University of Sofia until about 2009. Martin Shurbanov was and is a software engineer although he also has a property business. He lives in Somerville, Massachusetts, USA.

Mr Shurbanov had many accounts with FX between 2009 and 2014 in his own name, over 50 (in fact 59 according to the list at 3D/1656-1657) of which at least 24 were retail (29 according to the list). That is an unusually large number of accounts, according to Mr Cafe, even taking into account 5 year’s trading and the differing nature of some of the accounts, for example, the 11 “merchant” accounts which were not themselves tradable and 5 PAMM sub-accounts which involved trading for other clients. While Mr Cafe accepted that a trader might open more than one account in order to trade directly in say euros and dollars it was unusual (as here) to be opening accounts in 6 or 7 different currencies. And while accounts might over the years open and close, Mr Cafe said “not to this degree”. He said that a much lower number of separate accounts would raise a “red flag” about trading. He did not accept that the true position for Mr Shurbanova’s accounts was more nuanced (and by inference not suspicious or less suspicious). I accept Mr Cafe’s evidence. The only sensible explanation in my view is that because Mr Shurbanov was restricted in his trading activities on one or more accounts, he opened at least some of the other accounts in order to try and get round it. This is a paradigm example of where direct evidence from him would have been useful. In addition, on some occasions, as Martin Shurbanova admitted in evidence, the latter would call up FX on the telephone to make trades on his father’s behalf.

By 2011 FX suspected that Mr Shurbanov was abusively trading on price latency and was in possession of a very fast news feed i.e. he or the software he was using was picking up the result of news events faster than the platform on which he was seeking to trade. Also by now, the liquidity providers were complaining to FX that he was trading in this way.

FX took various steps to counter this. It did not exclude Mr Shurbanov entirely from trading but prevented him from trading CFDs on DD where there were the real opportunities to engage in price latency abuse. If Mr Shurbanov traded on FX’s professional platforms there was no or not sufficient price latency to enable him to make any real profits. Those other platforms would also not fill multiple small trades much more regularly than on DD.

FX also placed Mr Shurbanov on “Dealer Intervention” which was one of the restrictions permitted under Clause 24.2 of the TOB. This meant that FX had to approve each individual trade he wanted to make before it would be executed.

There are numerous contemporaneous internal emails at FX between, among others, Mr Patel and his colleagues, setting out their belief that Mr Shurbanov was price latency trading and describing the steps being taken to prevent him from doing it. But by opening new accounts Mr Shurbanov was also sometimes able to evade the ban on CFDs and it would take FX some time to catch up with him.

On 3 August 2012, Jason Shaheen, a senior dealer and head of CFD Dealing at FX, sent an email to dealers containing a list of known Shurbanov account details and said:

“If you come across any of these accounts with the following names make sure they are on DI. They’re not allowed to trade CFDs anyway. If somehow they do and they run us over we are taking the trades back. They use feeds that are faster than ours and even the banks.”

Martin Shurbanov explained that he used a particular computer and server. Others (including Mr Shurbanov and later his mother) could also access it by using remote- access software such as TeamViewer and thereby place trades remotely but which would show as having the same IP (Internet Protocol) address as Martin Shurbanov’s. In so doing, they would also use his software and the application which he had developed.

It is common ground, and Martin Shurbanov admitted, that he used what was then a freely available piece of software called Forex News Gun (“FNG”). This is an electronic newsfeed that receives the results of news events very fast and can “read” the relevant data, including, by way of example, the number of new jobs created as announced in the NFPD. Combined with Martin Shurbanov’s application, this meant that very positive results from NFPD would cause the appropriate buy and sell orders to be made automatically utilising FX’s API. The precise nature and extent of the orders would vary according to the extent of the positive results and this would all be pre-set.

This trading activity was in effect risk-free because the relevant trades would only be made if there were particular results announced. They were based not on intelligent predictions of those results but on knowledge of what those results were. The only conceivable risk would be that, contrary to the usual experience of how markets behave, for some reason the US dollar went down instead of up upon receipt of good economic news and vice-versa for gold. This was a not real possibility here. Even Martin Shurbanova accepted in evidence that the probability of making the profit is 90% or 95%. See 1/163-165.

It is true that at paragraph 83 of his report Mr Carse said that “… If an order is executed at a price which, once the markets are subsequently adjusted to the news that triggered the order, is above the then prevailing price, the trader would incur a loss.” I quite agree if the order was a “buy” order in respect of for example the Gold Trades. But on the assumption there is going to be good economic news, no trader would do this. Instead the trader would place “sell” orders and then close out with “buy” orders shortly after. Mr Carse went on to say that the market response to a particular news event might be short- lived. Again, I agree, but the context for the news trading here was a very narrow “sell and buy” window indeed. Any market stabilisation hours or even minutes later is irrelevant because the trades have been completed and the profits made.

There are two key factual questions here. First, what the nature of Mrs Shurbanova’s actual trading was on 8 November 2013 and secondly, the extent to which she was simply acting as a “front” for her husband and/or Martin Shurbanov. The latter question is not a legal one about principal and agent, rather it is relevant to FX’s belief that she was so acting and that this formed part of the reasoning behind choosing to revoke her trades. It is also relevant in a more general way to the question of abusive trading.

Something needs to be said about the background to the trades. A convenient place to start is the application form which she completed electronically in order to open a trading account with FX. This assumed particular importance because FX has alleged that she made a number of material misrepresentations in this form.

The form was submitted on 27 August 2012. She gave as her address Bigla 19a, Lozenets, Sofia (“Lozenets”). She said that she was employed as a University Teacher, and that she had one year’s trading in commodities, two years’ trading in CFDs and three years’ currency trading, all on a weekly basis. Her total annual income was said to be £50-99,000 with a net worth of £250-999,000 and liquid assets of £250,000-£1m. She said that she had heard about FX through a magazine. She also said that no one else would trade the account and that she understood all the terms of business.

There were a number of material inaccuracies in this form, in my judgment, although not picked up at the time. As to the given address, in fact, she lived (and lives), with her husband at 20 Tsanko Tserkovski, Sofia (“Tserkovski”). This is the address that she gives in her witness statement. As to why she used Lozenets instead she said that they owned this apartment as well and tended to go there in the summer. When it was pointed out to her that her husband’s application for a trading account made in June 2012 gave Tserkovski and not Lozenets, she then said that it may be that in June they stayed in Tserkovski and put the heating on before moving over to Lozenets later. I regarded that as implausible. They may have owned Lozenets but I do not accept that it was accommodation which they lived in as suggested.

As to prior trading experience, the form clearly indicated that she had some. But in paragraph 6 of her first witness statement, she said that her prior trading experience was limited to “demo” trading. This is in effect “dummy” trading for potential traders to gain experience of how the platforms worked but without committing any funds. So there is no actual trade and no risk. Demo trading is plainly not the same as real trading. In evidence she said that she had actually traded before but could not say when. She also said that she had traded using automatic devices and did not consider that this was real trading because there was no risk. But if they were real trades, then, risk or not, they could not sensibly be called demo trading. As to saying on the form that she had two years’ worth of trading in CFDs, in evidence she said at first that she did not know what CFDs were. At another point in cross-examination she said that she did in fact trade weekly. But that is not borne out by the evidence. So there are many inconsistencies here.

As to hearing about FX from a magazine, in paragraph 7 of her first witness statement she said that she had first learned about forex trading from her husband and her two sons. She said that she then later picked up the name of FX from the Forex Peace Army (“FPA”) website which dealt with forex trading and which had in fact made the FNG available to download. She said that the magazine was an online magazine but she could not recall its name. In paragraph 8 of her witness statement she said that she had had “discussions” with Mr Shurbanov and Martin Shurbanov about the trading. In evidence she played this down, saying that she may simply have heard them talking with each other and so learnt about FX - but that she had then discovered FX by herself. I regard that as extremely unlikely as I do the suggestion that as the apparent “occasional” trader in the family, she never actually spoke at any length or in any detail to her husband or Martin Shurbanov about FX prior to opening her account.

As to the income and asset figures given in the form, she said this was a reference to the income and assets of the “family” as a whole even though this is not what the form asked. In evidence she said that she and her husband practically treated their money as belonging to both of them although she also said she had her own bank account. She said that the income figures on the form related to her income only. In fact, Mr Shurbanov’s application form put the net worth as much higher going up to £4,999,999. If that was the correct figure for the family’s net worth, then it is inconsistent with the figure that she gave. Alternatively, if this was meant to represent Mr Shurbanov’s own worth, then he did not appear to treat his and her assets as belonging to both. The same must follow from his statement that the liquid assets were over £1m.

On her actual trading since opening the account, in her witness statement Mrs Shurbanova said that she learned the basics from her sons and husband, but in evidence said she did not speak to her husband at all about this and he did not know about her trades (including those on 8 November) until after her account had been closed. She said that all she told him was that she wanted to open an account with FX and he said that he knew FX and had traded with it and was happy that she should trade there also. In answer to the obvious question as to why Mr Shurbanov, by 2012, would be happy for her to trade with FX since they kept trying to ban him and restrict his activities, she said that he was not in fact recommending it and was not telling her to do so, but he simply mentioned it. All of this is wholly implausible.

Finally, although she said in the form that she was still employed, she had actually retired. In evidence she sought to explain this by saying that she was both employed and yet retired because she had an income from translating and private tutoring work. But even so that would be inconsistent with the statement on the form that she was still employed as a “university teacher”.

Quite apart from the yet further problems with her evidence described below, the above matters in respect the application form demonstrate clearly that Mrs Shurbanova was a most unreliable witness. I have no doubt that she gave information on the form she did to create the impression that she had her own independent income and wealth, was an experienced trader and was still working as a professional teacher. All of which made it likely that FX would allow her to open the account as it did once certain discrepancies over birth dates were sorted out. It is also absurd to suggest that she did not have more extensive conversations with both her husband and Martin Shurbanov about her prospective trading.

Finally, on the question of the application form and TOB, she complained in her second witness statement that FX should have given her a “crystal-clear” warning about its contractual rights, in particular in relation to gaming or abusive trades. But in evidence she accepted that she had read the TOB but did not understand them exactly. Later she said that she did, once Martin Shurbanov have explained had explained them to her.

I do not accept that the unsatisfactory nature of Mrs Shurbanova’s evidence can be explained away by saying that she was confused or had a poor recollection.

It appears from her account history that after the account was opened, she only actually traded with FX once, in July 2013. Here, she made a profit of £8,548.14. After this trade, there was £458,523.14 remaining in the account.

I should refer here to one other point on her credibility. Subsequent to the trades, she submitted replacement bank details to FX. These were of a bank account which in fact was that of her daughter-in-law, Martin’s wife, Irina Ignatova. She said she had done this because Ms Ignatova had wanted to trade and needed some funds to do it. But in fact, no more trading was done. Mrs Shurbanova then said that she had given Ms Ignatova’s details because she “needed it”. But that made no sense since this was not a gift of money but simply details for trading purposes. When that point was put to her she said that she did not know in fact why she had done this. None of this adds up.

I now turn to the trades in question.

In her first witness statement Mrs Shurbanova said she was familiar with FNG and also the application developed by her sons. She said that the particular reason for making the trades on 8 November was the market movements observed on previous NFPD events. She set the parameters for the trades herself and gave the relevant instructions a few days before 8 November. Although her witness statement suggests therefore that she really did everything to set up the trades, in cross-examination she said that Martin had suggested the size of each trade but she suggested how many.

Her evidence in cross-examination about the trades themselves betrayed a real lack of understanding in my view. First, she said that she was trading in dollars which was not correct; it was a product based on a dollar rate as against a basket of 4 currencies and she then said that she did not know what a “basket” was. Even more remarkably, she said that she had decided to embark on these particular trades because there had been a “good movement” on gold, in other words the value of gold would go up, to judge from previous NFPD announcements. She thought that gold would go up “again”. As to the dollar she thought that this would go up as well. When it was pointed out that in fact a gold went down after the 8 November trades (as it was bound to do if the dollar rose) she said that it had been a “good try”. But of course it was more than that - it was not a good attempt at winning, it was an actual win. Nor, of course, would the dollar go up if the price of gold was going up.

I do not think that Mrs Shurbanova had any real idea of what she was doing and the corollary of that here must be not that she was a novice acting alone, but rather that her trading activities were driven by her husband and/or Martin. The setting up of the account in her name was not because she had decided to trade in her own right; it was because it was another perceived way to get round the restrictions placed on Martin and her husband. And at first, it succeeded. She was probably unable to use a false name because it would be picked up on checks carried out by FX and if she had another flat at Lozenets then she could use that address in order to distance herself from her husband’s given address. She needed to establish her own financial ability to trade hence the misleading employment details given together with the financial information which was inconsistent with her husband’s details.

Martin Shurbanov’s evidence does not really assist his mother’s case as to the nature of her trading. Part of his witness statement was devoted to a justification for the kind of trading they did using FNG and his application. In his strong opinion it was justified trading but his opinion as a non-expert as to whether it was abusive or not is irrelevant.

Also significant was the fact that the IP address shown on his mother’s application form was that of Martin Shurbanova’s computer or one of his computers. So she must have made her online application from it. He said he let her use his computer remotely because she was having trouble with the form. But in fact the account which Mr Shurbanova opened in June 2012 also had Martin Shurbanova’s IP address on it. He denied that he had opened this account on behalf of his father but said that as he allowed his father to trade using his computer (connected to it using TeamViewer) he may well have been trading on it at the time and so did the application in the same session. That would have been advantageous because he apparently stored various documents on his server which may have been relevant to trades or account openings. There were other accounts opened with that IP address. See the FX email of 18 November 2013 ie after the revocation. They include two accounts in the name of his wife Ms Ignatova, said on the records to have been from Orel in Russia; he said she traded also. Then there was a Mr Poyet said to be from Mahe in the Seychelles. As to that he suggested that this was because he actually licensed out his application to third parties for a fee. And he may have been demonstrating it to them by using his computer. But that raised the question why that would require an account to be set up in those people’s names but with his IP address. When I asked him about this he said he had given some people his credentials to test the application but he did not know what they did on the server. They could have used it to trade also but he did not allow that. However it might have occurred because he let people have a “trial version” which they could use for 30 or 60 days. All of this evidence was highly implausible. The more prosaic but likely explanation for all of this is that Martin Shurbanova was trading for himself and/or his father using a variety of accounts which he set up in other people’s names.

Equally implausible was his downplaying in evidence of the extent to which he would talk to his father about trading.

As to the 8 November trades themselves, he said he had discussed trading with his mother after she had read the FPA information and learned about the figures, but he did not actually give her any figures. He then said in evidence that he was aware that she had been undertaking both demo and live trading before the 8 November trades. That immediately led Mr Khan, for FX, to take him to what he had said in his witness statement which at paragraph 21 stated that he was aware only of prior demo trades. Even before Martin Shurbanova was taken to that paragraph it was clear to me that he realised what was going to be put to him and he tried to qualify what he had just said. Ultimately he said that when he had just referred to knowing about live trades he meant the 8 November trades. That was not convincing (see Day 1/156-157).

Overall, I thought he was an unsatisfactory witness and provided no support for his mother’s case as to the nature and extent of her own trading.

It is clear enough from the matters referred to above that Mrs Shurbanova was a cipher for her husband and/or Martin. Her evidence to the court was untruthful on this point and was riddled with implausibilities and inconsistencies. That said, I have little doubt that she has been put under considerable pressure to give evidence in the way that she did from her husband, who has remained distant from these proceedings even though he could have given useful evidence. This conclusion is rendered all the more certain when one considers the nature of the trades themselves.

As to the 8 November trades themselves, the more technical aspects are set out below but as far as Martin Shurbanov was concerned, his application triggered them following the release of the news data through FNG. Mrs. Shurbanova said that she manually closed out the trades shortly afterwards (in fact very shortly after). Although it has been questioned whether, in the time taken, she could have closed out manually all 43 trades, I cannot conclude that she was wrong about this. It depends what software she had. It is possible that it enabled her manually to click once, in order to close out all the trades and then the software would deal with each individual trade. But on any view, it is clear from FX’s “global blotter” document recording details of the trades, that the opening trades had been placed between the first and second seconds following 8:30am New York time while the closings all took place by or within the thirty-first second following 8.30am New York time. The NFPD data had been released at precisely 8:30am New York time. It gave positive news in respect of jobs and was indeed better than expected.

An “off-market” trade is one where the price of the trade does not represent the underlying market price. In forex trading of CFDs, there is not one uniquely defined buy or sell price set by an exchange, but there are very good indicators and one of these is Bloomberg. I deal further with this below.

What could be observed very shortly after these trades was that there had been multiple orders with the Trades closed out very quickly and a substantial profit made. By 8:39am local time, Mr Shaheen had decided to put a stop on any withdrawals from Mrs Shurbanova’s account which by now had been credited with the profits. He had compared the Gold Trades’ opening prices with Bloomberg’s indicative prices and had observed a significant discrepancy. See the screenshots at 3B/745-748. The Bloomberg figures were timed at between 8.30.19 and 24 seconds New York time. That was the time at which her trades had both been executed and notified to the Global Blotter. The difference was about US$10 per unit which was very marked and unusual within such a short space of time. Mr Patel described it as significant. In addition these were multiple orders all placed within a second and the number of units in each trade was high. And the total amount committed on the trades (as opposed to profit on the differences) was very large – about US$130m. All of that indicated to Mr Patel that this was an abusive “price latency” trade. He also noted that this account had largely been inactive since being opened and he thought that the name Shurbanova was similar to Shurbanov. The trading itself was similar to that previously done by Mr Shurbanov.

Mr Shaheen wanted to revoke the trades but at that time, he needed partner approval first. In an email sent at 9:12am New York time the same morning, he said that the account had been traded “grossly off market price making 384K and that this should be revoked immediately.” He said that “the client in question” had a history of this sort of behaviour and he was not even sure why this person was still allowed to open new accounts. Obviously, Mr Shaheen had by then considered that Mrs Shurbanova and Mr Shurbanov were connected. Almost immediately, Mr Niv emailed back to say that he should do it and that this person (i.e. Mr Shurbanov) was only allowed to be on a different platform, namely EE (for professional traders) and “even the banks have complained about him and want him off so we have been moving him to fastmatch.”.

At 9.26am New York time, Mr Shaheen sent a further email identifying the trades that needed to be revoked and enclosed an Excel spreadsheet showing that all trades closed in the same minute whereas other clients were opening trades in the same second but at much lower rates, further strengthening the legitimacy of the revocation.

At around 10am New York time there was an email conversation between Mr Shaheen and another dealer, Mr Olukhov. It included a number of derogatory and unacceptable personal remarks about the Shurbanov family. It is not necessary to rehearse them here but the underlying point was that Mrs Shurbanova had been trading “grossly off price” and that she was probably connected to Mr Shurbanov who, it was assumed was now using an account in her name to keep on with his trades.

A little later, Mr Patel sought approval for the additional revocation of the Dollar Trades, and this was duly given. He then emailed Mr Shaheen and Mr Niv to say that the problem was that Mr Shurbanov was opening accounts in other people’s names and “today’s account was Daniela Shurbanova.”

At 4.11pm (11.11am New York time) Mrs Shurbanova wrote an email asking “could you please advise why some trades I did today… were removed?… Please note that reversing trades on retail forex accounts is illegal as per NFA regulations.… I would like an answer within 24 hours.” Mrs Shurbanova said in evidence that Martin had helped to write this email and that is fairly obvious.

FX provided a formal response to Mrs Shurbanova’s complaint on 27 November 2013 which maintained that it was justified in revoking the trades. This elicited a short reply which showed that Mrs. Shurbanova was still not satisfied.

One might have thought that given (from Mrs Shurbanova’s stated point of view) this unsatisfactory response to her complaint, she would not trade anymore with FX. Yet she opened four more accounts on 13, 20 and 29 November and 17 December 2013. Her explanation for this was that she thought they had just been a “misunderstanding” about the 8 November trades (which presumably had now been cleared up by FX) and she wanted to continue to trade. One of the accounts was with FX in New Zealand and she said that this had been opened by mistake. However, accounts take more than a few seconds to open, perhaps as much as 20 or 30 minutes when one considers the amount of information to be provided (see for example the opening form discussed above) so it is entirely unclear what sort of mistake Mrs. Shurbanova had made. Mrs Shurbanova was not aware of any further response from FX to her complaint but she did know that she could not trade CFDs.

Despite all of this she thought that FX had “removed the problem” thereby making her feel relaxed about doing further trades. However this was patently untrue. FX did not resolve the problem and there was no basis for her thinking that it did. Indeed it was FX’s failure to resolve the problem that led to the issue of these proceedings brought against it in her name.

Mr Stone submitted that it was surprising that Mr Shaheen did not give evidence on behalf of FX and it is certainly true that he was involved in the revocation process. On the other hand, he was not the person responsible for authorising the revocation and Mr Patel was senior to him. Moreover, for reasons given below, the lack of any direct evidence from Mr Shaheen makes no difference.

For the reason given below, it is not necessary to go further into the detail of the process by which FX came to revoke the Trades.

Against that factual background, it is necessary to set out in more detail, the issues which had arisen by the time of trial.

In answer to Mrs Shurbanova’s allegation that the revocation of the Trades was in breach of contract, FX contended as follows:

As against all of that Mrs Shurbanova alleged that:

However, in his helpful written closing submissions dated 11 August 2017 (“the Claimant’s Closing”), Mr Stone realistically conceded that if the Court concluded that there was in fact either “manifest error” within the meaning of Clause 26, or “abusive trading” within the meaning of Clause 27, then Mrs Shurbanova’s claim probably could not succeed, however it was put. This is because in that event, the revocation would have been expressly justified so there could be no breach by FX. And insofar as FX was in breach of any of the alleged implied duties (to act in good faith, rationally etc) or the statutory or regulatory duties alleged, that could not have made any difference since even if FX had acted “properly” the result would have been the same, namely there was manifest error and/or abusive trading; see paragraphs 83 and 84 of the Claimant's Closing. Hence it would be the end of the claim. Thus all the other points taken by Mr Stone on behalf of Mrs Shurbanova are now made only if I were to find that there was no manifest error or abusive trading. Otherwise, they become irrelevant.

I deal with these points in turn.

Clause 26.1 defines what a Manifest Error is. Clause 26.2 then describes the process by which FX determines whether there is a Manifest Error and the consequences.

Taken as a whole, I consider that Clause 26.2 means that it is for FX in the first instance to determine whether there has been a manifest error as opposed to having to show, after the event as it were, that there was in fact and objectively, a manifest error.

The fact that it is for FX to determine manifest error gives rise to a potential conflict of interest since it could (and would in this case) be in FX’s financial interest to revoke on this basis. That potential conflict is recognised by the express obligation upon FX here to act fairly towards the Client. Absent such an obligation, I consider that there would have been a duty to conduct the determination in a way which is not arbitrary, capricious or irrational in the public law sense. See the Supreme Court decision in Braganza v BP Shipping [2015] 1 WLR 1661 at paragraphs 27-31, 52-53 and 102-103. For ease of reference I shall refer to these limitations on the exercise of a contractual discretion as “the Braganza Duty”. The fulfilment of that duty will entail a proper process for the decision in question including taking into account the material points and not taking into account irrelevant considerations. It would also entail not reaching an outcome which was outside what any reasonable decision-maker could decide, regardless of the process adopted. As noted in Braganza, however, the duty does not mean that the Court can substitute what it thinks would have been a reasonable decision. Further it may well not be appropriate to apply to contractual decision-makers the same high standards of decision-making as are expected of the modern state.

Because there was an express duty to act fairly, however, that would encompass the rather more limited Braganza Duty anyway.

In my view, the November Trades cannot sensibly, rationally or fairly be described as a Manifest Error on the part of FX nor could they be determined by FX as such. Looking at the evidence, it seems to me that FX has always been struggling to fit the November Trades into this category.

Given that FX intended the DD to operate as it did, with prices changing relatively slowly, it is impossible to see how FX ever quoted the “wrong” price. Nor could any error said be said to be manifest. It is not like putting a decimal point in the wrong place or mistakenly quoting silver instead of gold.

The fact that the quoted price was “off-market”(see below) does not mean that it was a misquote in the first place. Otherwise all the trades at that price would be affected by the same manifest error irrespective of how the trades were devised and set initially.

In truth, and as Mr Kasapis eventually suggested in his evidence, the question of manifest error with regard to these trades involved a consideration of who the customer was and what they were up to. But that can no longer be in the realms of a misquote focusing as it does upon the nature of the trading concerned as opposed to the quoted price. I am sure that FX made a mistake in the very broad sense that had it known in advance what Mrs Shurbanova was going to do, they would not have traded with her at all. So it was an error that they did. But that is not manifest error by reference to the price.

The real complaint here, in my judgment, is that there was abusive trading and in the end this was the ground on which FX placed most emphasis in its written closing submissions (“the Defendant’s Closing”). I now turn to it.

Since the main debate between the experts was on abusive trading and also compliance, it is appropriate to make some general observations at this stage. Both experts had different principal fields. Mr Carse’s was regulation and compliance. As he said on a few occasions, he was not a computer expert. He had been involved in compliance for about the last 13 years during which time he did no trading. Before then he had been a broker and originally placed orders on a telephone basis but was still involved in trading after its electronic form was introduced in the early 2000s. However it was not clear how extensive his own involvement was because as he accepted in evidence, he may simply have been observing his clients effecting the trades on the computer screen.

By contrast, Mr Kasapis was not an expert in compliance but was an expert in electronic trading and news trading and to this extent his evidence was much more on point. He explained in cross-examination that when he started his career in 1988 he was engaged in designing computer systems for trading but was also a junior broker in foreign exchange and had some trading experience principally in forward rate agreements. Between 1990 and 1995 he did some trading in FX products which could be complex or “exotic”. He had to program his own models in computer languages C and C + because of the complex nature of the products. When at SoCGen until 1999, he was working in forex trading in fixed income, credit default swaps bonds and asset swaps. While that was the last time he was involved in forex trading generally, he later closed out large FX positions for Anglo Irish Bank and closed out a number of foreign exchange rate options and forward rate agreements for a millennium hedge fund and a large number of foreign exchange positions (including CFDs) and bonds for the MF global liquidation. Most of this was for institutions but he also structured and traded some FX options for retail clients of National Australia Bank. Subsequent to 2005 when he moved into consultancy, he still traded for himself with Credit Hedge Limited in CFD’s and commodities. He now structures credit default swaps and credit products, which have FX elements, for Duff and Phelps. He rejected the notion that his knowledge of retail trading was really some time ago. As for the suggestion that his more recent experience was far removed from the issues in this case, he said “not at all”. I accept that and in my view this was reflected in the authoritative and detailed way in which he covered the technical issues here. In fact, on some key issues, there was not much between the experts anyway following cross- examination.

Abusive trading under Clause 27 is a different kind of justification for revocation than Manifest Error under Clause 26. It is to do not with a determination by FX as to whether there was an error on its part with regard to price which affected the trade; rather it is to do with the nature of the trading itself.

This leads to the question whether FX, in choosing to revoke on this ground is exercising a Braganza type discretion at all, as opposed to having a pure contractual power and then deciding whether or not to exercise it. In Mid Essex Hospital NHS Trust v Compass Group UK and Ireland Ltd [2013] EWCA Civ 200, Jackson LJ put it thus at para. 83 of his judgment:

See also the decision of Blair J in Sucden Financial v Fluxco-Cane [2010] EWHC 2133 (Comm) at para. 50.

Mr Stone submits that FX had a range of contractual options open to it in the event of abusive trading including, for example, revocation, amendment or doing nothing. That characterisation is not quite correct. Under Clause 27.1 the simple right is to revoke and then there is a separate right to make any adjustments to the client’s account. I agree that it is probably correct to see an abusive trade within Clause 27.1 as also constituting an event of default for the purposes of Clause 24.1 in which case certain other remedial action will also be available to FX. But none of that turns FX’s power into a discretion of the Braganza type which is concerned with a determination of a substantive matter, or a judgment about or evaluation of some state of affairs which one party makes as the decision-maker, but which affects the interests of both, hence giving rise to a potential conflict of interest. See, for example, paragraphs 18-22 of the judgment of Lady Hale in Braganza. The need to find a “target” for the determination in question (see paragraph 105 of my judgment in Watson v Watchfinder [2017] EWHC 1275 (Comm) supports this. It is meaningless to talk of FX’s determination of its consequential or secondary contractual powers including revocation (arising in effect by reason of a defined contractual wrong on the part of Mrs Shurbanova) as a discretion of the relevant kind. If it were otherwise, then it could be said that a party’s choice as to whether or not to rescind a contract for misrepresentation as opposed to seeking damages (one of which may be very much more to the advantage financially of the party in default) was itself a contractual discretion subject always to a Braganza Duty. That cannot be right.

Mr Stone seeks to draw support for his contention from the reference by Jackson LJ in the passage from Mid-Essex quoted above to “choosing from a range of options” as opposed to a “simple decision whether or not to exercise a contractual power”. But this must be read in the context of the cases Jackson LJ cited as giving support to the duty later confirmed in Braganza. Thus the case cited in his paragraph 78 was all about the determination by one party whether a port was dangerous, in paragraph 79 the case was about the power of the employer to set a discretionary bonus for the employee, in Socimer, referred to in his paragraph 80 it was to make a valuation and in the case cited in paragraph 81 it was about the allocation of particular logical channel numbers for the two shopping channels granted to the other party. All such examples support my analysis in paragraph 90 and not Mr Stone’s contention.

The exercise under Clause 27 is not an assessment or judgment being made by FX as to a variety of outcomes, it is simply a decision whether it wishes to exercise an absolute contractual right to revoke (among other possible remedial options) which has arisen.

Mr Stone further contends that it cannot be a simple contractual power unfettered by any Braganza Duty because FX has to make an assessment of whether there is price latency (or more generally, abusive trading). I disagree. On the face of the clear words of Clause 27, and unlike clause 26, it is not for FX to determine finally whether there has been abusive trading. Rather the question is whether there was, objectively, or not.

This has three important consequences, in my view:

I agree that the latter part of Clause 27 refers to accounts which rely on gaming and/or abusive strategies being made subject to intervention by FX “at its sole discretion”. But that is not about the impugned transaction (which is all that is in issue here) but rather the future operation of the account as a whole. So that express discretion is irrelevant to the present dispute.

But one does have to deal with the last part of Clause 27.1 which says that “Any dispute arising from such quoting or execution errors will be resolved by the company in its sole and absolute discretion.” The reference to errors can only be to the first sentence of Clause 27.1 which refers to errors which sometimes create a situation where the price displayed does not accurately reflect the market rates. If this part of the clause was meant to confer upon FX an absolute discretion (even if subject to the Braganza Duty) to decide all the previous matters including whether there had been abusive trading or gaming, then it would run counter to the clear words preceding it and the need to establish those matters objectively.

I do not think that the last part of Clause 27.1 should be construed so widely. In my judgment, it governs the situation where there may have been errors on the prices quoted or executed and FX wishes to correct them but does not allege any abusive trading or gaming in respect of the trades. Then, FX is empowered to resolve the dispute if the client object to what it wishes to do, but that resolution exercise would then be subject to the Braganza Duty. Accordingly, I do not think that the situation contemplated by the last part of Clause 27.1 is relevant here.

Accordingly Clause 27 is not a case of a contractual discretion which attracts the Braganza Duty. It is a pure contractual power.

As it happens however, in this case, and for the reasons set out below, it makes no difference to the outcome whether abusive trading has to be established objectively, or is to be determined by FX but subject to its Braganza Duty.

A key point made by FX was that the 8 November Trades were executed at prices which were significantly “off-market” as described above. This is not accepted by Mrs Shurbanova and indeed a principal feature of the Claimant's Closing was that the Trades were not “off-market” at all. I therefore turn to this question first.

An initial point taken is that recourse to Bloomberg indicative prices is really to no avail because there is no one set market price for the Trades undertaken here, as there might be on an exchange. But even Mr Carse does not go that far. He says in paragraph 29 of his report that the OTC market (with which we are concerned) relies on firms such as Bloomberg or Reuters “to supply a stream of indicative prices” and he repeated that it was reasonable to rely on Boomberg in his evidence at 3/99-100. Nor was it suggested in cross-examination to Mr Kasapis (or Mr Patel or Mr Cafe) that using Bloomberg prices as a comparison for the purpose of ascertaining when trades may have been made at “off- market” prices was a valueless or pointless exercise. Just as it was never suggested that the concept of being “off-market” was meaningless. By contrast, Mr Kasapis gave positive evidence that a company like Bloomberg did give prices which could be used as “market prices” when seeing whether a trade was “off-market”. I accept that evidence. I also accept his evidence that the response time of FX’s NDD platform was significantly faster than the DD platform and would normally track, although not necessarily be the same as, the Bloomberg prices.

As I also explained in paragraph 63 above, the comparison between the executed trade price for the gold at US$1,303.36 as against the Bloomberg price of US$1,292 was done at 8.30.19-24am. By that time, the price quoted by DD had actually fallen to more or less the same as the Bloomberg price, in other words it had caught up. It was thus also by now similar to the faster moving NDD price - see Mr Kasapis’s chart at page 17 of his report. That of course does not give the market price at the time when the Trades were placed or executed. Rather this was the time at which the trade was confirmed to the global blotter and those timings are shown in its “AO” column. Although that column is headed “execution date” Mr Carse said that the time of execution must have been earlier. In the end, and as to some extent presaged in the experts’ joint statement, Mr Kasapis agreed with him. He said that the time of execution should be taken as the (earlier) “Submission Date” which is column Q in the global blotter. On that footing, the Trades were all executed between the first and second seconds after 8:30am.

If one goes first to the later confirmation time on the global blotter (which is when FX made the comparison with Bloomberg in respect of the Trades) the difference between the executed price of US$1,303 and the Bloomberg price was some US$10. If, instead, one takes the comparison time as the earlier execution time, then, and using the NDD price as a proxy for Bloomberg, the difference is around US$7. See Mr Kasapis’ clear explanation about that in the transcript at Day 3/152 and 176, where he also said that he would expect the Bloomberg price to be about the same as the NDD price at that stage. There is no reason why I should doubt that evidence. And in order to give it some context, according to Mr Patel, while different market-makers may give slightly differing quotations on gold (assuming no time-lag on their pricing) the difference would be in terms of a few cents not several dollars.

A point was taken in the Claimant's Closing (though not addressed by either expert and not put to either of them in cross-examination) that if one looks at Bloomberg figures which were subsequently obtained, in particular at page 1073 of bundle 3C at the top, it shows that the opening Bloomberg price of US$1305 was even higher than the quoted price and it then dropped down to US$1,294. But quite apart from the fact that this was not canvassed in evidence the suggestion that these figures show that the Bloomberg price was “lagging” just as much as the quoted price on the DD, or that it behaved in the same way within that the first second after 8:30am, has no evidential basis. What page 1073 does not tell us is precisely when the price of US$1305 dropped and here, milliseconds matter. This is the critical point because the quoted price on the DD remained at least until 1 second and 748 milliseconds after 8.30am whereas (for example) the NDD price had dropped before. See again Mr Kasapis’ chart at page 17 of his report. The key thing was to “catch” the quoted price on the DD before it had dropped. It was also suggested that at the time when the comparison should have been made i.e. the times shown in column “Q” there was no difference between the DD and the NDD prices. But yet again, that is contradicted by the chart at page 17. Indeed, Mr Kasapis was not really challenged in cross-examination as to the correctness of the difference between the prices of at least US$6-7. And in relation to all of this, Mr Carse was not really able to contribute to the debate because he did not provide anything like the same detailed evidence or analysis on the pricing point. As he himself said, for the most part, his expertise lay in the regulatory or compliance side of things.

In addition, in his evidence to the court, Mr Kasapis said that the Bloomberg price, which had already reacted to the news event, would also have been at about $1,292 at the time when the Trades were actually placed. He may not have that exactly right if the lowest that the Bloomberg price fell to between the first and second seconds was US$1,294. But for the purposes of the argument, that is close enough. And although the DD price itself had caught up by 8.30.19-24am, this did not matter because the Trades had been executed at the original quoted prices. Further, as Mr Kasapis said, any slippage (and here there was none) would have occurred before the time given in the “submission date" column Q. See Day 3/169-177.

As a further point, although again not put to either of the experts, Mr Stone referred to Appendix C to Mr Kasapis’s report. In a table headed “Market Value of Trades”, he works out the total value of the Trades on a gross basis as US$131,034,331 in order to make the point at paragraph 9.35 of his report that this was not a size of trades with which individual traders would usually be comfortable. It is pointed out that within this box, for the Gold Trades, the trades are put at US$1293 whereas the Bloomberg price is put at US$1,303. In other words, the Bloomberg price was the same as the Gold Trade price hence there could be no “off-market” dealing here. There is nothing in this. The Gold Trades were not executed at US$1,293 but US$1,303. The buyback was then at US$1,293, hence the profit. So I suspect that all that has happened here is that the names and figures have been put the wrong way round. In any event none of this was put to Mr Kasapis in cross-examination.

In addition to the Bloomberg price taken as a reference point, it is the case that other gold “sell” trades done in the same second were at significantly lower prices (other than Investom dealt with below) and this indicated that the Gold Trades were effected using a very fast news source. See paragraphs 30, 31 and 67 of Mr Patel’s witness statement which I accept.

A yet further point was taken that Mr Kasapis’ Graph 2 at page 18 of his report really showed that there was no material difference between the DD and the NDD prices at the relevant time. But there is nothing in that either. Here, one cannot simply speak of what might have happened within a range of one second. We know that both of these prices had approximated towards the end of the first second after 8:30am. But the issue is what happened beforehand. Graph 2 shows that the DD price stayed at US$1,303 much longer (in relative millisecond terms) than the NDD price. It is true that something happened over an extremely short period of time which made the NDD and later the DD price spike (the spike does not appear on the chart at p17 of Mr Kasapis’s report). It is not clear what that is but it makes no difference to the general point about the obvious delay in reaction time of the DD price.

It is right to say that on the evidence, the divergence between the prices at the time of execution by reference to column “Q” is about US$7 as opposed to US$10. But that is quite enough of a divergence to constitute off-market pricing and there is no evidence to the contrary.

All of that seemed logical and sensible to me and indeed Mr Carse did not really dispute it.

For all those reasons there is simply no warrant for the assertion made in paragraphs 57 and 69 of the Claimant's Closing that the evidence showed that there was no divergences at all between the price quoted for the Trades and at which they were executed, and the equivalent NDD or Bloomberg prices, so that there was no off-market trading.

The above points deal with the Gold Trades, which produced 83% of the total profits from the Trades.

As to the Dollar Trades, a preliminary point taken in paragraph 72 of the Claimant's Closing is that there could not be any abusive trading because, as these bespoke FX products are not available on the NDD, a trader in the position of Mrs Shurbanova would simply have no other “market price” with which to compare the quoted price on the DD feed. This misses the point which is that assuming that the dollar (including a dollar measured by reference to a basket of other currencies) would go up in the event of good news a profit would be virtually guaranteed provided that the “slow” quoted price on the DD lagged behind.

Although Mr Kasapis did not have “market” figures with which to compare the actual price at which the Dollar Trades were executed (ie the “buy” orders), in section 8 of his report and over 3 pages, he gave a careful and reasoned analysis as to why one could infer that the DD quoted prices must have been off-market here, too. He was not cross- examined on any of this. Accordingly the points now made in paragraph 72 of the Claimant's Closing cannot be given much weight. But in any event, I accept that because there is an inverse correlation between the dollar and gold and that the actual “market price” could only be sourced from a fast feed, it was inevitable or at least highly likely that the quoted prices on the Dollar Trades of US$10,528 were off-market. Mr Carse gives no relevant evidence on this point.

At paragraphs 84-86 of his witness statement, Mr Patel recounts that his investigations at the time found that the price difference achieved between the buy and sell prices of 43 to 45 points were atypical and more indicative of differences one could expect to see over a whole day rather than the very short period here. Under cross-examination he did not resile from this view nor was it really shown to be wrong – see 2/90-96.

Here, reference is also made to the Bloomberg prices, this time for the “straight” dollar (as opposed to the dollar instrument traded here) at 3C/1077 Those prices range from US$10,535 to US$10,548 (Mr Kasapis had said that the operative time range for the trades’ execution was 13.30.01.921pm and 13.30.19.530pm). While not the same as the execution price of US$10,528, being somewhat higher, it could be said that the price differential was not as marked as it was in the Gold Trades. I see that but first, the comparison is not exact because the Bloomberg prices are not for the dollar measured against four currencies and I am not prepared to accept without evidence that the differential on the correct basis would have been exactly the same. Second, and as with the Gold Trades, none of this was put to Mr Kasapis.

Accordingly, I find that the Dollar Trades were also executed at off-market prices.

In my judgment, what happened here was classic abusive trading. The use of FNG in conjunction with the application used by the Shurbanov family so as to trigger the trades (and trigger them only) upon receipt of positive news from NFPD meant that those trades were placed with knowledge of the outcome. They were not based on predictions and there was therefore no risk.

It does not matter that Mrs Shurbanova herself may not have known what the actual NFPD results were until later because the application, through its connection with FNG, did. Nor does it matter that she was not using a fast price feed at the time with which to compare the DD quoted prices. She did not need to because the Trades would be placed automatically at particular levels upon receipt of the designated data. That was the whole point of the FNG. So while (as Mr Kasapis accepted) a news feed is not the same as a price feed, it makes no difference here. It is not suggested (by anyone) that Mrs Shurbanova had a very fast price feed which in effect would show market prices already below the quoted prices for the Trades. On the assumption (proved correct) that there was a time lag on the DD quoted prices all one needed was the ability to place orders extremely quickly after the NFPD announcement in the direction dictated by the announcement itself. That was what FNG did as a fast news feed. Nothing else was required and Mr Kasapis did not suggest otherwise. See the totality of his evidence at 3/136-137, and at 181-184. Accordingly the points made in paragraphs 75, 76 and 78.10.1 of the Claimant's Closing are incorrect.

I thought that the combined evidence of Mr Cafe, Mr Patel and Mr Kasapis was compelling as to there having been abusive trading here. Moreover, and crucially, Mr Carse himself accepted twice in cross-examination that the circumstances described in paragraph 11.18 of Mr Kasapis’s report could amount to abusive trading. Headed “Latency Trading” this passage said as follows:

“In this case a trader uses a high-speed feed and a low-speed execution platform, the high-speed feed used to gain information on the current market price or market direction, the trader then places orders on the slow execution platform to take advantage of its stale prices.”

The high-speed feed here was FNG acting upon the application. On Mr Kasapis’s definition, there was Latency Trading here and in the end Mr Carse did not really dispute this. His concerns were more compliance-related.

While both experts agreed that the application (or the application combined with the FNG) was not a sophisticated algorithm, that is beside the point. Its functionality enabled the abusive trades to be effected.

The timing of the Trades and the “off-market" prices at which they were made are hallmarks of abusive trading, according to Mr Kasapis. Another hallmark, he said, is the placing of many orders at smaller amounts as opposed to one very large order. Overall there has to be a very substantial value in the Trades so as to make a significant profit when they are closed out. But placing a single very large order would run the risk of the trades failing altogether or a less favourable price being offered. Smaller trades command more favourable prices even if there are many of them and this could be achieved on the DD platform precisely because it enables the genuine low-level trader to participate. Again, in the end, in his oral evidence, Mr Carse did not really challenge this. The point is that it is the combination of many small orders placed within an extremely short time frame facilitated by software like the FNG which made the Trades abusive. The fact that, taken by itself, for example, a large number of small trades might not be abusive is beside the point. It is correct that the Trades were not closed out in milliseconds as Mr Patel had said in his witness statement. But that makes no difference to the abusive features here. Nor, in my judgment, did this affect the general reliability of his evidence. Nor does the fact the comparison made by FX was at the confirmation rather than the (earlier) submission point.

Equally I do not accept that the very large size of these Trades on a gross basis, of US$130m was not significant, because Mrs Shurbanova may simply have been confused or unaware of the risks as a new trader operating on her own. That is quite unrealistic here for the reasons given in paragraphs 37-71 above. Again, it is all a question of context and here the context pointed in only one direction – see the evidence of Mr Kasapis at 3/146-147.

A further point was made that too much emphasis is placed by FX on the simple fact of a large profit, and its “trigger” point where a customer seeks to withdraw US$75,000 or more from his account. In that event FX will investigate before allowing the withdrawal. But I do not think there is anything in this. It is all a question of the context in which the profit is made, as Mr Cafe explained. If profits are small even if made through abusive trades, FX might led them go especially if the overall trading relationship is profitable for FX. The US$75,000 threshold recognises this, but in any particular case FX may choose to intervene (or not) anyway. None of this suggests that FX have no interest in seeing whether a particular trade is off-market (as an element of abuse along with the others). After all, on my analysis of the Clause 27 power of revocation, if that is the only basis for revocation, FX need to make sure that there actually was abusive trading.

Mrs Shurbanova also relied upon the fact that similar trades had been made on the same day by Investom and yet those trades were not revoked. Thus it is said that there could not have been any abusive trading by her. There is nothing in this point. The Investom trades were very much smaller and for far fewer units per trade, being only 5 (a total of 1,900 as opposed to 37,500 units) and all bar one were at much lower opening prices, they were in the main only partially filled and the profits were very much smaller. See 3B/584. They did not present the same collection of “red flags" as the Trades did. See the evidence of Mr Kasapis at Day 3/166-167. As noted in the previous paragraph, therefore, the context was quite different.

It was further contended on behalf of Mrs Shurbanova that the Trades could not be an abuse when what was being taken advantage of was FX’s own slower prices which it had deliberately set up, so as to enable the non-professional individual traders to trade more easily. I disagree. The whole point is that taking advantage of the price latency inherent in the “throttled” price offered on DD was literally an abuse of the system put in place. It was not intended that users of the DD should act as professionals and make risk-free profits, based not on intelligent predictions of market movements, but purely on the inherent time lag before the DD prices accurately reflected the news event. As Mr Kasapis put it, “...because you know where the market is going. You have a five second preview into the future. Assuming the latency is five seconds between the two.” The advantage is because the trader already knows the outcome.

Equally the fact of occasional trading combined with other factors can indicate abusive trading, as Mr Kasapis said.

A number of criticisms were made of Mr Kasapis which I do not consider are justified. First, it is said that his opinion that the FNG was an extremely fast tool was unsupported by evidence. But he referred to the test results he carried out in paragraphs 9.15 – 9.17 in the absence of the continued availability of the FNG software itself. But first he was not specifically cross-examined on the data he saw and second even Mr Carse accepted in cross-examination that FNG was certainly an effective tool and he knew that it had been sold as a product that had the ability to read data releases from the exchanges “extremely quickly”. Here, what was mainly put to him in cross-examination was the distinction between a fast news feed and a fast price feed. But as a explained above that point is misconceived. Moreover Mrs Shurbanova adduced no positive evidence to suggest that FNG did not move very quickly so as to give her an advantage over the “slow” DD price.

Other points were made like the fact that at paragraph 4.1 of his report Mr Kasapis said that the NFPD was released at 13.30 while in paragraph 7.19 he said it was 13.30.01- in context nothing turns on this, nor does the fact that at paragraph 7.17 he referred to the close-out price for the Gold Trades being US$1,292 whereas the exact figures in his own table were between US$1,292.98 and US$1,293.22. That is a hopeless point since at paragraph 7.17 he was not dealing with the precise figures. Yet again neither of these points were put to him in cross-examination. Nor was an alleged error in how he described Mrs Shurbanova’s references in her witness statement to her “demo” trading. I accept of course that Mr Kasapis did change his position about when the price comparison should be made. But he presaged this before the trial and I do not consider that this affected his reliability generally.

For all the above reasons, it is plain that there was abusive trading in respect of the Trades. That is clear from the nature of the Trades themselves and how they were made. While the true role of Mrs Shurbanova (as I have found it to be) as a cipher for her husband and/or Martin Shurbanova supports that conclusion, because FX had already taken action in particular against Mr Shurbanova for the same reason, it is not necessary for it. But it would have presented an incomplete picture not to have dealt with the allegations about her role. And it is also relevant to the misrepresentation claim (see below).

Accordingly, the revocation was justified and one need look no further. Even if (contrary to my conclusion) the power to revoke under Clause 27 for abusive trading was a determination to be made by FX but subject to its Braganza Duty, the result would be the same, as accepted now by Mr Stone – see paragraph 77 above.

Accordingly Mrs Shurbanova’s claim against FX must be dismissed.

In these circumstances it is neither possible nor appropriate to make alternative findings on the hypothesis that in fact there was no abusive trading. First my finding that there was is a factual and evidential question for the Court. Second, no sensible further finding could be made without knowing why in fact the Trades were not abusive.

In the light of all my conclusions above, it is not necessary to deal with this issue because I have found that there was abusive trading within the meaning of Clause 27.

However, I would make some brief observations. The first way the claim is put is to say that there was a breach by Mrs Shurbanova of Clause 23 of the TOB where she gave a number of contractual representations and warranties. I agree because Mrs Shurbanova plainly made false representations about her prior trading experience, and her personal financial position. As to prior trading she clearly did not do the regular trades previously which she described. Indeed, in my judgment, she could not have done so by herself. And whatever in fact her personal financial position was (which is not clear) I am quite satisfied that it was not in the amounts put forward. Indeed she accepted she put forward what she said was the joint or “family” position. And overall she misrepresented her own role – in truth she was not acting for herself but for her husband and/or Martin Shurbanova. There was therefore a breach of Clause 23 which itself constituted an event of default which would have justified the revocation had that been necessary.

The alternative head of misrepresentation under s2(1) of the Misrepresentation Act 1967 would also have succeeded. Mrs Shurbanova’s pre-contractual misrepresentations were plainly material and serious.

And I accept the evidence of Mr Cafe at paragraph 80 of his witness statement that FX would not have opened the account if it understood that Mrs Shurbanova was associated with Mr. Shurbanov and certainly would not have permitted her to trade CFDs which is what these Trades were, and this was not in fact challenged in cross-examination. So causation and reliance are shown. So again, had it been necessary, FX could have claimed damages under the first or the second heads of misrepresentation claimed. That would have cancelled out any liability on its part for damages to Mrs Shurbanova.

In the event, the claim is dismissed. I am most grateful to Counsel for their assistance and comprehensive submissions.