Citadel Securities (Europe) Ltd v GSA Capital Partners LLP & Ors

The issues which I have been asked to decide in this costs and case management hearing are:

Mr. David Craig QC appeared for the claimant with Ms. Rogers and Ms. Taunton and Mr. Simon Devonshire QC appeared with Mr. Forshaw for the defendants. I am grateful to both legal teams, including the instructing solicitors, for the thoughtful, thorough and helpful way in which their respective positions on the issues were presented. I am also grateful for the constructive way in which they approached the resolution of those issues at this hearing.

The claimant is a global financial services group of which GSA is a global competitor. Defendants 3 to 6 are senior officers of GSA and the seventh defendant is an in-house recruiter. The part of the claimant's business with which the Claim is concerned is algorithmic trading, that is, the use of computer algorithms and statistical models to trade in the financial markets. Algorithmic trading businesses develop trading strategies and then translate those strategies into computer commands through a combination of C++ Code, related configuration files and other relevant computer systems and programmes. The key aim is to devise combinations of signals which identify price dislocations in the market which have not been identified by competitors, and to exploit them by trading profitably without undue risk. The trading strategies and the principles and logic on which they are based are therefore highly confidential and, where the strategies are successful, highly commercially valuable. They are also complex and expensive to develop.

These proceedings concern one of the claimant's trading strategies, known as the "ABC Strategy", which is highly profitable. Mr. Cologlu is the claimant's former Head Trader for ABC Strategy in Europe and was at the material time a London-based Quantitative Researcher employed by the claimant. It appears that between the fourth quarter of 2018 and June 2019 he was involved in discussions with GSA with a view to his recruitment, and that the plan was that he would establish an algorithmic trading business for GSA.

The claimant's case is that in the course of those discussions Mr. Cologlu copied and disclosed its trade secrets and confidential information to the defendants in three key documents which were created by Mr. Cologlu and provided to the defendants in hard copy. These were:

The DDQ was prepared in answer to a set of 96 questions which were asked of Mr. Cologlu by the defendants under ten headings including, for example, "Strategy overview", "Signal research" and "Risk considerations". The DDQ has 10 corresponding headings and it answers each of the 96 questions in turn. It appears to contain information about an existing strategy as well as setting out what would be required for GSA to establish such a strategy. The claimant says that it contains highly confidential information as to the "core portfolio logic used in the ABC Strategy and its related strategies".

The claimant also relies on three emails which were sent to GSA by Mr. Cologlu. However, it is clear that the DDQ is its primary concern.

I will refer to these six documents the "disputed documents".

Mr. Cologlu's activities were discovered on Friday, 7 June 2019 and on the following Tuesday, that is 11 June 2019, injunctive relief was ordered against Mr. Cologlu by Butcher J under section 44 of the Arbitration Act 1996. The claimant has also brought proceedings against Mr. Cologlu in the London Court of International Arbitration in accordance with arbitration provisions in his contract of employment. A liability hearing is listed in those proceedings in February 2021.

The defendants have also delivered up the documents which they say they have received from Mr. Cologlu. By letter dated 14th June 2019, they gave undertakings confirming that no copies of the disputed documents had been retained and that no use had been or would be made of their contents. The fourth to seventh defendants have also sworn affidavits which explain what they say is the limited use which they accept they made of the documents and information provided to them by Mr. Cologlu, and they have given further undertakings confirming that they will not make use of the information contained in them in the future. GSA and the fourth to seventh defendants, as well as the recruitment consultant who assisted them in the recruitment process -- that is the external recruitment consultant -- also provided pre action disclosure on 23rd August 2019, they say, of all relevant documents

The proceedings in the Queen's Bench Division, with which I am concerned, were issued on 16th December 2019. The claimant alleges that information which Mr. Cologlu provided in the disputed documents was copied or derived from its ABC Strategy and related strategies, that this information was highly confidential and that the defendants were aware of this and solicited the provision of such information by

In relation to remedy, the claimant claims declarations, final injunctive relief "negotiating damages" and "moral damages" as well as damages for losses such as wasted management time and forensic IT costs, equitable compensation and/or an account of profits. Its financial claims are set out in a Provisional Schedule of Loss dated 7th August 2020, which values its claim for negotiating damages at not less than US$30 million and its claim for moral damages at not less than an additional $10 million. These figures, the claimant says, are on the basis of the defendants' admitted use of the information in the disputed documents. The claimant does not accept that the use was limited to what has been admitted by the defendants and it says that it is to be inferred that there was greater use than was admitted. Indeed, at paragraph 55 of the particulars of claim, it specifically pleads as follows:

On this basis, the claimant pleads that the true value of the Claim is likely to be "very substantially higher" than the figures I have mentioned.

The defendants' position is that they were involved in routine recruitment discussions with Mr. Cologlu in which they considered a pitch by him to set up a new high frequency trading strategy trading in European equities. They say that the discussions were at a high level of generality. Mr. Cologlu did not disclose any trade secrets or confidential information, nor any information which would enable GSA to reverse engineer Mr. Cologlu's proposed business strategy. Nor did they induce him to do so, or think that he was doing so. They say that the sort of information which he disclosed is widely known in the sector and/or is in the public domain. As a matter of common sense, they say, Mr. Cologlu would not disclose information which was confidential or of significant value as this would undermine his own negotiating position.

The defendants put the claimant to proof that any information which Mr. Cologlu disclosed was copied for derived from the ABC or related strategies. They point out that passages in the DDQ suggest that there would be a period of build-out after Mr. Cologlu was recruited and that there would therefore be no trading for 12 to 15 months after recruitment. They say that this is indicative of the fact that there was no plan to transfer the claimant's models to GSA. They say that the documents provided by Mr. Cologlu were kept confidential within GSA and the DDQ had barely been considered given that it was received on 6th June and handed back with the other disputed documents on 13th June 2019. They also say that they did not have a European equities algorithmic trading strategy at the time of the discussions with Mr. Cologlu and that these discussions were brought to an end by the claimant's actions in June 2019. He was not recruited. The defendants have not sought to set up a high frequency trading equities strategy since then and nor have they any plans to do so. They say that they have made a clean breast of what took place through delivery up, the affidavits and the pre-action disclosure which has been provided. There is, therefore, no inferential case that they did more than they have admitted and the disclosure shows, and the claimant’s financial claims are grossly inflated in any event.

At paragraph 28(c) and (d) of its reply and defence to counterclaim, the claimant has responded to this line of argument as follows:

There is also a counterclaim in which the defendants say that the claimant was guilty of soliciting breach of confidence when recruiting a Mr. Taylor from GSA in May 2019. Mr. Taylor was GSA’s Co-Head of Macro Research and Trading. The relief sought is a declaration and an injunction. No claim for damages is made. The counterclaim appears to be a rhetorical in nature and designed to illustrate that inevitably in a recruitment process a certain amount of information about the employee's skills and experience will be sought and provided, and that this will inevitably include information about skills and experience gained by the employee at their current employer.

Mr. Craig QC began by emphasising the complexity of the evidence in the case and drew particular attention to paragraphs 46 and 47 of the witness statement of Mr. Weiner which state:

Mr. Weiner's evidence as to the degree of complexity of the evidence at trial does not appear to be materially in dispute.

Mr. Craig also confirmed that the claimants' case of copying and derivation is not based on cutting and pasting from a particular document. The claimant has set out its case as to which information in the disputed documents has been copied or derived and which is confidential in an Annex to which I will refer as “the Annex”. The Annex breaks the text of the disputed documents down into lines or sentences and identifies by use of crosses in boxes whether the information in each line or sentence is said to have been copied or, alternatively, derived from the ABC or related strategies and whether it is said to be confidential in itself or confidential taken in aggregate with

The claimant originally agreed standard disclosure on all issues and did not suggest that there be an approach based on sampling. Mr Craig said that in the light of the claimant’s experience in the arbitration proceedings against Mr. Cologlu, it has reflected on the feasibility of this approach and has concluded that it is not proportionate given the complexity which it entails.

On 20th October 2020, the claimant therefore made an application for an order that each side identify 60-line entries on the Annex. Under this proposal, the claimant would identify 60 items which it contended were confidential and the defendants would nominate 60 which they contended were not. These entries would then be the subject of disclosure and would be litigated at the trial.

On questioning from the court about the detail of this proposal, Mr. Craig confirmed that:

Mr. Devonshire QC took instructions and indicated that he would be potentially be minded to agree Mr. Craig's proposal provided:

Mr. Craig responded that he thought it may be possible to reach agreement on this basis. He made clear, however, that he would wish to rely on the 96 questions which the defendants asked Mr. Cologlu, to which the DDQ provided the answers, in support of his case that the defendants were knowingly inducing breaches of duty by Mr. Cologlu. The hearing was therefore adjourned overnight to allow counsel to attempt to reach agreement and, in this event, draft accordingly.

On the second day of the hearing it emerged that agreement had not been reached. The key concerns were as to:

At Mr. Devonshire's request I adjourned for the parties to discuss whether these issues could be addressed by agreement but, unfortunately, they could not be. I was therefore invited to rule on the claimant's application and decided to reject it, both in the modified form in which it was presented orally and in its original written form which, as I have said, asked for an order that the defendants nominate 60 items of their own.

My reasons for rejecting the modified application can be summarised as follows.

First, as to complexity, I accept that the evidence in this case will be highly complex from a technical point of view, but I do not accept that at trial the judge is likely to adopt a line-by-line approach as has been adopted by the claimant in the Annex. While it is likely to be necessary to consider whether certain specific items of information are confidential or trade secrets, in respect of others the case is already pleaded in terms of their effect when taken in aggregate with the other information in the disputed documents is confidential. I think it likely that, where appropriate, the judge will look at the information provided under the ten headings about which the defendants asked questions and/or Mr. Cologlu provided information in the DDQ and/or that the judge will look at particular paragraphs rather than adjudicate the quality of all of the information line by line. Indeed, both parties agree that witnesses of fact and experts and advocates are likely to adopt a similar approach.

Thus, although the evidence is undoubtedly complex, I do not accept that the process of adjudicating the issues will be as detailed as it has been presented in the Annex.

Second, for similar reasons I do not accept that any approach other than Mr. Craig's will result in a six-week trial on liability. Even if it would, the claim is valued by the claimant at not less than US$40 million and it puts the reputations, and potentially the livelihoods, of at least the four individual defendants at stake. The prospect of such a trial is therefore not disproportionate.

Third, I have a very real concern about the fairness to the defendants if, as Mr. Craig advocates, they are not to be permitted to ask the court to find that the bulk of the information in the documents was in the public domain and/or of a general rather than a confidential nature. This matters because the central issues in the case include the defendants' knowledge, belief, understanding and intentions in their dealings with Mr. Cologlu as well as the extent of their use of the information which they received. As I have noted, they are accused of bad faith and indeed mendacity in relation to these issues and they face a very substantial financial claim. It is said, in effect, that they were obviously seeking “weapons grade” information, that that is what they received, that they considered the disputed documents carefully, that they were well aware of the nature of what they received, that they made more use of the information than they say and that it is of great value to them in the future. Their position, on the other hand, is that even if the claimant established that the items which it nominated were collectively or individually confidential information or trade secrets, the vast majority of the information which they received was of general and high-level nature (as they put it), they therefore did not give it the attention alleged or make particular use of it. They say that their actions and the credibility of their evidence about their actions should be judged in that light. In this connection I note that the most confidential information in the DDQ has been redacted but it is apparent from the redactions that the information appears to be short passages from what, overall, is a relatively lengthy text.

Fourth, I have concerns about creating pitfalls in relation to other aspects of the case, including disclosure and evidence going to issues other than those which Mr. Craig's proposal seeks to address. By way of example, Mr. Craig was clear that under his proposal there would be no constraints on either party in relation to the use of other material in the case to make or refute charges of wrongdoing:

These sorts of issues also seemed to me to have the potential to complicate an already complicated disclosure process, with the parties wrangling about whether the material did relate sufficiently directly to the 60 nominated items.

Fifth, as Mr. Devonshire pointed out, it is within the gift of the claimant to reduce the scope of its case. There is nothing to stop it from limiting its positive case to a narrower category of information.

As to the claimant's intermediate proposal, which involves the defendants also nominating 60 items of information, but this time information which they say is not confidential, to my mind this meets with essentially the same objections in terms of potential unfairness, artificiality and the creation of evidential pitfalls. But it also requires the defendants to nominate without first having had disclosure. This is not quite a “lucky-dip” approach, as Mr. Devonshire described it, given that the defendants say that they are able to plead a case as to whether the sort of information which the disputed documents contain is high level and generally available. But it does add to the risk of unfairness given that the claimant is currently in possession of the requisite information to make its choice and prove its case on copying and derivation whereas the defendants are not; or at least not entirely.

I therefore refused the Claimant’s application.

Before turning to the question of a split trial, I should say something about the proposed expert evidence in the proceedings.

The parties are agreed that there will potentially be two types of expert evidence. First, there is a proposal for experts in algorithmic trading. For the claimant this will be a Professor Clifford and for the defendants this will be a Dr. Thomas Auld. The issues which these gentlemen will address have not been finally agreed but are currently as follows:

The claimant is sceptical about the use of expert evidence to address these issues and has reservations about issues 5 to 7 in particular, but it has been agreed that any arguments as to the admissibility of the reports of these experts should be dealt with after reports have been served. I am content to adopt this approach.

Second, there will be accountancy experts. The issues in relation to these experts are agreed as follows:

It will be seen that these questions go exclusively to the claimant's claims for financial remedies including valuation for the purposes of the negotiating damages claim.

The competing positions under this heading are that:

I was referred to CPR 1.42(h) and (i) and reminded that I should exercise my discretion in accordance with the overriding objective. Both parties took me to judgment of

Mr. Craig also relied on Leaflet Company Limited v Royal Mail Group Limited [2009] UKCLR 232 to illustrate the point that where there may be a range of outcomes in relation to liability, a split between liability and quantum may be appropriate so as not to overburden the court and the parties with the need to prepare for and consider the implications of multiple different hypothetical permutations in terms of the outcome, the majority if not all of which may never eventuate.

Mr Craig also referred me to the decision of Morris J in Marussia Communications Ireland Limited v Manor Grand Prix Racing Limited [2017] EWHC 901 (Com) to illustrate the conventional approach in trade mark dispute cases particularly where there is a claim for an account of profits as here. In the Marussia case an application to deal with liability and quantum at one hearing was refused despite the fact that it seemed more likely the claimant would elect for damages rather than an account of profits and damages were most likely to be decided on Wrotham Park principles.

Mr Devonshire relied on Road Chef Limited v Ingram-Hill and another [2013] EWHC 939 (Ch), but this was in truth an illustration of the application of paragraph 5 of the judgment of Hildyard J in the Electrical Waste case to the particular circumstances of the Road Chef case. Mr. Devonshire also referred me to the very helpful summary of the principles at paragraphs 25 to 32 of the judgment of Bryan J in Daimler AG v Walleniusrederierna Aktiebolag and others [2020] EWHC 525 (Comm). His analysis of the Leaflet Company case at paragraph 60(3) and his observations about judicial resources at paragraph 61.

Mr. Craig accepted Mr. Devonshire's point that, in relation to liability, in broad terms the court would need to consider:

Taking its findings on these matters into account in the context of the evidence as a whole, the court would then assess the seriousness of the defendants' conduct and the value of the information provided to the defendants in coming to a view on the quantum of any negotiating or moral damages.

Mr. Craig pointed out that, as matters stand, there are numerous items of disputed information: in the order of 300 were said to be confidential. He said that the court could reach a range of findings as to what or was not copied or derived, was or was not confidential, as to degree of confidentiality and as to the degree or extent of the defendants' use of that information.

Mr. Craig's argument was that the trial should be divided into two parts. In Part 1 the witnesses, including the experts on algorithmic trading, would give evidence about the questions at paragraphs 49 (i) to (v) above and the judge would then make findings which would then form the basis for their decision on liability. Any decision about financial remedy would then be dealt with in Part 2. In Part 1 the judge would also deal with the claim for injunctive relief.

Mr. Craig said that at the first stage the claimant may lose, or the case may clearly not warrant a financial remedy (although, of course, there is a certain irony about his reliance on these possibilities given that the claimant says that it should and will win, and that substantial sums should be awarded), or it may settle. In any of these events, the litigation would be at an end and the costs of preparing the evidence on the financial claims would be saved.

If the litigation did not come to an end at this stage, there would be a decision by the claimant as to whether the remedy it sought was an account of profits or negotiating and moral damages. In the event that it was the latter, the accountancy experts would give valuation and other evidence in the light of the actual findings of the court. This would mean that their evidence was given on an actual, rather than a hypothetical, basis and would be of assistance to the court -- the evidence would deal with the case as found rather than a hypothetical outcome or range of outcomes -- and it would therefore avoid wasting costs. The alternative was that the accountancy experts prepared valuation and other quantum evidence at substantial expense, only for this ultimately to prove to have been unnecessary, or they prepared evidence which was not particularly helpful as it did not address the case as found and the accountancy experts were, therefore, obliged to carry out a second exercise in the light of the actual findings of the court.

Mr. Craig pointed out that the issue was being raised early in the proceedings and said that the demarcation line in terms of the evidence which would be adduced in Parts 1 and 2 respectively was clear and it did not involve any overlap. He said his approach also maximised the chance of a speedy resolution of the litigation. On Mr. Craig's approach, he said that liability would be determined at a hearing in January 2022. This might well bring an end to the litigation but, if it did not, the quantum hearing could be dealt with by the end of the year particularly if disclosure on all issues took place in the context of Part 1 of the proceedings, which he agreed it would.

Mr. Craig said that there would be no prejudice to the defendants if the question of quantum were delayed as they have no financial claims. The prejudice was all to the claimant which would have been kept out of its money in the event that (a) it won and (b) the Claim is financially valuable.

Mr. Craig rejected the suggestion that his approach would make settlement less likely as there would be no evidence dealing with quantum. Disclosure would take place on all issues, the parties would seek expert opinion if they wished and they were in any event sophisticated parties and well able to judge the realities of litigation.

He also said that the risk of any bifurcation of the appeal process was overstated. I should not manage the case on the basis that there might be an appeal. Even if there were an appeal on liability, it could be stayed pending the outcome on remedy. He said that the arguments about the possibility of an appeal were not all one way. He raised the possibility of the claimant losing on a trial of all the issues and the judge then not making findings on quantum or having to re-do the whole of the case on the basis of a successful appeal on liability because the judge's findings on liability were flawed.

As to the defendants' point that the judge in Part 1 might not be able to deal with costs without a view about the value of the case, Mr. Craig said that the judge could, if necessary, reserve the costs.

On balance, I have decided to reject the application for a split trial. I accept that the factors which fall for consideration overlap and can be relied on to support competing conclusions, but for the following reasons I do not consider that the overriding objective will be best served by the course proposed by Mr. Craig.

I accept that Mr. Craig's proposal as to the demarcation of Parts 1 and 2 is, in principle, capable of clear definition, but this is not a case in which the findings on liability are separate from the considerations which are relevant on quantum. Obviously, they never are, but this is not a case where, say, the court is asked to make a finding of liability for negligence or breach of contract before going on, at a later stage, to make essentially discrete findings as to the claimant's injuries or the financial consequences of the breach of duty. It is a case in which the court would be asked to make findings which directly impacted on the valuation of the information disclosed and moral damages and then, rather than apply those findings, postpone reaching a conclusion so that experts could give opinions as to what flowed from them.

There also seems to me to be a significant risk that, in the event that liability was established, even the most conscientious and thorough judge would not address, in the liability judgment, all of the questions of fact which would need to be answered in order for the accountants to carry out their valuation exercise. The judge could not guarantee to do so, precisely because what the experts specifically regarded as necessary to their task would not be known. There is, therefore, a substantial risk that the proposed valuation exercise carried out by the court in Part 2 would involve evidence and argument which might have been deployed on one occasion but are spread over two hearings. In this event, witnesses might also need to give evidence twice.

In this connection, I also note that the draft questions for the algorithmic trading experts (which I have set out above) potentially straddle liability and quantum in the case of questions 5 to 7 but particularly question 7. Although no final position has been reached as to admissibility in relation to the experts' answers to these questions, as matters stand the evidence will be prepared and it does appear to me that there is a close interrelationship between the answers to these questions and the evidence on which the accountants will base their opinions.

I also accept that in principle a split trial may lead to a saving in costs in that if liability is not established, or the case is disposed of by agreement after the liability hearing, the costs of the expert accountancy evidence will not have been incurred. If liability is established, the expert accountancy evidence is also likely to be more focused if it is prepared after judgment is handed down given that the evidence would be based on the actual case rather than the hypothetical cases or a range of scenarios.

But against this I weigh a number of factors in addition to the ones already mentioned. First, in the real world it seems to me that there is a significant risk of duplication of work and costs if the case has to be prepared twice. In theory this ought not to be so, and in the course of the hearing I also suggested that to guard against this risk I could cap the budgeted costs at the level of the agreed costs of the unified hearing. Mr. Craig was willing to agree to this approach, but it then emerged that the claimant would wish to revise its budget in relation to any unified hearing in the light of my ruling on sampling. Its budget had apparently been prepared on the basis that I allowed its sampling application.

Second, it seems to me to be highly desirable for the parties to gain a clear

Third, the decision as to the appropriate financial remedies for the court to make (see Marathon Asset Management v Seddon and Others [2017] ICR 791 at paragraph 223). It seems to me that the overwhelming likelihood is that the only remedy available to the claimant in the event that liability is established will be negotiating damages and moral damages. There has been pre-action disclosure and affidavits have been sworn and there is currently no evidence or pleaded case that the defendants have taken any step since 13th June 2019 to establish a relevant algorithmic trading business. I appreciate that the claimant says that they made more use of the information than they admit prior to this date, but, of course, that does not deal with the position since that date and I also have to consider the management of the case as currently pleaded.

Fourth, although there is a range of possible outcomes on liability, and therefore in theory a range of possible results in terms of negotiating and moral damages, I note

Fifth, no doubt the experts will also evaluate the defendants' best case as to the nature of the information disclosed by Mr. Cologlu by reference to their witness evidence and the expert evidence of Dr. Auld, as well as evaluating the claimant’s best case if it is right about the issue of confidentiality and there is any basis for alleging that the activities of the defendants went beyond what they have admitted. These steps will provide additional useful parameters for the parties and the court. The experts may well address other scenarios. I therefore do not agree that expert accountancy evidence as part of a single trial would serve no useful purpose, particularly given that, as the claimant itself pleads in the schedule of loss, the approach to negotiating damages "is necessarily impressionistic".

Sixth, I do not accept there will be no prejudice to the defendants if Mr. Craig's approach is adopted. They may win outright but they may not, and, in this event, there seems to me to be a substantial risk that settlement is not possible on Mr. Craig's proposal and the claimant then pursues its claims to the next stage. The litigation would then be hanging over the defendants personally, as well as the GSA business, for a significantly longer period.

Seventh, as to the saving of time, the parties agree that unless the litigation is resolved in Part 1, it will necessarily take longer to complete the proceedings if there is a split hearing than if there is a unified one. I have listened to the detailed arguments on the likely timetable but there is a high degree of guesswork involved in predicting how long the stages of the litigation will take and what complications will arise along the way. Suffice it to say that I am not convinced that the case will be ready for trial on Part 1 in January 2022 and quantum could then be decided 12 months thereafter given the complexity and volume of the evidence, which both sides emphasise, and given the limited progress which the proceedings have made in the best part of a year since they were issued. It seems to me that there is a substantial risk that each stage of the litigation process will take longer than the claimant predicts. On the other hand, the parties agree, as I have noted, that a unified trial could take place from October 2022.

Eighth, the burden on the trial judge is a relevant consideration, as Mr. Craig points out, but I am not at all sure that the trial judge would thank me for splitting the trial. This would potentially tie their hands in relation to remedy. I agree with Mr Devonshire that the approach to negotiating and/or moral damages will not be scientific in the sense that a formula can be applied. The approach will be impressionistic as the claimant pleads and, in my view, the judge is likely to be able to come to a conclusion in Part 1 on what, if anything, to award even if the accountancy experts do not directly address the precise scenario which the judge finds on the evidence. It would therefore potentially be very unhelpful for the judge to be prevented from making a decision on damages and costs because the necessary evidence was not available. I also take into account the potential complications in relation to judicial availability and resources given that quantum would not be determined until 12 to 18 months after the trial on Part 1.

Ninth, the bifurcation of any appeal process is a concern in my view. Mr. Craig's proposal to stay any appeal on liability is a potentially treacherous course to take although, of course, he did not intend it to be. As Mr. Devonshire pointed out, in the event of a defeat for the claimant, there would be delay whilst any appeal ran its course, with the potential need for a re-run if the appeal succeeded. If the claimant won, but either party considered that the judge had got the findings on liability wrong (the claimant thought the judge did not go far enough or the defendants thought that the judge went too far) staying an appeal until the decision on quantum had been made would not be an attractive course. Such a stay would mean that the parties then incurred the costs of the quantum hearing whilst running the risk that the conclusion on appeal was that quantum had been decided on a false basis and the costs of Part 2 had been entirely wasted. In this event Parts 1 and 2 would also potentially have to be re-run.

So for all of these reasons, I decline to split the trial at this stage. The defendants' application for orders for further information

I was reminded of the well-known passages from the judgment of Lord Woolf MR in McPhilemy v Times Newspapers Limited and Others [1999] 3 All ER 885 at 792 and Hall v Sevalco Limited [1996] PIQR 344 at 349 which deprecate prolixity of pleading and pleading battles. They also emphasise that the approach should be for the pleadings to contain what is necessary and proportionate for the fair determination of the issues in the proceedings, and that disproportionate expense should be avoided.

The approach which I am required to take is also clearly set out in paragraph 1.2 of CPR Practice Direction 18 which states that requests for further information must be "reasonably necessary and proportionate to enable the party to prepare his own case or to understand the case which he has to meet".

I also agree with the commentary in Blackstone's Civil Practice 2020 paragraph 30.1 which states:

As far as claims for breach of confidence are concerned, I was reminded by Mr. Devonshire of the authorities which emphasise the need for particularity of pleading where the breach of confidence is alleged. He referred to John Zinc Company v Wilkinson [1973] RPC 317 and, in particular, Russell LJ's emphasis at page 724 on the need to plead "the matters that are said to be trade secrets or confidential matter and not simply areas in which or matters in respect of which it is said that the plaintiffs were, so to speak, proprietors of trade secrets and information was confidential".

Mr. Devonshire also referred to the following well-known passage from the judgment of Laddie J in Ocular Sciences Limited v Aspect Vision Care Limited [1997] RPC 289, 360 where he said at page 359:

Then at page 360:

These cases therefore emphasise the need for clarity as to what information is said to be confidential. But I do not accept that, as Mr Craig appeared to submit, they contain an exhaustive account of what degree of particularity is required in all breach of confidence cases – provided the items of alleged confidential information are specifically identified the claimant cannot be required to do more by way of pleading. Having considered these authorities and other authorities referred to by Mr. Craig, in my view the requirements which were reiterated by Laddie J should be regarded as an application of the generally applicable test now set out in paragraph 1.2 of Practice Direction 18 in the particular context of breach of confidence claims. Even in that context a claimant may therefore be required to do more where the application of that test to the particular circumstances of the case requires it.

These requests relate to a number of contentions in the particulars of claim that information provided by Mr. Cologlu to the defendants was "copied or derived from and/or were the claimant's confidential information related to the ABC Strategy and related strategies" (see paragraphs 28(2), 35, 44 and 45 of the particulars of claim). The defendants are not able to make any positive case that the information was not copied or derived from the claimant's confidential strategies but, as I have noted, they say that the sort of information provided by Mr. Cologlu was unremarkable to any informed industry insider and available from publicly-available sources. The defendants have therefore agreed to provide particulars, by reference to the items identified by the claimant in the Annex, of which items comprise the sort of information which they maintain is publicly available and where they such information can be found, and to do so by 22 January 2021.

In a request for information dated 19th March 2020 the defendants asked for further particulars of the claimant's case on copying, derivation and confidentiality. By way of example, Request 2 asks:

Is it alleged that the 16 October 2018 email contained the

The claimant's response has been to create the Annex to which I have referred. As I have said, this takes each document in turn and breaks it down line by line indicating by way of crosses in boxes which information is said to have been copied and/or derived and/or is confidential and/or confidential in aggregate. But they have declined to plead the information from which it is said that the information in the disputed documents is copied or derived so that there is, in effect, a pleaded comparison between source and extract.

After some exchanges with the court, ultimately I understood Mr. Devonshire to accept that this part of his application gives rise to an issue of principle, the determination of which will dispose of all of the items under this general heading and that that issue is whether, having particularised the information which the claimant says is copied, derived and/or confidential the claimant should also be required to say from what information belonging to the Claimant it was copied or derived.

I then heard argument from Mr. Devonshire as to why the question of principle should be answered in his clients’ favour. For reasons which will become apparent I need not summarise those arguments, but they are set out very clearly in his skeleton argument.

In his skeleton argument, Mr. Craig resisted this part of Mr Devonshire’s application with equal clarity and vigour. But in the light of provisional views expressed by the court early in his oral submissions, agreement under this heading has been reached, and I can therefore deal with the matter significantly more briefly.

As I have noted, paragraph 1.2 of Practice Direction 18 sets out the test. My provisional view was that proportionality in relation to this part of the Defendants’ application was not an issue given the size and alleged value of the Claim and given that the claimant, through Mr Weiner’s witness statement, rightly accepts that it will have to set out this aspect of its Claim in due course. As Mr Craig immediately accepted, the real issue was therefore as to when it should do so and in what form.

My provisional view was therefore that the defendants should make clear the basis of its copying and derivation case by the stage when disclosure was provided. This was because the defendants, their advisers and their experts need to understand this aspect of the case against them by that stage in order to process the disclosure efficiently. The claimants' disclosure is likely to involve a very substantial amount of material and that material will be complex. The defendants and their representatives and advisers should therefore know what they are looking for when they read and process it. This will also enable them potentially to narrow the issues and to take other decisions about the conduct of the litigation.

When I expressed these provisional views, as I have said, Mr. Craig very sensibly indicated it might be possible to reach agreement on this basis and Mr. Devonshire, equally sensibly, agreed. In the result, agreement has been reached and will no doubt be recorded in the draft directions but I understand that the agreement is that on inspection the claimant will re-serve the Annex with an additional column identifying, for each line entry, the disclosure identification reference and internal page number evidencing the relevant aspect of the ABC Strategy and/or related strategies.

The application under this heading relates to Requests 8, 10(5) and 11(2) to (4) of the defendants' request for further information dated 19th March 2020. The requests relate to allegations by the claimant that passages in the DDQ set out specified matters:

Mr. Devonshire submits that in accordance with the authorities on breach of confidence, referred to above, this information should be provided.

Mr. Craig disagrees. He submits that the Annex sets out in granular detail each of the items of information alleged to be confidential and that the defendants therefore clearly understand the case against them and/or are in a position to prepare their own case. He argued that there was no need for the claimant to be required, as he put it, to parse the documents further than they have already been parsed.

I accept Mr. Craig's submissions. I am satisfied that it is not necessary or proportionate for the claimant to be ordered to provide the further information referred to. The defendants’ requests appear to me to be rhetorical in nature and designed to show that the pleaded propositions are not supported by the disputed documents. Even if they are not rhetorical, the claimant's case is sufficiently clearly pleaded. The defendants, their legal teams and experts are well capable of reading the disputed documents and deciding whether they accept the claimant's contentions as to what those documents show and of preparing their case accordingly.

The requests under this heading relate to passages in the defendants' reply and defence to counterclaim. The defendants' request for further information is dated 8th July 2020.

Request 1 asks a series of questions about the identity of price dislocations which are as follows:

Mr Devonshire’s arguments in support of this part of his application were essentially the same as in relation to the items under the heading above and so is my answer. I accept Mr. Craig's submission that there is no need for further pleading in relation to

Request 2 relates to the apparent allegation of falsehood in the affidavits of Messrs. Ward, Muldrew, de Lioncourt and Skinner at paragraph 28(d) of the reply and defence to counterclaim which I have set out above. The claimant says that they are not obliged to identify the respects in which it contends that those affidavits are false. They say that they have pleaded their inferential case at paragraph 55 of the particulars of claim and elsewhere as far as they can at this stage. Their position is that in so far as their pleaded case overall is inconsistent with the contents of the affidavits of Messrs. Ward, Muldrew, de Lioncourt and Skinner, it follows that those affidavits are not accurate and that the claimant does not accept what is said in the relevant passages of those affidavits.

Mr. Craig also agrees that if there is anything further which is specific in nature -- for example, in relation to the claimant's inferential case -- that will have to be pleaded after disclosure, and he resists the application for an order under this heading on that basis. In the course of his submissions, he indicated that the inaccuracy and/or falsity which the claimant has in mind includes the defendants' assertion that the information provided by Mr. Cologlu was not confidential in nature and is in the public domain. His inferential case also, of course, suggests that the claimant will seek to persuade the court to infer that they made more use of the information than they admit.

Having considered Mr. Craig's arguments, I have concluded that the claimant should identify the paragraphs in the affidavits of the four individuals named, which it says are false or inaccurate and should briefly explain why by reference to its existing pleaded case or otherwise. So that this order should not be considered vague, I will explain the rationale. It does seem to me that potentially very significant reputational damage can be caused, particularly in this sector, by allegations of the sort under discussion under this heading being flung about. I am sure that that is not what the pleaders of the claimant’s case are doing, but it does seem to me that the risk of reputational damage will be minimised if, in what are ultimately public documents, the claimant is specific as to any points on which it says these individuals have given dishonest evidence and they are therefore very clear as to the case which they are required to address. Of course, the effect of so doing or being required so to do may be that some of the heat is taken out of the litigation. It may also be that some narrowing of the issues is feasible.

Request 7 seeks detailed information about the claimant's account of the negotiations with Mr. Taylor, which are the subject of the defendants' counterclaim. I have very little hesitation in rejecting this application. The claimant has provided an answer to the request and has indicated, at least in broad terms, if not more specific terms than that, its position in relation to this aspect of the defendants' pleaded case. It seems to me that the detail and the nuances of where the parties differ as to what was said and done is plainly a matter of evidence which can be provided in the course of trial preparation if, indeed, the defendants conclude that it is worthwhile to pursue this aspect of the case. So, in my view, this aspect of the claimant's case is adequately pleaded and it would be disproportionate to order the claimant to do more.

There is a large measure of agreement as to the terms of a confidentiality protocol which will apply to confidential information and evidence which is disclosed, referred to or created this these proceedings. That information has been divided into two categories. In broad terms category 1 "restricted information" is less sensitive, although it remains highly confidential, and the restrictions applicable to this category of information are therefore less stringent. Category 2 "restricted information" is subject to greater restrictions and is defined as follows: "5. Category 2 Restricted Information shall mean:

There is also a mechanism for resolving disputes as to what is or is not category 2 information: see clause 6.

The outstanding issues on the terms of the confidentiality protocol relate to some of the requirements in relation to category 2 restricted information which are sought by the claimant to be placed on the expert witnesses. The defendants object to the application of these particular requirements to their expert, Dr. Thomas Auld. The issues which are disputed, although many restrictions are agreed, are as follows:

In considering the competing arguments under these headings I have taken into account the obvious need to preserve the confidentiality of any confidential information and trade secrets which are the subject of this litigation. For present purposes, and without in any way binding the trial judge, I accept the evidence of Mr. Weiner as to the claimant’s concerns about the high degree of confidentiality of the information which will be shared in these proceedings, the very high commercial value which it has, the very high level of investment in that part of the claimant's business which there has been and the potential impact of misuse of that information given that the business model is built on identifying price dislocations which the market has not identified.

I also accept Mr. Weiner's evidence that the claimant has taken rigorous steps to protect the confidentiality of that information. Merely by way of example, Mr. Cologlu was only one of 15 individuals in the claimant's entire global business, comprising in the order of 3,000 employees, who had access to the information in question. There were then various other measures which were taken to ensure that there was no breach of confidence, whether deliberate or inadvertent, by those who had access to that information.

I have also considered the authorities to which I was referred, including the well-known judgment of Hamblen J in Libyan Investment Authority v Societe Generale [2015] EWHC 550. Of particular significance in the present case are the following passages:

I was also referred to the Roussel UCLAF v ICI [1990] RPC 45 case including the following passage:

In addition to that, I was reminded of paragraph 21 of the judgment in IPCOM GMBH v HIT Europe Co Ltd and others [2013] EWC 52 (Pat):

Mr. Devonshire also referred me to paragraphs 27 to 42 of the judgment of Roth J in Infederation Limited v Google LLC and Others [2020] EWHC 657 (Ch).

Finally, by way of legal materials, I was referred to the Trade Secrets (Enforcement etc.) Regulations 2018 and in particular Regulation 10, read in the context of recitals 24 and 25 of the Trade Secrets Directive (EU 2016/943).

Mr. Craig draws attention both to how little is known about Dr. Auld by the claimant or the court and to the following evidence as to what is known about Dr. Auld.

I emphasise, as Mr. Craig did, that none of these points is relied on to suggest any lack of integrity on the part of Dr. Auld and nor is there any such evidence before the court. The concern which Mr. Craig was seeking to illustrate was about the risk of inadvertent use or disclosure, or Dr. Auld being placed in an awkward situation in the context of future work, where temptation or conflicts of interest may arise.

More broadly I understood Mr. Craig to be making the point that given that the claimant goes to great lengths, in the context of its own workplace and its own workforce, to protect the confidentiality of the information for the reasons which I have summarised, it is unsurprising that it is unwilling for that information, or at least

Mr. Devonshire, for his part, draws attention to the practical impact of the disputed restrictions. He emphasises that the burden is on the party seeking to limit inspection by the other party to the litigation to demonstrate the necessity for the restrictions proposed. Mr. Devonshire also rightly emphasises that these sorts of measures are exceptional and should be limited to what is absolutely necessary to achieve fairness between the parties and in particular fairness in terms of protecting the confidentiality of the information which is involved in the litigation.

My conclusion on this issue is that the restrictions proposed by the claimant should be applied. I have outlined what is proposed, above. Importantly, it does not involve additional cost to the defendants. I understand and accept from Mr. Craig that the same approach is being applied to Professor Clifford, the claimant's expert. An issue was raised about the ability of Dr. Auld to communicate with the solicitors instructed on behalf of the defendants but, ultimately, I understood it to be agreed that that issue could easily be resolved.

Mr. Devonshire was to some extent facing a case that was moving towards him rather than away from him when the application relating to this issue was made. On the footing that ultimately what was being said was that Dr. Auld should work in a local solicitor's office using methods which are secure and leaving the documents in a secure environment, ultimately, I was not persuaded that there were any practical objections which outweighed the need for data security and warranted my rejecting the proposed restriction.

Ultimately, Mr. Devonshire's objection was that Dr. Auld might wish to work outside office hours and/or might have thoughts about the case in the bath or elsewhere but would not be able to write them down without creating category 2 restricted information and, as a result, being in breach of the protocol. These seemed to me to be issues that could readily be overcome from a practical point of view. For example, as Mr. Craig pointed out, in this type of event if he needed to, Dr. Auld could phone the solicitors instructed on behalf of the defendants and ask them to make a note or check a point. In so far as he wished to work out of hours, I anticipate that arrangements could be made. I also note in this connection that if real practical difficulties do arise, the confidentiality protocol provides for applications to vary.

As I have indicated, the issue under this heading was in broad terms as to a mechanism that would provide safeguards in the event that, quite properly, Dr. Auld had a change of heart or decided that he wished to move back into algorithmic trading or related work. Initially, the proposal was that he should undertake to notify the claimant of any intention to take up such an appointment. That proposal, following constructive discussions between the parties, was modified so that Dr. Auld agreed that he would be prepared to provide an undertaking to show the undertakings into which he had entered under the confidentiality protocol to any prospective work provider.

The issue therefore narrowed to the question for how long he should be under such an obligation. Mr. Devonshire contended that the period should be not longer than one year. The basis for this contention was that Mr. Cologlu, a central figure in relation to the relevant strategy was, under his contract of employment, subject to a notification requirement, in the event that he was to take up alternative and related work, which applied for a year. Accordingly, Mr. Devonshire proposed that the period should be one year from the last point at which Dr. Auld had any involvement in the case.

Mr. Craig contended for a period of five years. He said that this was on the basis that, of course, Dr. Auld is not and never has been an employee of the claimant. The claimant, therefore, has not had the opportunity to scrutinise him in the way that it would do a recruited employee or to vet that employee from the point of view of security; accordingly, the period should be five years. I pressed Mr. Craig to say whether there was any particular rationale for the five-year period, and there is no criticism of him, but he was not able to do so. He maintained that the information would remain confidential far beyond the five-year period so that the “shelf life” of the information was not ultimately part of the rationale for the five-year proposal.

Doing the best, I can, I have come to the conclusion that the undertaking should last for a period of two years from the last point at which Dr. Auld has any involvement in the case. I reach that view on a very broad-brush approach. Essentially it is my best guess as to how long, given the other safeguards in the confidentiality protocol which prevent any retention of the documents, notes etc, it can it reasonably be expected that Dr. Auld would retain important details of the relevant trading strategy in his head such that possession of that information would then potentially pose a threat in the event that he accepted work or related work from another work provider. So that is my ruling on Issue 2.

Issue 3, as I have said, is whether in addition to the undertakings to the court Dr. Auld should give undertakings to the claimant which would be contractual in effect. I have come to the conclusion that he should not be required to do this.

In my view, the undertakings to the court which Dr Auld will give and which are extensive and detailed and, more importantly, underpinned by the threat of criminal sanction in the event of breach, are sufficient to discourage him from any conduct which might give rise to the risk of the confidential information being leaked. I do not consider that contractual undertakings are necessary or proportionate in the circumstances of this case.

Mr. Craig asks the questions, "What difference does it make if he is going to comply with the undertakings in any event? If he is going to comply with the undertakings, then why does it matter? If he is going to breach them why should the claimant not have a claim in damages?". In my view, the answer is that the threat of criminal sanction does have the requisite deterrent or cautionary effect. The more serious the breach of the undertakings in the protocol and/or the consequences of such breach, the

In relation to Issue 4, the practicalities of communications between Dr. Auld and the defendants' legal team, as I understand it, those issues are resolved by my rulings on Issue 1, or at least will be capable of resolution by a sensible approach on all sides and by use of technology.

The confidentiality protocol will now be finalised in the light of my rulings and submitted for my approval in the usual way.

Given my ruling on the sampling approach, the parties sought permission to revise their budgets before submitting them for approval and I agreed with this course. An agreed direction on this question will be submitted for my approval.

These were agreed in the light of my decision, following a short adjournment. The proposed directions will be submitted for my approval in the usual way.