516-227-0700

First Department Protective of High-Frequency Trading Algorithm: Considerations in Handling Discovery Requests for Propriety Code, Software, and Algorithms

January 30, 2020

As we continue to see increased litigation over electronic programs, apps, and algorithms, courts are increasingly called to consider discovery requests for the coding behind that technology.  These requests highlight the tension between the need for broad discovery and the litigant’s proprietary interest in secret, commercially valuable source code.  And as a recent First Department decision highlights, Courts are acutely protective of this source code.

The First Department, in BEC Capital v. Bistrovic, recently reversed the trial court’s ordering a coder to either produce his trading algorithm subject to the Commercial Division’s standard confidentiality order or abandon his claims, holding instead that the trial court should have ordered the algorithm produced “for Attorneys and Experts’ Eyes Only.”

The discovery dispute in BEC Capital arose from a failed high-frequency trading joint venture.  The defendant Bojan Bistrovic, through his company MCM, entered into an agreement with Plaintiff BEC where Bistrovic would integrate his proprietary trading algorithm into BEC’s trading platform, and he and BEC would share in the gains or losses of Bistrovic’s algorithm.  Bistrovic’s trading algorithm was built for speed: executing trades in a fraction of second to exploit fleeting market trends and inefficiencies.  Consequently, the algorithm required an efficient trading platform—every process, line of code, or mile of cable that a trade order had to traverse increased the time between when Bistrovic’s algorithm directed the trade and the time it was executed, and that lag gutted the efficiency of Bistrovic’s algorithm.

Less than six months into their agreement, the parties’ venture had performed so poorly that both effectively abandoned the agreement.  BEC pinned the poor performance on Bistrovic’s algorithm; Bistrovic blamed serious problems in BEC Capital’s trading platform.  When a dispute arose about the allocation of trading losses, Bistrovic allegedly told others in the high-frequency trading industry that Plaintiffs were fraudsters who had stolen money from him.

When BEC and its principals sued for defamation and breach of their NDA, Bistrovic brought counterclaims for breach of contract, alleging that BEC breached their agreement because they failed to provide Bistrovic with a satisfactory high-frequency trading platform into which his algorithm could have been properly integrated.

In light of Bistrovic’s counterclaims, Plaintiffs demanded the coding behind Bistrovic’s high-frequency trading algorithm in discovery.  The coding was necessary, Plaintiffs argued, to rebut Bistrovic’s counterclaims that the failure resulted from BEC’s platform—i.e., to show that the poor performance resulted from Bistrovic’s algorithm itself, not BEC’s platform.  When Bistrovic objected to the discovery of his trading algorithm, the trial court (Ramos, J.) held that Bistrovic must either (i) disclose the algorithm subject to the standard Commercial Division confidentiality order, or (ii) face the risk of having his counterclaims and defenses based on his algorithm stricken.

In November, the First Department reversed the trial court, holding that:

The production of defendants’ source code, which is a trade secret . . . should have been ordered to be produced for ‘attorneys and expert eyes only.’  Plaintiffs’ assertion that they have the expertise to review and opine on the source code and should not be subjected to retaining an expert, does not support unfettered access to defendants’ confidential algorithm.

The First Department’s ruling constitutes a rare reversal of the discretion afforded to the trial courts to oversee the discovery process.  See Don Buchwald & Assoc. v Marber.

In so holding, the First Department continues to show its interest in ensuring that confidential source code be produced according to appropriate terms and limitations, including, where necessary, an “attorneys and experts’ eyes only” designation.  See MSCI Inc. v. Jacob (reversing trial courts’ denial of discovery into confidential source code and ordering production “for attorneys’ eyes only”).  More generally, the First Department remains actively protective over confidential source code and—in this area more than others—willing to substitute its own discretion for the trial court’s, ensuring the deliberate development of legal authority over issues relating to confidential source code.  See also, e.g., People v. Aleynikov (reinstating the tossed conviction of the now-infamous Goldman Sachs programmer Sergey Aleynikov for uploading portions of Goldman’s high-frequency trading code to a German code repository).

Practical Considerations

BEC Capital provides some welcome guidance on how the First Department views the interplay between proprietary computer code and the need for “open and full disclosure as a matter of policy.”  MSCI.  Going forward, litigants can expect trial courts to take a thorough and critical look at requests for discovery into proprietary source code, including analysis of the following considerations:

  • How central is the source code to the claims? Not all requests for disclosure of proprietary code are created equal, and, of course, the more central the disputed source code is to the issues in the case, the more compelling the argument for disclosure.  For instance, a copyright case concerning the disputed code (where the party asserting the claim must prove the originality of the work) might favor disclosure, see Fonar Corp. v. Magnetic Resonance Plus, Inc., 93-cv-2220, 1997 WL 689462 (S.D.N.Y. Nov. 3, 1997), but only where the claim directly concerns the specific code requested, see Abarca Health, LLC v. PharmPix Corp. (denying discovery of portions of source code that were not at issue in infringement claims).  Likewise, a federal court has granted discovery of a carmaker’s proprietary code in a products liability action alleging harm directly caused by the code, Burnett v. Ford Motor Co., while another has denied discovery into proprietary source code in a false advertising case where plaintiffs claimed that the code created the misleading content, Congoo, LLC v. Revcontent LLC.
  • Is the code needed offensively or defensively? In Viacom Int’l Inc. v. YouTube Inc., the Court denied plaintiff’s request for discovery of defendants’ source code based in part upon the fact that Viacom sought YouTube’s source code offensively—to support its own claims.  The Court observed that defendant “should not be made to place this vital asset in hazard merely to allay speculation.”
  • What are the parties up to? Where the parties are in the same industry—such that the disclosure of the source code even subject to strict confidentiality restrictions may result in a competitive disadvantage—the case for non-disclosure or an “attorneys and experts’ eyes only” designation is stronger.  See MSCI (attorneys and experts’ eyes only designation appropriate where employee left plaintiff company to build a competing platform for defendant-company); ABC Rug & Carpet Cleaning Serv. Inc. v. ABC Rug Cleaners, Inc. (“Ample precedent exists for limiting disclosure of . . . proprietary information to attorneys and experts, particularly when there is some risk that a party might use the information . . . to gain a competitive advantage over the producing party.”).
  • Have other creative disclosure frameworks been considered? In addition to an “attorneys eyes only” designation, litigants should consider whether some other discovery framework is appropriate.  See RGIS, LLC v. A.S.T., Inc. (appointing special master to review confidential source code); Princeton Mgt. Corp. v. Assimakopoulos, 1992 WL 84552 (S.D.N.Y. April 10, 1992) (limiting disclosure to two designated individuals within plaintiff’s organization).
  • Is the code uniquely ill-suited to an “attorneys’ eyes only” designation? While the First Department in BEC Capital rejected the plaintiffs’ argument that BEC (and not its experts) needed to view the source code because they had the expertise to review it, other courts have favored this argument in holding that an “attorneys’ eyes only” designation was inappropriate.  See Metropolitan Life Ins. Co. v. Bancorp. Servs. LLC, 2000 WL 1644488 (S.D.N.Y. Nov. 2, 2000).  It remains to be seen how the First Department would view a case where the code sought is so particularized that an “attorneys eyes only” designation would be effectively useless.

While courts will continue to consider the discoverability of proprietary source code on a case-by-case basis (and subject to their broad discretion to oversee discovery), recent First Department guidance suggests that litigants would be well-advised to prepare for a deeply thorough, fact-specific inquiry focused not only on the necessity of the source code to the claims or defenses, but also on the commercial value of the source code and the appropriateness of alternative discovery frameworks.