In Rosgonki ANO v Formula One World Championship Ltd [2026] EWHC 2899 (Comm), the Commercial Court held that the defendant should in principle have security for costs against a Russian-resident claimant, but only for the costs of issues raised by the claim and not by its substantially overlapping counterclaim. The decision shows how the Crabtree principle operates where the overlap is substantial but not complete, and how an application should be framed.
Background
Rosgonki, resident in Russia, had been host and promoter of the Russian Grand Prix under a race promotion contract with F1. F1 served a force majeure notice and the 2022 event was cancelled. Rosgonki sued for repayment of about US$24 million paid towards the 2022 event fee, its prima facie entitlement being common ground [18].
F1 defended on sanctions and public policy grounds and, as a logically prior defence, contended that its termination of the contract meant no refund was due. It counterclaimed for the 2022 and 2023 event fees under clause 28.2(b) [21], a sum that dwarfed the claim [24].
F1 applied for security for costs under CPR 25.26. It was common ground that the CPR 25.27(b)(i) condition was satisfied, and the judge found a high risk that an English costs order would not be enforced in Russia [2], [3]. The corporate impecuniosity gateway was not passed [8].
Costs Issues Before the Court
The judge said that, if the proceedings comprised only Rosgonki’s claim, it would be a straightforward case for ordering security [8]. The real question was how the discretion should be exercised where a counterclaim overlapped almost entirely with F1’s logically primary defences [14]. Following the White Book, the judge asked what the substantial position of each party in the litigation was [15]. A further question was whether Rosgonki’s undertaking to allow F1’s costs to be paid from the Relevant Sums answered the application [25].
The Parties’ Positions
For Rosgonki, Mr Alliott opposed the application on the basis of the counterclaim [14]. He argued that Rosgonki’s undertaking should be regarded as appropriate and sufficient security [69], and that there was a cynical or at least tactical flavour to F1’s contractual case [71]. He also criticised F1’s quantification as not much better, and no better evidenced, than the sticking of a finger in the air [72].
F1 relied on a solicitors’ letter emphasising that it had never wanted to litigate [45]. It characterised its counterclaim as having independent vitality [65], and said the undertaking responded only if one of its sanctions or public policy defences succeeded [26].
The Court’s Decision
The judge started from the Crabtree principle, stated in the White Book at note 25.27.6: the court may refuse security for costs in respect of a claim where the same issues arise on a counterclaim in the same proceedings [27]. It derives from the Court of Appeal’s decision in BJ Crabtree (Insulation) Ltd v GPT Communications Systems Ltd [28].
The judge identified the primary and principled concern in Crabtree: security should not be ordered for a defendant’s costs that are at least as much costs of pursuing its counterclaim as costs of defending the claim [38]. Agreeing with Abbotswood, he accepted that where claim and counterclaim raise the same issues, that is likely to be a very significant, and in many cases probably determinative, factor, albeit that any other considerations will also have to be taken into account [43].
Had this been a full Crabtree case, it would have been on all fours with Anglo Irish Asset Finance plc v Flood and another, where the Court of Appeal held, obiter, that security had rightly been refused [40], [41]. The judge then set out seven propositions describing the parties’ substantial positions [59]-[67].
First, he assumed in F1’s favour that, as a matter of fact, the parties were always going to be this way round [59]. Second, in substance they would now more naturally be the other way round: the counterclaim was for a vastly greater sum, and F1 was the effective claimant even on Rosgonki’s claim [60]. Third, each claim went further than merely defending the other [61].
Fourth, this was not a full Crabtree case. The sanctions and public policy case at paras 50 to 86 of the amended defence was extensive and might well be expensive to litigate, but was substantially unnecessary for, and unrelated to, the pursuit of F1’s claim [62].
The List of Common Ground and Issues confirmed this. Issues 24 to 29 were the key issues raised by F1’s claim, so security covering them would be wrong in principle [63]. Issues 16 to 22 reflected the sanctions and public policy issues raised by Rosgonki’s claim but not by F1’s claim [64].
Fifth, the application should have the same outcome as if the parties were reversed. Andrew Baker J held that F1’s “independent vitality submission, therefore, had the wrong focus and so got things the wrong way round” [65]. Sixth, the right question was whether Rosgonki’s claim added materially to F1’s costs of defending F1’s claim, which the fourth proposition answered firmly in the affirmative [66].
Seventh, Rosgonki’s claim raised substantial factual and legal issues not the subject of F1’s claim. Other things being equal, security would be appropriate, but limited to the costs of those additional issues [67]. This applied the principle noted in Dumrul, citing Shaw-Lloyd and Newman [34], [68].
As the one-sided post-dismissal litigation problem referred to in Crabtree was not a possibility, a Dumrul undertaking could not take the matter anywhere [52]. On the basis on which security would be ordered, such an undertaking would be unnecessary, inappropriate and pointless [68].
Rosgonki’s undertaking was not a countervailing consideration. It gave no security if the US$24 million was unrefundably F1’s money; it was F1’s contractual defences, not the counterclaim, that deprived it of value [69], [70]. The judge proceeded on the basis that F1’s claim was raised in good faith on at least arguable grounds [71].
The just outcome was in principle that F1 should have security for its costs of Issues 16 to 22 [77]. The judge declined to work from F1’s overall figure, derived by stripping out costs referable to its own claim [72]-[74]. He identified the expert reports phase as one clear-cut qualifying item [75], and left quantum for further argument [77].
Practical Implications
This is a first instance decision, but it gives a structured route through applications for security for costs where a counterclaim is in play. Defendant applicants should map claim and counterclaim issue by issue, using the pleadings and any list of issues, and identify defences that do no work in the counterclaim.
Security for costs should then be sought for those issues, not for overall costs with counterclaim costs removed by estimate. The costs schedule should be built around the qualifying issues. Phases tied only to them, such as expert evidence confined to the non-overlapping defences, are the easiest to establish; rough apportionments of a wider total risk the court declining to work from the figure.
Claimants resisting should test whether the case is a full Crabtree case, and whether the defendant’s argument has the right focus: the relevant inquiry is whether the claim adds materially to the defendant’s costs of its own claim. An undertaking to pay costs from the sums claimed will not answer the application where the defendant’s defences, if well-founded, would leave those sums without value. A Dumrul undertaking adds nothing where the one-sided litigation risk does not arise.
▶ Watch the case summary
Security For Costs | Proportionality And Foreign Enforcement
Security For Costs In UK Civil Litigation
Related service: costs advocacy at detailed assessments and costs hearings.
