In a Mary Carter agreement, a plaintiff settles with one defendant in a multi-defendant case, but that defendant remains a nominal party at trial, often aligned in interest with the plaintiff against the remaining, non-settling defendants. The settling defendant's actual payment obligation is typically reduced, capped, or partly refunded depending on how much the plaintiff recovers from the co-defendants, creating a financial incentive for the settling defendant to help the plaintiff win against the others.

These agreements are controversial because they can distort a trial's adversarial structure: a defendant who appears to be defending itself may actually be financially motivated to assist the plaintiff. Many jurisdictions require disclosure of the agreement to the jury, restrict its use, or prohibit it outright, while others permit it subject to careful management by the trial court.

Juricratic models a Mary Carter-style arrangement as a conditional payoff function tied to the outcome against the remaining defendants, which is a natural fit for the engine's game-theoretic framing: it changes the settling party's incentive structure and payoff curve, not just its dollar exposure, and that shift is exactly the kind of strategic realignment a multi-party simulation is built to surface.

Plaintiffs' counsel in multi-defendant cases consider a Mary Carter-style settlement to secure a guaranteed recovery from one defendant while preserving a cooperative witness against the remaining defendants, while defense counsel for non-settling parties move to have any such agreement disclosed to the jury or excluded, since it can affect how the jury perceives the settling defendant's testimony.