A high-low agreement is a contract both sides sign before a verdict comes in, setting a floor the plaintiff is guaranteed to receive regardless of the jury's decision and a ceiling the defendant will never have to pay past, regardless of how large the verdict turns out to be. If the jury returns a number inside that corridor, the actual verdict controls; if it falls outside the corridor in either direction, the agreed floor or ceiling applies instead. It is a way to keep the upside of a jury trial while removing the tails of the outcome distribution that make trial frightening for both sides.

Negotiating one well requires the same expected-value and variance thinking that goes into any settlement decision, just applied to a narrower question: how much variance is each side willing to pay to remove, and where should the floor and ceiling actually sit given the real range of plausible verdicts. Modeling the underlying verdict distribution before you propose numbers, rather than picking a floor and ceiling that simply feel reasonable, is what keeps a high-low agreement from quietly transferring more risk to one side than either party realizes they agreed to.

A high-low agreement is a private, binding contract, typically kept from the jury so it does not influence deliberation, that both sides sign before the verdict is returned. It does not settle the case outright; the trial proceeds normally, evidence is presented, and the jury still reaches its own verdict. What changes is what happens to that verdict once it is in: any amount above the agreed ceiling is reduced to the ceiling, any amount below the agreed floor, including a defense verdict of zero, is raised to the floor, and any amount in between is paid exactly as the jury decided.

A high-low agreement earns its cost, meaning the certainty it removes from both sides in exchange for protection from the tails, when the range of plausible verdicts is genuinely wide. It fits poorly in cases where the outcome is already fairly predictable, since there is little tail risk left to insure against, and the negotiation over floor and ceiling numbers can consume time better spent elsewhere. It fits well where liability is reasonably clear but damages could land almost anywhere, or where one side is exposed to an unusually severe downside it wants to cap regardless of what the jury ultimately decides.

The floor should track the plaintiff's BATNA, not an aspirational number pulled from the best-case scenario. If the realistic alternative to trial is a modest but certain settlement, the floor should sit at or above that figure, since agreeing to a high-low with a floor below your actual walk-away number gives up the protection the agreement is supposed to provide. Calculate the floor the same way you would evaluate any settlement offer, using expected value net of the cost of continuing to litigate, rather than treating it as a separate, looser negotiation.

The ceiling matters most to the defense, and it typically tracks available insurance coverage or the point past which a verdict would create exposure the defendant genuinely cannot absorb. Insurers and defense counsel frequently set the ceiling at or near policy limits specifically to eliminate the risk of a verdict that exceeds coverage and creates personal or bad-faith exposure beyond the policy. Confirm the actual limits and any excess coverage before proposing a number, since a ceiling set carelessly above real exposure defeats the purpose of the agreement for the side that needed the protection most.

Before signing, run the case's likely verdict outcomes as a distribution rather than a guess, then overlay the proposed floor and ceiling on top of it to see how much of the realistic outcome range actually falls inside the corridor versus how much gets clipped at either end. A high-low that clips a meaningful share of the upside for a small reduction in downside risk may not be worth signing for the side giving up more than it is getting back. This is exactly the kind of comparison a simulated outcome distribution is built to support before commitments are made.

Draft the agreement to specify exactly which verdict components are covered, how and when payment is due once the verdict is known, whether either side retains any right to appeal, and whether the agreement's existence and terms remain confidential from the jury and, in some cases, the public record. Ambiguity in a high-low agreement defeats its purpose, since the entire point is to remove uncertainty, and a poorly drafted corridor can create a new dispute about what the agreement actually covers right when the verdict comes in.