Minimax regret is a decision criterion used when a party wants to avoid the worst possible sense of having made the wrong call, rather than purely maximizing the mathematically expected outcome. For each available choice, regret is calculated as the difference between the outcome actually achieved and the best outcome that could have been achieved under the same circumstances with a different choice; the minimax regret rule picks whichever option minimizes the largest possible regret across all scenarios.

This approach differs meaningfully from a pure expected-value decision rule, because it can lead a party to reject the option with the highest average payoff if that option also carries a scenario with an especially painful worst-case regret. It is particularly relevant for risk-averse or reputation-sensitive parties, such as public companies or government entities, who weigh the optics of a clearly avoidable bad outcome more heavily than the raw expected math would suggest.

Juricratic reports a minimax-regret view alongside its default expected-value and maximin recommendations in the decision-walk analysis, letting a user see whether the strategy that looks best on average also happens to carry an outsized worst-case regret, which is often exactly the information a risk-conscious client needs before committing to a recommended move.

Litigation strategists use minimax regret analysis for high-visibility or reputation-sensitive decisions, where a client cares less about the average outcome across many hypothetical repeats of the case and more about avoiding a specific, foreseeable, and avoidable bad result that would be hard to justify after the fact.