Minimax regret

Minimax regret is a Game Theory decision rule for uncertainty: choose the option that minimizes your worst possible regret, compared with the best choice you could have made in each outcome.

What is minimax regret?

Minimax regret is a decision rule in Game Theory for situations where you do not know which state of nature will happen. Instead of asking, “Which choice gives me the highest payoff if I’m right?” it asks, “Which choice leaves me with the least regret if I’m wrong?”

Regret here means the gap between the payoff you get and the payoff you could have gotten from the best action in that same outcome. So for each possible state, you compare every option to the best option in that state, then record how much each choice falls short. The result is a regret table, which is usually built from a decision matrix.

Once you have those regret values, the minimax regret rule tells you to look at the worst regret for each option. Then you choose the option with the smallest of those worst regrets. That is why the method is called “minimax”: you minimize the maximum regret.

A quick way to see the difference is that this rule is not trying to maximize payoff directly. It is trying to avoid making a choice that you would kick yourself for later if a different state had happened. That makes it a conservative but not purely pessimistic strategy.

In a class problem, you might see a payoff table for a business choice, an investment, or a strategic move in a simple game. If one option does very well in one state but badly in another, minimax regret checks how bad you would feel missing the best possible result in each state. That makes it useful when you care about bad surprises, not just low payoffs.

Why minimax regret matters in Game Theory

Minimax regret matters in Game Theory because it gives you a way to make decisions when probabilities are missing or unreliable. In uncertainty problems, you cannot always justify expected utility, so you need another rule for comparing choices. Minimax regret is one of the main alternatives, along with maximin and maximax.

This term also shows a different way to think about rational choice. Instead of focusing only on gains or losses, it focuses on comparison with the best possible action after the outcome is known. That makes it a nice bridge to regret theory and loss aversion, where the feeling of having chosen “the wrong” option changes behavior.

The concept also shows up in how you read payoff tables. If you can turn a decision matrix into a regret matrix, you can see whether a choice is fragile, safe, or exposed to big disappointment in one scenario. That is a useful skill in problem sets because the method is mechanical once you know the steps.

Minimax regret also helps explain why real decision-makers sometimes avoid bold moves even when the average payoff looks good. The rule models caution in a very specific way: not fear of every loss, but fear of choosing an option that looks much worse than the best alternative after the facts are known.

How minimax regret connects across the course

Regret Theory

Regret theory is the broader idea behind the feeling that drives minimax regret. Instead of treating choices as only payoffs and probabilities, it looks at how people compare what happened with what could have happened. In Game Theory, that comparison helps explain why some decision rules avoid choices that might later feel like a mistake.

Decision Tree

A decision tree can organize the choices and possible outcomes before you compute regret. It is especially useful when a problem has several stages or branches, because you can map the states first and then build a regret table from those outcomes. The tree shows structure, while minimax regret shows how to choose.

Expected Utility Theory

Expected utility theory chooses the option with the best weighted average outcome when you know or can estimate probabilities. Minimax regret is different because it does not need probabilities. If a problem gives you no reliable likelihoods, the two methods can lead to different answers and different strategic interpretations.

maximin criterion

Maximin looks at the worst payoff for each option and picks the choice with the best worst case. Minimax regret also looks at a worst case, but the worst case is regret, not payoff. That means maximin is about protecting against low outcomes, while minimax regret is about avoiding the choice you would most wish you had changed.

Is minimax regret on the Game Theory exam?

A problem set or quiz question usually gives you a payoff matrix and asks you to apply the rule step by step. First, find the best payoff in each state, then subtract every other payoff from that best value to build the regret table. Next, identify each option’s maximum regret and pick the option with the smallest maximum.

If the course uses written responses, you may also need to explain why minimax regret fits an uncertainty setting better than expected utility. In discussion or short-answer work, a good answer often compares it with maximin, since both are cautious but they measure risk differently. Watch for questions that hide the real task in a scenario about investments, pricing, or strategy, because the method is the same even when the story changes.

Minimax regret vs maximin criterion

These two are easy to mix up because both are cautious decision rules under uncertainty. Maximin focuses on the worst payoff each option could give you, then chooses the best of those worst payoffs. Minimax regret focuses on how badly you would miss the best possible outcome in each state, so it compares disappointment rather than raw payoff.

Key things to remember about minimax regret

• Minimax regret is a Game Theory rule for choosing under uncertainty when you do not know the probabilities of outcomes.

• It works by building a regret table, where each payoff is compared to the best payoff available in that same state.

• The rule chooses the option with the smallest maximum regret, not the highest expected payoff.

• This method is cautious, but it is about avoiding future disappointment rather than simply avoiding low outcomes.

• If a problem has a payoff matrix and no probabilities, minimax regret is one of the first decision rules to try.

Frequently asked questions about minimax regret

What is minimax regret in Game Theory?

Minimax regret is a decision rule for uncertainty where you choose the option with the smallest possible worst-case regret. Regret means the payoff you gave up by not choosing the best action for that outcome. In Game Theory, it is used when probabilities are unclear or unavailable.

How do you calculate minimax regret?

Start with a decision matrix and identify the best payoff in each state of nature. Subtract each option’s payoff from that best payoff to get regret values. Then find the maximum regret for each option and choose the option with the smallest of those maximum regrets.

How is minimax regret different from maximin?

Maximin looks at the worst payoff for each option and chooses the option with the best worst payoff. Minimax regret looks at how far each option falls short of the best possible payoff in each state. So maximin protects against bad results, while minimax regret protects against the feeling that you picked the wrong option.

When would you use minimax regret instead of expected utility theory?

Use minimax regret when you do not have trustworthy probabilities for the possible outcomes. Expected utility theory needs probabilities to weight outcomes, but minimax regret works without them. That makes it useful in uncertainty problems where you want a cautious choice based on possible disappointment rather than average value.