Marble Game Payoff Analysis

This brain teaser presents a simple two-player, simultaneous-move game where each player chooses between two colored marbles, and the payoff to one player depends on the combination of revealed colors. The twist is that the payoffs come partly from an outside source and are not strictly transferred between the players, so the game is not purely zero-sum. Candidates are asked to judge whether either player is structurally favored or whether the setup is fair, based only on the strategic symmetry and payoff structure, not on any iterated or adaptive play.

Solving it leans on building and interpreting a payoff matrix, reasoning about symmetry, and comparing expected values under different strategy choices. Interviewers watch whether the candidate can abstract away from the narrative, formalize the situation into a simple game, and then argue clearly about fairness or advantage without getting lost in individual outcomes. It also reveals how comfortably they handle simultaneous decisions, non–zero-sum payoffs, and the distinction between intuitive "looks fair" reasoning and a more systematic expected-value analysis.

What it tests

In simultaneous-move games with symmetric options and payoffs determined by both players' choices, the core principle is to analyze the expected value of each player's payoff under optimal or random strategies. The structure is governed by the payoff matrix, which encodes all possible outcome combinations and their associated rewards or penalties. Fairness or advantage is determined by whether the expected values (averaged over all possible strategy profiles) are equal or skewed. The reason this works is that, in the absence of exploitable asymmetries or dominant strategies, rational players' choices average out, and the game reduces to its statistical properties—mean and variance—rather than any single outcome. This approach generalizes to any game where outcomes depend on both players' simultaneous, independent choices and the payoffs are not strictly zero-sum.