In many real-world settings—such as policy-making, legal disputes, and lobbying—agents hold strongly polarized preferences, yet only a single outcome can ultimately be implemented. This creates a fundamental strategic choice: agents may seek a mutually acceptable compromise, or engage in competition, risking a worst-case outcome. We study how individuals make compromise–versus–competition decisions when paired with an agent holding an opposing preference. Compromise is implemented through a bargaining mechanism, while competition is modeled as a Tullock contest. If both agents choose to compromise, a guaranteed joint outcome is implemented; if either rejects compromise, the contest determines the final outcome. This structure yields a clear theoretical prediction: as risk aversion increases, compromise becomes relatively more attractive than competition. We build on Morgan, Orzen, and Sefton (2012) by replacing the outside option with (i) a fixed equal-split compromise in one treatment, and (ii) an endogenous bargaining process in another, where compromise materializes only when both agents reject competition. We address two main research questions: (1) when are individuals willing to forgo competition in favor of compromise, and (2) how do risk preferences shape this choice? Standard theory predicts similar contest entry across treatments; systematic deviations therefore indicate behavioral influences. Our results shed light on why polarized environments sometimes produce agreement—and why they often result in persistent gridlock.