We examine the relationship between private capabilities and war in a new class of crisis bargaining games. Whereas traditional models consider interactions that end in either war or a peaceful bargain, we assume actors can also engage in low-level, costly policy options that shape final political outcomes in their favor (e.g., sanctions, airstrikes, cyberattacks). We analyze this new class of models using the tools of mechanism design. In contrast with standard monotonicity results in crisis bargaining models, we identify general conditions under which a greater private willingness to fight is associated with a lower probability of war or expected utility from settlement. Theoretically, our results identify when including multiple conflict options breaks key tenants regarding the role of private information in conflict. Substantively, our results suggest that improving specific conflict capabilities can perversely lead to more war or worse outcomes for the state making the improvements.