A simplified study of the complex structure of arms races in the real world by the assumption of two contenders, each having one policy variable: weapons stocks. With such simplification, arms races are studied by means of reaction curves that indicate optimal weapons stocks, given the weapons stocks of the opponent. The ratio goal and Richardson models are formulated in terms of reaction curves. Allowance is made for the neglect of strategy and constraints by these models. The analysis is suggested only and does not consider the many dimensions of policy.