The problem of determining effective allocation schemes of underwater sensors for surveillance, search, detection, and tracking purposes is a fundamental research area in military operations research. Among the various sensor types, multistatic sonobuoy systems are a promising development in submerged target detection systems. These systems consist of sources (active sensors) and receivers (passive sensors), which need not be collocated. A multistatic sonobuoy system consisting of a single source and receiver is called a bistatic system. The sensing zone of this fundamental system is defined by Cassini ovals. The unique properties and unusual geometrical profile of these ovals distinguish the bistatic sensor allocation problem from conventional sonar placement problems. This study is aimed at supporting decision makers in making the best use of bistatic sonobuoys to detect stationary and mobile targets transiting through an area of interest. We use integral geometry and geometric probability concepts to derive analytic expressions for the optimal source and receiver separation distances to maximize the detection probability of a submerged target. We corroborate our analytic results using Monte Carlo simulation. Our approach constitutes a valuable "back of the envelope" method for the important and difficult problem of analyzing bistatic sonar performance.