Accurate estimation of the transport probabilities among regions in the ocean provides valuable information for understanding plankton transport, the spread of pollutants, and the movement of water masses. Individual-based particle-tracking models simulate a large ensemble of Lagrangian particles and are a common method to estimate these transport probabilities. Simulating a large ensemble of Lagrangian particles is computationally expensive, and appropriately allocating resources can reduce the cost of this method. Two universal questions in the design of studies that use Lagrangian particle tracking are how many particles to release and how to distribute particle releases. A method is presented for tailoring the number and the release location of particles to most effectively achieve the objectives of a study. The method detailed here is a sequential analysis procedure that seeks to minimize the number of particles that are required to satisfy a predefined metric of result quality. The study assesses the result quality as the precision of the estimates for the elements of a transport matrix and also describes how the method may be extended for use with other metrics. Applying this methodology to both a theoretical system and a particle transport model of the Gulf of Maine results in more precise estimates of the transport probabilities with fewer particles than from uniformly or randomly distributing particle releases. The application of this method can help reduce the cost of and increase the robustness of results from studies that use Lagrangian particles.