Until recently system error has usually been measured in feet or meters and stated in terms of average errors or percentages of position estimations which are within a certain distance of the true position. Systems such as FLAIR pose new problems, however, in analyzing, modeling, and interpreting system errors. These systems use an in-car odometer and compass to provide a crude form of inertial guidance. The somewhat noisy information from these instruments is transmitted at one-second intervals in the FLAIR system to a central receiver. The computer may correct the estimated location back to the most likely position whenever the estimated position is infeasible. Sources of error include random error, systematic error which creates a bias in the odometer readings, the street patterns of the city in question, the number of binary digits used to transmit information on vehicular heading and distance, and the sampling interval. Open loop tracking and system subvertability also have a bearing on system performance. Diagrams and tables are included. (Author summary modified)