A procedure is described for calculating two statistics, the mean and the coefficient of variation, for largest live load effects in a bridge structure caused by highway traffic during a reference period. These statistics are pertinent in the selection of highway live loads and calibration of a statistically based design code. The reference period depends on the limit state being calibrated.Using weights and dimensions of trucks from a 1982 survey in Ontario, congested traffic moving through one bridge length is repeatedly simulated for various “truck densities.” Herein truck density is the number of trucks on a length of highway divided by the maximum number that could occupy the length. When simulating traffic of one density crossing one influence line, the largest structural effect per traffic move has a frequency distribution that, with reasonable accuracy, is idealized as Gaussian.Observed multiple truck presence is modelled by a modified Poisson distribution that has three parameters. For one set of these traffic parameters, mean largest structural effects are shown to vary in a consistent manner with truck density, truck volume, and one parameter characterizing the tributary area of the effect. Based on these observations, more concise procedures for simulation and calculation are recommended. Key words: statistics, probability, truck survey, bridge loads, structural effects, bridge design, code calibration.