A beef cattle herd simulator was designed and developed as a tool for testing the consequences of farm management practices on herd performance. This discrete-event simulator is stochastic and individually based. It integrates a detailed reproduction model together with parameter-settable practices. This article details the various core components of the simulator. We then used it on virtual experiments with a herd of 70 cows. The first virtual experiment compared an autumn calving system with a herd management system based on winter calving. Adapting the herd management rules made it possible for the two systems to give similar outputs. In order to show the importance of the number of mated cows and of the reproductive period length on the performance of the herd, two different strategies were tested in with two new virtual experiments. The first one consisted in reducing the number of mated cows while the second one deals with reducing the reproduction period when the bull is present. The first of these scenarios led to drop in herd size in proportion with the female reduction over the years, whereas under the second scenario, a gap is simulated and the herd was wiped out after 10 years. We also focused an in-depth analysis on the variability of the results. The model shows that an average result not only masks variations from replication to replication but also masks variations in performances from year to year.