Group Identification in N-Body Simulations: SKID and DENMAX Versus Friends-of-Friends

Three popular algorithms (FOF, DENMAX, and SKID) to identify halos in cosmological N-body simulations are compared with each other and with the predicted mass function from Press-Schechter theory. It is shown that the resulting distribution of halo masses strongly depends upon the choice of free parameters in the three algorithms, and therefore much care in their choice is needed. For many parameter values, DENMAX and SKID have the tendency to include in the halos particles at large distances from the halo center with low peculiar velocities. FOF does not suffer from this problem, and its mass distribution furthermore is reproduced well by the prediction from Press-Schechter theory.