The fault plane of an earthquake can be estimated in many different ways. One is to examine the distribution of a well located aftershock sequence to see if a fault plane is evident. Here we present a new method where we take an earthquake aftershock sequence with distinguishable off fault clusters and vary the slip distribution to achieve a good correlation between regions of increased Coulomb failure stress (CFS) and aftershock occurrence. We find with the Lake Tennyson earthquake that simply taking an area outlined by aftershocks to estimate a fault plane may not be valid. If we do this, patterns of increased CFS have negative correlation with off-fault aftershock occurrence. However if we restrict slip to only deeper regions, and over a smaller area, then the correlation between regions of increased CFS and aftershock off-fault occurrence is good. ¿ 1999 American Geophysical Union