From the study of cluster distributions, we know that 0.01 is good enough cutoff for dt = 0.005. However, after calculating the average speed of blocks in microscopic events according to theoretical deduction of Carlson and Langer( close to n*\nu/(2*\iota^2) for an event with n blocks), it is order of 100 times smaller than the loading velocity \nu. If the loading velocity \nu is 0.01 used in my simulations, the average speed is of order 1E-4. So if I use 0.001 as velocity cutoff, I may underestimate the number of microscopic events by taking the moving blocks with speed less than 0.01 as stuck blocks. Table 1. shows distributions with different cutoff values for the same system as Table 2. It is surprising that 0.01 is reasonable for both dt =0.005 and dt =0.001.
| moment | energy released | lifetime | number of involved blocks | number of events | |
| dt = 0.005 | figure | figure | figure | figure | figure |
| dt = 0.001 | figure | figure | figure | figure | figure |