I have 2 scripts, let's call it ScriptA.m and ScriptB.m. ScriptB.m is supposed to be the same as ScriptA.m except with improvement on the algorithm for speed. Both scripts make use of a lot of random numbers, but at different places due to the speed improvement of ScriptB.m. I want to make sure that statistically the two scripts produce the similar result, so I run each one of them 5x using matlab -batch 'run xxx.m', and look at the variation from run-to-run and script-to-script. I noticed that the output of each script is exactly identical across the 5x runs, which is a little strange to me. Does this mean Matlab initialize the random seed for each -batch run? Note that I do not use the function rng in my script for seed initialization. Thanks.