It seems that when using multiple cores for WFO training, Zorro 1832 does not close other instances(cores) automatically as older versions do (i.e. 1748). So when the TESTNOW flag is set, strategy will not be tested after training, since training is not "stopped" (training is in fact stopped where par files seem created properly, but Zorro is stuck). I do not know if anyone else is experiencing the same issue.