Non-Markov Policies to Reduce Sequential Failures in Robot Bin Picking