Again, simple questions for everyone: what would prove the quitting-while-ahead theory wrong? And what is wrong with running simulations to test it?