What is the absolute minimum number of physical trials that would need to be conducted in order to validate that a game which has a Supposed negative expectation of 1/2 % per trial really does have that expectation in fact, with a 97% degree of certainty.