You only get a floating point result if your operands to the division operator are floats themselves. So:
See if that works. It forces the operation to use floating point numbers, which will result in a correct output. You see, when all is right, you get the quotient between two equal integers, which is 1. If AddCountRight decreases, it is always <1. Since the operation is based with integers, that value is truncated, so any 0.xxxxx value will become 0. If you force the operation to be float based, you'll get your results.
(((float)AddCountRight) / ((float)TotalAddProb) *100.0f)
 Actually, I am not sure if it is enough to have a float operand, so I always play safe and have both of them be treated as floats. Someone please correct me about this issue or advise a better technique.