Using float or decimal will work ... but you need to do the floating point arithmetic by making at least one of the operands decimal / float / double:
decimal x = ((decimal) 204) / 5;
decimal y = ((decimal) 200) / 5;
if (x > y)
Or use floating point literals:
decimal x = 204m / 5;
decimal y = 200m / 5;
It doesn't matter which operand is a floating point:
decimal x = 204 / 5m;
decimal y = 200 / 5m;
float / double also works:
double x = 204d / 5;
double y = 200d / 5;
So what happens if you just use 204/5? Well, consider this statement:
double x = 204 / 5;
. , , . double x. , , , , .