Still, the comparison ratio == threshold has very little chance to turn out true.
When you want to check whether A >= B (where A and B are numbers represented as floats), you would probably want to use the check A >= B - epsilon.
When you want to check whether A >= B (where A and B are numbers represented as floats), you would probably want to use the check A >= B - epsilon.
So I'm not sure you can argue against a piece of code on the basis of likelihood.
A relevant question is what epsilon should be.