However, if one parent is only heterozygous dominant white and not recessive white, I/i+ C+/C+, and the other parent is recessive white without dominant white, i+/i+ c/c, then about half of the offspring will still be non-white, i+/i+ C+/c. It does not require both parents to be hetero dominant white, in other words.
All we know at the moment is that your male has leakage, so must be dominant white instead of recessive. Your hen didn't seem to have any, so most likely is recessive white. Either or both could carry blue, though it sounded like the male had stray black feathers and not blue ones so it sounds like the blue dilution is most likely coming from the hen. So on the surface, this looks like (and thus far is correct for) hetero dominant white crossed to recessive white.
With a small sample size (it was 4 chicks from your whites, only 2 of which were thought to be from the frizzle hen, correct?), there is often a skewing of the results; this is why statistical analyses generally are not considered significant unless a certain threshold of samples is entered into them. Think of flipping a coin two times. Pretty often, you'll get the same result both times, right? But that doesn't mean that your coin only ever lands on that one side every time. The more times you flip that coin, the more equal your results become, because both sides have a roughly equal chance of coming up in each flip. That's the same in this case with your birds. You've 'flipped' the dominant white 'coin' twice and gotten not-white both times with your frizzle hen's chicks, but if you 'flip' it again, you might just get white. And if you 'flip' it a dozen more times, you might get a little of both!