It now seems I won't have time to write the long strategy post even sometime this week, so I'll just give some cliffnotes here (and we can discuss this more when I make the complete write-up):
Players don't treat their range as continuous. I call it "chunking." What it comes down to is that their calling range is essentially unaffected within certain bet size intervals. In order for $7 to be a better size than $10 on this flop, it has to be true that his range is a lot wider when we bet the smaller amount. I don't think his range for calling $7 vs. $10 changes hardly at all, and certainly not enough to make the $10 bet worse.
The reason for that has to do with the texture; there are few weak hands with which he might consider peeling for a small bet but will fold to a bigger bet. This is actually true most of the time on flops, which is why I think that, in general, a lot of people would win quite a bit more if they just bet more on the flop for value. For similar reasons, there's no reason to bet $10 when you're bluffing
since his calling range is the same as it is for $7, but you get a better price for it.
Obviously, this is exploitable, if you come across an opponent observant enough to notice your different bet sizes on the flop with different hands. If you're worried about this, ask yourself if you consider yourself an observant opponent. Next, ask yourself how often you take notes on your opponent's flop c-betsizes.
HUDs show your flop betting frequencies, not your sizes. Once you start playing a lot, lot, lot with good regulars, I think playing an exploitable but profitable strategy of "bet slightly bigger for value, slightly less for bluffs" is greatly +EV.