[awm]

It's worth remembering that a lot of the time, your score will not be much different if you open 1NT or 1M on these hands. And when the score is different, it is often for a somewhat random reason that could go either way (which defender has the right "natural" lead against 3NT, or how well the major suit happens to break effecting whether 4M or 3NT is the better game).

My guess is that the difference is at most a few percentage points either way. So if Cascade's simulation is viewed as indicating that having a 2-card holding in the other major makes opening 1NT lose roughly 0.6% of boards, this might actually be significant!

And his assumption that you "always find all major fits on game hands" is not necessarily realistic: most people transfer and then bid 3NT on 53(32) game forces opposite a 1NT opening and a 5-3 heart fit could obviously be lost in this way. Certainly you could change your methods so as to find all such fits, but this would entail other losses because you lose some sequences that might be put to better use if you were not obsessing about finding all major suit fits opposite a 1NT opening that includes a five-card major.