I've got to wonder about simulations and percentage computations around a given hand playing 10 other "random hands." I don't think the numbers are of any particular value, because THE CARDS ARE NOT PLAYED IN A STATISTICAL VOID.
Here's what I mean.
A-A (red) can raise to 4 times the BB and get called by two players (neither of them the blinds).
Flop comes Q-10-7 with 2 spades and a club.
A-A bets out 3/4 of the pot (trying to make the pot odds wrong for a flush draw). One opponent folds, the other calls.
Turn comes 8 of clubs.
A-A bets out 3/4 of the pot again.
River comes 9 of clubs.
In this scenario, a certain set of hands that were statistical losers against A-A would win if they were willing to make calls that were bad pot odds, but not if they weren't. What if the aces were to slow-play at the flop or turn or both? This would change everything. Then, hands that shouldn't have called the opening raise would have the opportunity to beat the hand that had them statistically beaten. Hell, 2-3 of clubs would win if it was willing to stay for river, as well as K-J and J-6 and 6-5.
I don't believe the odds of a given hand improving against RANDOM hands takes into account the types of hands that will call the big blind or a raised or re-raised pot and how those hands would stack up against a given hand. It also doesn't take into account which types of hands can be played differently after the flop. Isn't A-K is an easier hand to play against 9-5-2 rainbow than 4-4?
It's why I think it's going to take a lot longer for a supercomputer to beat a 10 person sit-and-go of strong pros than it took for it to beat a grand master at chess.
Dedicated player of "Drew's Arlo" (Black 6, Red 4) in honor of the late Drew Podolak.