Wow. I'm experiencing some serious cognitive dissonance here.
(1) Mad Genius is a way better poker player than I am. OK.
(2) Mad Genius advocates going all-in can be a good thing to do with a nut flush draw on the flop, in certain circumstances.
But (3) My gut tells me that this is a total loser of a play.
So I'm going to try to sort through this. Let's make the imaginary hand more concrete.
BB is $1.00. Everybody has $100 in front of them to start the hand.
[Th] in my hand,
[4h][2c] on the board.
There's $19 in the pot from the preflop action minus the rake. Preflop raise was to $4.00, with a total of four callers including both blinds. (Not so likely to get both blinds with a raise like that, are you? But that's the situation. OK.) I'm on the button.
My analysis at this point is that I am clearly going to lose this hand unless another heart hits. Wiith 5 people in a raised pot, somebody has an AK AQ or AJ, and somebody has a pair higher than TT, so I have zero A outs and zero T outs. (This assumes it goes to showdown with nobody folding.) If another heart hits, I'm golden unless the board pairs in which case I'm a little worried. The odds of a heart hitting on the T or R is about 35%.
My further analysis is that this flop is just about as good as I could ask for (other than ATT or a flopped flush). And I have position for the rest of the hand. Therefore, if there's not some pretty large +EV play available to me here, I should never be playing AT suited again.
Blinds check, EP bets $12 (about 2/3 pot), MP calls. There is now $43 in the pot. I still have $95 in front of me. Mad Genius says he goes all-in at this spot.
Ignoring the rake the rest of the way, as I calculate it, here are the possibilities:
No callers: I win $43. Nice outcome.
One caller, I am risking $95 to win ($43+$95) = $138. My odds of winning are around 35%. So 35 times I win 138, 65 times I lose 95. +4830 vs -6175 = -$13.45 average over 100 runs. Not so good.
Two callers, I am risking $95 to win ($43+$95+$95)=$233. 35 times I win 233, 65 times I lose 95. 8155 vs -6175 = +$19.80 average. Better.
Well, OK, maybe it's not as bad as I thought overall.
Except this isn't quite right. Anybody playing a flopped set or two pair (ok, 2 pair is unlikely with this flop on a raised pot!) would be likely to call and would have outs even if I hit my flush. Anybody with a hand less good than that would be more likely to fold anyway. Say my odds of winning are really only 32% with 1 caller and 30% with 2 callers (more out cards to beat me), then my EV ... [works it out on calculator] ... is ... let's see ... -$20.44 (1 caller) and +$3.40 (2 callers). Much less good, even with the +$43 if everybody folds.
OK, this is all theory. But it seems to me that I have a much better chance for a +EV here if I see the turn/river as cheaply as possible, so that I can bet when I'm winning and fold when I'm not.
The basic issue, I think, is that if I play like Mad Genius recommends I will be throwing a lot of money at a pot that I have significantly less than a 50/50 shot of winning if I'm called. When there are 4 others seeing a flop with me and a big pot, an all-in is going to look to somebody like a steal attempt -- so I have to expect to get a caller, don't I? In this situation, I prefer to bet big with a made hand and hope for calls, not bet big with a hopeful hand and hope for folds. My gut tells me that's got to be a bad play most of the time.
But I'm sure I can be corrected by the smarter & more experienced people here. Thanks in advance!