Quote:
Originally Posted by ArtyMcFly
I usually open for 2.5x and I don't think I'm missing out on much/any EV by folding. A3o is likely also a GTO button fold (and A7o is probably -EV in the CO) even in low rake environments like 2000NL. (There's a video somewhere of Negreanu analysing some 2kNL hands with holecards face up and when one of the endbosses folded A3o on the button, Dnegs thought it was a misclick, but it's actually a standard fold in tough games.)
It's possible that the weak offsuit aces are profitable for you as CO/BTN opens if you check with filters in your tracker (I didn't even win with ATo/A9o on the button last year, that's probably just variance), but I've always found ragged aces to be breakeven at best. Exploitatively (against blinds that fold too often), they should be profitable stealing hands, but they don't belong in a balanced opening range. You've already got millions of Ax combos for board coverage on Axx.
FWIW, I also think A3o is probably a GTO fold at 3x+ RFI, I think at 2.5x it's really close. I'm kind of coming from an in-practice perspective. I bought some datamined hands awhile ago and compiled the stats of the winningest regs at my stake at that time (although the datamined hands were on a different site) and it showed stealing on the BTN wide (compared to say Snowie) was very profitable. Similar to you, my db shows A6o losing, but A2o winning, but I just assume that's variance/sample size issues. I do agree that opening a hand like A2o isn't printing or anything, and also that a hand like 64s will realize equity better, but I still think most micro to lowstakes online environments call for opening this wide.
People in these pools just aren't fighting for pots hard enough to not justify opening these offsuit aces. Maybe they defend their SB/BB at the right frequency, but most people are x/r OOP at pathetically low levels compared to solver recommendations, probing too little with way too much merge value and wayyyyyy too few bluffs on a lot of textures, overfolding to turn barrels, and delay cbet. All this stuff just makes me think you overrealize pretty significantly.