Playing small pairs?

[counterspell]

i raise if i know the limper and can isolate often, otherwise overlimp.

[palamedes]

passive table: I usually call for set value (esp if the table is also loose) -- I think that the implied odds are clearly there, especially since it's fairly easy to play this hand nearly perfectly post-flop against many players except in the rare cases where you face set over set, and a few tough decisions such as a set on a 3-flushed board against a tricky player.

aggressive table (i.e., I'm likely to get raised or even raised then 3-bet pre-flop): fold. The analysis below suggests why.

I'm not crazy about isolating here unless I have good reads on the player who is likely to be isolated. Otherwise (i,.e., without reads), I think I probably play badly in this case, so that's a personal problem, and if you're a great heads-up player (I'm not) and can get it HU, then sure, why not? Maybe then your EV is better than it is playing for the set against a larger field.

I've heard it argued that raising here is good if you can be relatively certain that the pot will be at least three-handed and preferably 4-handed (or more). The argument goes that this move increases your implied odds, because it ties players to the pot post-flop when you hit a set, but it remains very easy for you to get away from the hand if you don't flop a set. For example, flop comes J82. A player with something like 86 (forget about why they called 2 bets with that hand...) might stick around if the pot is large enough, but not if it's small. I'm not very enamored of this argument though. Perhaps it makes sense if I happen to know that the table is EXTRA loose and I can also count on somebody else to do some betting for me post-flop. Here's a VERY(!) rough analysis of the two cases (raise vs. no-raise)

LIMP ALONG: suppose the pot will be 4-handed, with SB and BB in (+ you and UTG of course). Pot is 4SB. You flop a set that will eventually win showdown (suppose this happens 10% of the time -- I'm eyeballing it), and proceed to get, on average, say, 3 callers on the flop, 2 on the turn, and one on the river. (Remember, this table is by hypothesis loose. And sometimes you'll be able to put in a raise on the turn or river, thus increasing the average pot size to make up for the times when you bet the flop and get just one caller, for example.) So for your initial investment of 1 SB, you win 13SB, which works out fine for you. (Yes, I'm ignoring the times that you flop a set then lose. I guess that happens around 2.5% of the time or so. Some of those times you lose a lot (set over set) and other times not so much (on obvious straight or flush boards where at the very least you'll slow down if not fold).) Let's say it all breaks even-ish in this case.

RAISE: Surely on average you'll lose some callers here. Let's be somewhat optimistic and say that on average there will be 2.5 callers, so the pot will be 3.5-handed. So you put in 2SB pre-flop and there are 7SB in the pot (on average). You hit your set. The theory here is supposed to be that you have tied people to the pot, so let's suppose that the theory is true, and that you get 3 callers on the flop, 2 on the turn, and 2 on the river (on average -- again, these bets could come in other ways against fewer players, for example, by raising, but remember that we are not supposing that these players are maniacs). You win 18SB. You invested 2. You come out worse in this case than when you limped, and in fact you are now -EV, even without taking into account the times that you hit your set and lose.

Obviously I'm making all sorts of assumptions here, many of which could easily be challenged (though many of them are neutral between the two cases, I think). And sorry for engaging in an analysis that nobody here has specifically requested, but it was recently an annoyance of mine when I heard the argument made, and even though the analysis above is rough, it does strike me as prima facie good evidence against the 'raise to tie them to the pot' theory.

As for the 'raise to isolate' play, as I said, good HU players might want to try it. I'm not, so I don't.

By the way, just to see, I checked my PT stats on 22 over the past 15K hands, just to see. I have in fact pretty faithfully taken the line described above. The only other thing to mention is that I raise first in with 22 from the button or the cutoff. In the past 15K hands I've been dealt 22 91 times (quite a bit more than expected -- I guess I'm lucky, lol) and VPIP is 52%. (Detail: Button 70%, cut-off 76%, UTG+1 17%, UTG 9% (that was one hand and I'm not sure what I was doing there), BB 33%, SB 83%.) On the button and in the cut-off I raise first-in, and my wins there normally come from just winning the blinds. Otherwise, with one exception, my wins have come from hitting a set. (I also once lost to quad over quad with it. I didn't need to be reminded of that hand...) Overall BB/hand is .42, but looking at just the times that I limped, BB/hand is .09, which suggests that the analysis above (which seems to suggest that limping under the right conditions is about break-even) is about right -- most of the value comes from raising in late position and stealing the blinds or following through post-flop to win.

By the way, my VPIP in the SB is way too big - I just discovered a leak here. (That's the only position from which I'm losing money with 22, and now it's clear why -- I play it too often from there...)

To put these numbers in context, I am a decent but not great SH player. I generally play 3/6 or 5/10, and over those same 15K hands I am winning 2.9BB/100, according to PT.

Anyway, thanks for the opportunity to think through the lowly ducks. I never gave them much thought before (especially, apparently, in the SB!)