Just The Facts: Dissecting Deuces

Poker Tracking software is an amazing thing. There's really no end to what you can learn about the game; that is, if you bother to spend the time trying to cull out the gems. I wouldn't be at all surprised if every one of us could delve into some minutia of the game and come up with some nugget of wisdom that very few players, even pros, are privy too.

Lately I've been thinking a lot about pocket pairs, small pocket pairs to be more precise. I had in mind running a series of reports on how to play small pairs 22-66 facing a raise. I came to the conclusion that for the most part, calling a raise of 3bb's or less with small pairs was safe enough but even given the range of 22-66 I noticed enough variation between the high and low to conclude that I had to narrow my investigation even further and landed squarely on Deuces, the smallest of the small.

So without further ado, here are my conclusions on how to play deuces based on 128,000 hands of Tournament No Limit Hold 'em. I have tried to stick only to the facts based on my data base and there is very little of my opinion in here.

Total Hands dealt = 127,787
Total Number of times dealt Deuces = 615 =1/208 hands (A little low but not suspiciously so)
Win % = 27.64
BB/Hand = -.22

My win rate is higher than I would have guessed. The -.22bb/hand would suggest that I'm better off not playing it at all. But rather than just chopping it out of my playbook altogether lets take a look and see if there aren't specific situations where I'm winning or losing. Lets start by position.

0(Button) = -.18bb/1
1 ob (CO) = -.86bb/1
2 ob (HJ)= -.82bb/1
3 ob = -.13
4 ob = -.22
5 ob = -.33
6 ob = +.56
BB = -.12
SB =+.31

So we can see that except for the blinds playing 22 even on the Button or the CO is an overall loser. There is that one strange +EV in position 6 but I can't see any reason why playing deuces from UTG has an advantage over the button.

So are we done? Not by a long shot. I am a firm believer that given the right conditions no hand is unplayable. So what am I doing wrong? The above stats are taken from the complete database irregardless of preflop action. It mixes in limped deuces with those raised, called, limped, check and folded. We need to break it down a bit more.

What if we just limp, open or otherwise? Ugh! a loser in almost every position with a couple of nonsensical positives thrown in to add legitimacy. Even the button is a disaster at -.49bb/1. Total all positions = -.53bb/1 win rate = 18%

If limping is bad, calling a raise must be worse right? Yup, but not much worse.
0 = -.36
1 = -1.18
2= -5.68
3= -1.24
4= -0.01
5= -1.76
6= -0.20
BB= +0.48
SB= +0.22
Again this hand is a total dog in every position other than the blinds. It would seem that the blinds provide us with the opportunity to tailor our calls to the correct pot odds (or perhaps the correct opponent) without much risk of facing a raise. It's also interesting that we actually did better in the blinds (much better) facing a raise than we did just limping or even checking the BB. I could be variance or it could be that when we hit a set in a raised pot we're much more likely to score vs a laggy opponent. This would also explain why the blinds do better than the late positions. Yes the pot odds are better, but that can't be the whole story. My guess is that the blinds offer checkraising opportunities that the cut-off and the button do not.

So I guess that means we only call raises from the blinds. OK fair enough but perhaps we can do better. Can we call any raise? We need another break down. Now here's where I really wish I had a much larger database. Even at 130k hands the sample size just isn't large enough to draw any solid conclusions based on specific raise sizes, but here's what I have.

If I remove all raises above 2.5 BB's 22 becomes an overall winner at .13bb/hand and solid from several positions (sadly not the blinds or the button). Again, sample size is small but I think there's a pattern. At the very least we can conclude that calling a small raise isn't a big mistake and may even have pos EV. Again, there's that discrepency between calling and limping. It seem counter intuitive that calling a raise would increase our EV. It's just a guess to say that while our direct odds decrease, our implied odds more than make up for it. I think it's fair to say that if somone isn't excited enough about their hand to raise preflop their very unlikely to be too keen on getting stacked post flop without actually hitting a hand that can beat your bottom set.

I've got more to say on the matter but I've got a house full of hungry kids and I've already spent more time on this than I anticipated. When I come back I'd like go into the subject of raising with deuces and how the size of the blinds effects the profitability or lack thereof.

Interesting post... I suspect it would only be worth playing dueces if you know your opponent will fold to a large amount of bets and/or they are deep enough for set odds.

This would be most likely because you are raising UTG... your hand is getting quite a deal of respect from your opponents... whereas on the btn your hand gets a great deal less respect from the blinds who most likely defend widish, so you either need them to fold or set to win the hand in most cases.

Interesting theory. There could be something to that. I'll take a closer look tonight.