What’s Your Play? Over-Limp-Back-Raised Results

I was pleasantly surprised by the great discussion of this week’s What’s Your Play? I nearly didn’t post it because it was “just a pre-flop spot”.

How Likely Are Better Hands?

Everyone was rightfully quick to recognize that it’s rare to see people overlimping the biggest pairs in late position. A few of you asked a very good question about whether there was someone in late position or the blinds who could be expected to raise with a high frequency; there was not. This is indeed a good reason to think KK/AA are unlikely.

However, Villain is also committing 64, and quite probably 145, big blinds to the pot before the flop. Though there could be other explanations for this, which we’ll get into in a moment, this is evidence of a big hand. The more money someone is willing to put into the pot, the stronger he’s likely to be. Sometimes hand reading is that simple.

We have no reason to think this is a spazzy or gambley player. Quite the opposite, he’s very young, usually plays smaller stakes, likely relying on poker income, likely undercapitalized for this game, and initially bought in short. There’s nothing here to indicate that he has the stomach for putting so much into the pot on a lark.

As I indicated in the comments, the “big bets are weaker than small bets” line of thinking is a dangerous one in that it flies in the face of the fundamental mathematics of the game. Yes, it seems a little odd that Villain seems to be trying to shut down the action, but that doesn’t have to indicate a non-nut hand. Perhaps he’s just nervous about playing post-flop and wants to end the hand pre-flop one way or the other even when he has the nuts and the pot is already large. Perhaps he thinks worse hands will give action even to a bet this large. Perhaps he thinks we will think that a large bet will look weak. Unless you have a very good reason, you should continue more often against small bets than against large bets.

In this case we have some conflicting evidence. The overlimp suggests a non-nut hand, but the big raise suggests a big hand. Here’s a good rule of thumb: the more money involved in a particular decision, the more valuable the information gleaned from that decision. In other words, unconventional or seemingly suboptimal plays are a lot more likely when Villain is making a $5 decision than when he is making  a $320 decision.

How Likely Are Worse Hands?

Gareth cuts straight to the heart of the matter:

I think we are weighing a few events of unlikely probability.

Event 1: villain overlimps in this spot and then takes the line they did. What has happened before our eyes is an event of very small probability.

Event 2: Villain overlimped in the first place with KK/AA.

Event 3: Villain overlimped in the first place with {not QQ+AK} and is now losing their mind for no apparent reason. And when calling is a viable option, having a reasonable price in a reasonable position (in position postflop that is).

All three of these events have small probabilities. Yet Event 1 has already happened.

This is recurring theme of these WYP posts: Villain has to have something. It’s not enough simply to say “better hands are unlikely, so let’s stick it in”, because sometimes worse hands are also unlikely.

That said, I think Gareth understates a bit the possibility that Villain could have slightly weaker hands than QQ that he still thinks are quite strong. Hands like TT, JJ, AQ, and AK all seem plausible as hands that he didn’t feel like raising the first time around but that he’s now willing to play aggressively in the hopes of either locking up an already somewhat large pot or commit. Overall I think he’s less likely to make the big raise with JJ than with KK, but I also think he’s less likely to overlimp in the first place and thus have the opportunity to make the big raise with KK.

What About Calling?

I didn’t even consider calling. Villain has virtually half his stack in the pot, so he’s extremely unlikely to fold to a shove. What’s more, as Gareth points out, we’d actually like him to fold a hand like AJ. Assuming Villain ever folds postflop, he’s far more likely to fold correctly (say, with an AJ that whiffed the flop or TT on AQx) than incorrectly.

I also didn’t feel confident enough in my ability to get away on any particular board. With so much money already in the pot, it’s certainly possible that Villain would just ship the rest even with hands worse than QQ on an Axx board, which means check-folding even some of the worst flops a questionable proposition. It’s true that with a very specific grasp of Villain’s range, you may be able to do better by calling than shoving, but I think that’s mostly a theoretical matter, and here a specific grasp of his range is exactly what I lacked.

Results

Hero’s equity against {JJ+,AK} is not great. Assuming Villain never folds, we need better than 45% to get it in, and we have 47%. If you allow for a small chance of the UTG player waking up with KK+, that makes a shove vs that range basically neutral EV. So the big question is how to weight it: are some hands in that range more likely than others, or is there some chance is range is a bit wider than this?

Nate addresses this subject well:

we have to discount AA twice: once given that it didn’t raise the first time and again because it raised so big. A further problem is that tendencies to do these are often negatively correlated, so we have two unlikely and perhaps negatively correlated events that have to have occurred for the Villain to have AA. Now, overlimping JJ is also perhaps a nonstandard play, and raising this big with it too, but these are (IMO) less unlikely events taken individually, and (also IMO) they arise from *positively* correlated tendencies. Similar comments apply to TT and maybe even AK.

I argued before that it was dangerous to rely heavily on an assumption that a big raise is less likely to be a big hand, so it’s important to note here that we are looking at only a small deviation from otherwise correct play. If we thought AA was a little more likely than other hands in the range, we’d have a close fold. Now, looking at it as a little less likely, we have a close shove. In neither case are we deviating wildly from the otherwise correct play.

Now if you were to rely very heavily on this assumption and exclude QQ+ from Villain’s range entirely, you could end up shoving TT here, which I think would be a huge mistake.

The other bit of information I chose not to rely heavily on was my read that Villain was an undercapitalized nit playing above his head. My evidence for this weak, but if I’d had more confidence in it, I would have folded.

Basically I decided QQ was the worst hand I’d move in with here, and if he happened to have set some weird trap with KK or AA and I woke up with exactly QQ, then good on him. I shoved and apparently he’d set some weird trap with KK, so good on him.

Thanks again, everyone – this turned out a lot better than I expected, and everyone who commented deserves the credit for that.