I wrote awhile back about differentation in poker. A few days later, I started a thread with what I consider a typical differentiation practise, and what I feel is the most important difference between the two situations I described. Aysak posted and asked me to explain what I meant with “overcounting outs” which is a good question, and something that I don’t think people consider enough when they make poker decisions. Today, I’ll try to explain it.

As per the general idea that you should always strive to make your opponents make mistakes while avoiding making them yourself, it’s clearly great if you can get your opponent to fold a better hand, or call with a worse one. The trivially easy example is if you’re holding the nuts - clearly you should want people to call when you bet, and the bigger a bet they will call, the better. Even if you don’t hold the absolute nuts, you still often want your opponent to call. Like if you have a set, or a small straight, or any other kind of good - but not monster - hand, you generally expect to win money when you get called.

But there’s a break-even point. There’s a point where, unless your opponent is drawing dead (has no outs), the bet-size in proportion to the pot-size is going to make you wish that he fold his hand even if you’re currently ahead. With a really big pot, even hands with very few outs would be correct to call a small bet - and that’s what you want to avoid happening. A more common way of phrasing this is “don’t give your opponent sufficient odds to draw.” Clearly, this is more difficult to achieve in limit hold ‘em than in no-limit, but the general concept applies to both games.

So, if we can agree that you want to bet enough that your opponent would be wrong to call, what amount should you bet when you have, say, two pair?

The answer is of course dependent on what cards your opponent has, and what the board is showing. For instance, if your opponent has a flush draw, you should strive to bet enough that he would be wrong to call with his 9 out draw. If you start examining different situations, you’ll very soon discover that how much you should minimally bet depends entirely on how many outs to beat you your opponent has. And this is where it gets interesting!

You know what hand you have. Your opponent, presumably, doesn’t. You, on the other hand, don’t know what your opponent has but you can usually be fairly correct about two things: If you’re currently likely to be ahead or not, and if you’re ahead what the largest number of outs your opponent can have to beat you.

To give an example, if you have AQ and the board shows A-T-4-7, different suits, and from your read on your opponent, you’re sure that he does not have an ace, and unless he’s slowplaying, you currently have the best hand. Since he rarely slowplays, you feel that it’s very likely that your top pair, second kicker, is ahead. So how many outs could your opponent have to beat you? A random hand like K-6 is actually drawing dead at this point. Jack-ten has five outs to beat you (any jack, any ten). K-Q has four outs (any jack) to make a straight; same with QJ (but not with KJ, which has only three outs since you have one of the queens). The hand with the largest number of outs to beat you is 6-5, which would have eight outs (treys and eights).

So if you’re still ahead, in the worst case scenario your opponent has eight outs, making you just a little bit better than an 80% favorite to win. In this case, your opponent must know that you currently have the best hand (because he has six-high) and he’ll know exactly how many outs he has and that he will hold the nuts when he hits. He can make a correct decision as to whether to call or fold when you bet based on the size of the pot and the size of your bet. You’re allowing, in a sense, your opponent to make a perfect decision. Since you should strive to make your opponent to make mistakes, this is not an ideal situation for you - it’s a good situation (being 80% to win is never bad) but it’s not ideal. Ideally, you’d be 80% to win versus an opponent who has reason to think he’s in much better shape than he really is. So let’s change the board a little bit.

If you, instead of holding top pair, second kicker, turned the nut flush when your opponent turned a straight draw, he would suddenly be drawing dead. He might think that he has 8 outs to beat you, and he may therefore think that he’s making the correct move to call a small bet, but he’d be wrong. He’d make a mistake in calling any bet when he’s drawing dead, but he doesn’t know that he’s drawing dead! He’s, as the title of this post suggests, wrong about his outs. This is an excellent situation for you to be in. Now, clearly it’s rare that someone plays 6-5 the way he did here, but the general point still stands: If your opponent thinks he has more outs than he really does, that can cost him dearly.

One more, slightly less far-fetched, example: If you raise preflop with AK, and make a heads-up continuation bet on a rag flop (a flop with only small cards, like 7-4-2), then - at least in limit hold ‘em - you will often be called by random overcards. Someone holding JT might “peel one off” to see if he pairs up on the next street, if he gets good odds on the flop. In order to profitably do so, he must get odds better than 6.5:1. So if the pot is 6 small bets (limit hold ‘em), he checks and you bet, he should call. He has six clean outs to beat you. But what if he has KQ? He’s as likely to call with KQ as with JT (even more likely, obviously), but getting 7:1 he’s correct to call with JT but not with KQ. Because his King outs aren’t live, but he doesn’t know that. He doesn’t have a six out draw, he has a three out draw. This doesn’t have much of a practical application since you would almost always make a continuation bet with AK anyway, but right now I’m trying to focus on getting the general principle of what it means to “overcount your outs.”

And now for the other side of the coin: People may undercount their outs. When that happens, they make bad folds - and your opponent making a bad fold is like christmas for you. For example - and I’m going to use limit hold ‘em again because it’s easier to grasp the concept - if you hold 8-8 in the big blind, the aggressive button open-raises and you make it three bets to see the flop, the pot will be ~6 small bets (depending on rake) when you get to the flop. If the flop comes ace-high, you bet, and your opponent folds a hand like KQ, he will have folded six clean outs in a 7-bet pot. He - incorrectly, but understandably - did not think that he had as many as six outs, but is even likely to think that he’s drawing almost dead to a pair of aces.

So being wrong about outs in the “other direction” can be expensive as well.

What’s the lesson here? That you should try to induce mistakes of this kind. This is why, in the forum thread I linked at the top, I advocate betting when the board has three-to-a-flush, but checking when it’s only two; despite you having the nut flush draw in both cases. Your opponents are much more likely to make a bad call, thinking that their number of outs is greater than it really is.

In closing, I got back from my bachelor party (stag party for the brits among us) yesterday, and am still kinda groggy and worn out, so I apologize for any mistakes I may have made in writing this up. I had a great time, though - and since I know that at least one guy who attended reads this blog regularly: Thanks, Stefan!