Poker strategy: betting hard with a medium strength hand

Short, Average, Tall

At the low stakes poker tables I play at, a strategy I often see is for a player to flop a medium strength hand (middle pair, say) and then to bet really big to try to take the pot down.

Before rubbishing this play, let’s pause to think why this might have some merit:

1. Aggression is generally good at the poker table. (Well, not physical aggression . . . )
2. Many of our pots are multi-way, so middle pair is unlikely to be best. It therefore seems a reasonable candidate to semi bluff.
3. Players with thousands of hours of experience at the felt seem to think it’s a good play.

So maybe this is a positive Expected Value (EV) play against a recreational field. Maybe. But I would argue in the case of 1. and 2. that there are better candidate hands to play aggressively.

For point 3., as I have noted many times in these blogs before, the feedback loop in poker is terrible. One can make negative EV, losing plays, for many years and never realise it.

We all only get our fair share of equity

One way to think of poker is that over time we’ll be dealt poor hands, medium hands and strong hands. One way to measure a hand strength is the raw equity of a hand — the probability that the hand will win if it goes to showdown. In a 9-handed game we’ll be dealt hands with equity of 1/9th of the pot on average, in the long run.

Hands with high equity are — needless to say — good. We can bet them for value. Often the EV of these hands will exceed their equity as they win their share of the pot plus future bets that go into the pot. Low equity hands are also good, at least in the sense that it’s easy to exceed their raw equity. We can take a hand with 0% equity and bluff with it.

The trouble with bluffing medium strength hands is that we cannibalise the equity they have. Only hands that beat them call, and if the opponent folds then we could just as well have bluffed them with a zero equity hand.

The theory would say that we should play our medium strength hands passively and try to get to showdown as cheaply as possible. We try to realise what (limited) equity they have.

Four types of hands

Broadly speaking, we can place any range of hands into four categories:

1. Premium made hands: these are the hands that we’d happily play for stacks. Sometimes our opponents will put the money in for us, but often we need to put the money in ourselves (i.e. we need to bet).
2. Marginal made hands: I’ll come back to these.
3. Draws: these hands are good candidates to play aggressively as a semi-bluff. They have little equity unimproved. So when we bet, raise, or check-raise and everyone folds, we’re happy with this outcome, but we still have some outs when they call. Good players may even bet the ‘junky’ draws harder than the good draws. We’re ‘cannibalizing’ less equity to bluff, say, a gut-shot straight than to bluff a flush-straight combo draw, so this makes sense.
4. Junk: just let the air-balls go. Of course, I’m simplifying a lot. A computer that’s playing game theory optimally will always be perfectly balanced, and in cases where it needs to ‘find’ bluffs, then it will bluff the junk too. That said, multi-way pots encourage more honest play, and there are seldom situations where it makes sense to bet a junk hand without even backdoor potential to improve.

Which brings me back to the ‘marginal made’, or medium strength hand. Our middle paired hands. The theory says that these should be played passively. Sure, we can play them as a semi-bluff, but assuming they are only called by stronger hands, we generally only have maybe 5 outs to improve (viz. we can improve to 2-pair or trips). Straight or Flush draws are better candidates, with 8 or 9 outs respectively.

Unfortunately, this means that medium strength hands tend to underperform their equity (EV<EQ), especially when played out of position. When the opponent is betting and we are calling, they will have the polarised range and we will have the condensed range. They either have the nuts or nothing. We just have the bluff catcher. If they construct their betting and bluffing range perfectly, they can leave the bluff catcher with an EV of zero — indifferent between calling and folding. This is not an enviable position to be in, granted, but we may lose even more money by trying to bet these hands hard from the flop.

Betting into the nuts (e.g. monotone boards)

A related interesting play I see is when the flop is a monotone board (all diamonds, say). The theory would suggest that the preflop aggressor (‘PFA’, who should generally be the one continuation betting on the flop) should bet small, 25% to 33% of the pot. Yet I often see players betting big, 66% to 100% of the pot. Again, before rubbishing this play, we should consider why this might have merit:

1. Many people (myself included) don’t like playing monotone boards, unless we have a good flush draw - preferably the nut flush draw (Ace of diamonds, in this case). It’s therefore an opportunity to steal the pot from weak opposition.
2. A monotone board therefore acts a bit like a scare card.
3. It’s actually quite difficult to flop a flush, even if it’s a monotone board, as the board removes 3 cards from the suit we need. The same is true of paired board flops such as K55. I see players bet big to represent having the 5. This play works (mostly!) because it is quite difficult for any other player to have a 5 — the board has removed two 5s from the deck.

The theory says that monotone boards should be bet small: either we have the nut flush and no one will call a big bet, or we don’t have the flush and someone else does! The paired board situation is a bit more nuanced: as the PFA we will generally have the range advantage and should bet frequently, but if the board is K55 it’s possible that the defender(s) has more 5s in their range than we do.

If we happen to have the nuts (the flush on a monotone board, or the 5 on a K55 board) then it’s hard for our opponents to have anything; we’re unlikely to be called.

If we don’t have the nuts it’s still fine to bet frequently but betting large is akin to betting large with a medium strength hand. We’re likely to get called by a monster hand too often for it to be worth our while.

In summary

Medium strength hands such as middle pairs are uncomfortable to play. It may therefore be tempting to turn them into bluffs. Unless our player pool really rewards aggressive play however (by folding a lot), the best EV play with medium strength hands is to play them passively.