Am I betting for value enough

[Louie Landale]

You are on the river with a reasonable hand and the opponent has checked. Theory says you should bet any hand (for value) that has a 55% of being the best hand when called (5% covers getting raised). Fine.

Now lets say you want to evaluate your betting strategy on the river. It seems to me you can keep track of your value-bet attempts and successes, and if your sample is large enough get a good idea if you should be betting more or less often. I was wonder, however, how you would use these statistics to gauge it:

[] If you are betting the hand that is at 55% then you are also betting hands that are 60%, 70%, etc.

[] If you are winning 55% over-all then you are clearly betting FAR too many hands, since to counter the 80% hands you would need to be (incorrectly) betting the 30% hands.

[] I don't think you can "average" 55% through 100% to get a target of 77.5% since there are generally more (a lot more) marginal hands in the 55% range than "nuts" in the 100% range.

I'm guessing your target should be around 70%; that is if your river value bets win 70% when called you are betting for value optimally.

But invite comments on that.

- Louie

[Copernicus]

I would think anything in the 70-80% range might be reasonable if you are just counting % of wins, including folds.

If it is only showdowns then the lower end of that range or even into the 60s makes sense to me. First of all, on hands you bet with the high likelihoods of winning, you will get many more folds, so they dont go to showdown. Also, there are bound to be times when you have over-estimated your likelihood of winning, which means some hands will actually be a less than 55%, dragging things down.

[MaxPower]

To determine if you are value betting optimally, wouldn't you have to also look at the times you don't value bet? When you check it down, you can see the percentage of times you win the showdown, but you don't know how many times your value bet would have been called.

[Mike]

Hello Louie,

I kind of think you can't put value betting in such neat boxes. Two different scenerios if you will:

I am at a tough table where I know I am going against a good hand on the river no matter who holds it, I tend to value bet less.

I am at an easy table, I tend to value bet more.

The reason for the different betting characteristics is the value in the value bet. A tougher table does not give as much value on the river as an easy table.

The tough games I find myself in are more of pushing the same size pot around the table than any player having a clear advantage over the others. The way I see it if there is no clear advantage the value is also minimal. If the value is minimal, you are then paying too much for an item. Hence no value betting.

I do agree with your last statement about betting optimally. Now that I said all that, have I missed your point completely?

[ramjam]

Some thoughts (not necessarily answers):

-The 5% add-on for the risk of being raised by a better hand seems to be an essentially arbitrary add-on. I would think an analysis of your "value betting" performance would have to consider how well you anticipate raises and deal with them - which can't be captured by a straight win percentage.
-When faced with the decision to value bet, you are deciding between betting or checking behind, check-calling, check-folding and check-raising. You can't really assess whether your value betting strategy is optimal based solely on the percentage of times when your bets are called by a better/worse hand - or, more generally, just by looking at the results of your actual value bets. In many cases, you might gain more by checking and inducing a bluff or encouraging a weak hand in late position to mistakenly bet out - hands which might not have called your value bet (or not overcalled an intermediate caller).[Consider following moderately plausible example. Player A has 2nd pair. Players B and C have 3rd and 4th pairs or 3rd pair with weak/strong kickers. If A bets, B may call and C fold. If A checks, B may bet, C call and A makes a winning check-overcall.]
-I would think that you would have to adjust the percentages for multiway pots / multiple callers. If you bet out into a large field, (ignoring possible raises) your value bet will have positive expectation if there is a 50% chance that it gets called by one worse hand or a 33.3% chance that it gets called by two worse hands and so on.

[Louie Landale]

[1] At 66% you break even vis-a-vis a raise since you are giving up 2:1 for your 2:1 shot. Adding 5% resulting in 55% may be arbitrary, but its embraced by the Authors and I see no reason to disagree.

[2] Excellent points: gauging your river bet-for-value ratio is only ONE element of a perfect river strategy. Other categories include [a] hand isn't worth betting but I check-and-call [b] hand is worth betting but I check-raise [c] hand is worth betting but I check-call to induce a bluff [d] I'm advertising for affect so this hand doesn't count.

[3] Multi-pots adjustment is easy: if you get 2 calls then add 2 to your called-river-bets column.

- Louie

[Louie Landale]

No. Uncalled bets don't count at all. (In fact, they end up counting against you in the "should I bet or check and induce a bluff" tactic.) If you have a king high straight flush showing its foolhardy to bet it for value, even though you are going to "win" 99% of the time.

Yes, this raw statistic will NOT distinguish a "perfect" strategy with one that [] bets too conservatively most of the time AND [] makes some hopeless bets at other times.

- Louie

[Louie Landale]

Yes, noticing % wins when there was no bet is interesting, but I really don't know how someone can use that information objectively to improve the strategy. Other than bet more or less often and gauge the bet-for-value ratio, the topic of this thread.

- Louie

[George Rice]

Essentially what you are trying to do is use statistical analysis to determine if your estimate of a 55% better hand rate if called is accurate. Some thoughts.

You should also be bluffing on the end sometimes. Both because you have a reasonable chance to get a better hand to fold and to make your opponent call more often when you value bet. So an analysis will have to distinguish the value bets from the bluffs. This is easy to do manually, but if you are relying on a computer program that just keeps track of river bets, then that would be a problem. I'm not suggesting that you would do this, but some may not realize this at first. Also, notice how bluffing increases your sampling size for your value bets and changes what a 55% hand would be.

It may be more accurate to logically estimate the chances based on the cards and the play of the hand so far, assuming a somewhat predictable opponent, than using a sampling, because the sampling will probably be too small to be meaningful. It may take many years to obtain enough hands for a meaningful analysis.

You may be betting hands with at least a 55% chance of winning if called, but are you betting ALL such hands? A statistical analysis won't necessarily tell you this. In other words, your sampling may only be a subset of the actual number of hands you could have bet, yielding the same percentage of winners but at a lower money gain. Of course it's more likely to miss a 60% hand than a 90% hand, which would raise your winning percentage and tell you something is wrong, but you don't have to miss that many 90% hands to offset a bunch of 60% hands either.

You may be missing certain types of hands, or including certain types of hands you shouldn't, but a statistical analysis might only identify THAT you are doing this, not WHERE you are doing this.

I think the best way to check your value bets is to analyze hands played, hand by hand, using probability and logic to determine the possible hands your opponent held and whether you would have won 55% or more. This will also give you an intuitive feel for it.

[Louie Landale]

I disagree, this IS a great "neat little box" statistically. The situation is clear as is the outcome, and there IS a "perfect" ratio against which you results can be measured. Perhaps not "perfectly" measured but certainly "usefully": if the perfect is 72% and your results are 85%, you should value-bet more often.

Your tight/loose game adjustment is sound. However, it you think it matters to this topic then just have 3 groups of numbers [] value betting in "tight" games, value betting in "moderate" games, and value betting in "loose" games.

Lets say you raise with KJ, flop a K, bet it twice, and here you are at the river... This is a no brainer bet against typical loose fish; but against tight sensible players it gets marginal fast.

But not for the reasons you gave. [1] the tight player is already MORE likely to have a hand that can beat you (KQ in this example), [2] Less likely to have a bad hand that can call you all the way (say a pair of 4s with an Ace), and [2] the tight (and sensible) player is LESS likely to pay you off since he can more realistically read your hand accurately. The fact that you and the tight player have no particular advantage over each other in the long run is irrelevant.

- Louie