I know what implied odds are and use them in appropriate situations. However, they just seem so nebulous and are completely subjective. How can one use them in any real sense? They just seem like a potential excuse to justify a borderline -ev call/raise. How do you avoid the pitfalls of making poor decisions due to the “implied odds” you were getting? I don’t know, they just seem a little off to me. Maybe I’m doing it wrong.

you are right, a lot of players justify -ev call with having implied odds, when really having reversed implied odds.

i.e. flatting with 42s multiway because you have great implied odds is bad because acutally you have reversed implied odds. you will make flushes, trips etc - and still loose at showdown.
yet a very good player vs a couple of very weak players can make that call because he acutally has implied odds. He knows when to get out of the way and fold a flush because he is beat or when to call a jam on the turn with bottom two pair. but these kind of postflop edges where hero can make super exploitive calls /folds are mostly to find in live games.


it is also very range dependend. when you know Villain has AA/KK , then 65s has a lot of implied odds because you know when you have the best hand aund you will get paid when you have the best hand. but once his range gets wider you dont know anymore whether your flush, straight, 2pair, trips are good, also you dont get paid every single time, cause Villain missed on 552 and gives up... the next time you loose a buy in on 552 because Villain had A5s.

Depends if you know the game play of your opponent for the future node(s). This would allow you to make approximations with regards to EV of a certain hand.

If you are working away from the table you can make some assumptions about your implied odds using combinations from villain's range.

There's more going on with draws than the discussion suggests. BrokenStars kind of hinted at it. I'll try to elaborate:

There are more ev sources(both debit and credit) than (hitting the best hand and winning a big pot). You might also successfully bluff the next street. You might unsuccessfully bluff the next street. You might hit a pair and win unimproved. You might hit a pair and successfully bluffcatch. You might hit a pair and unsuccessfully bluffcatch. You might hit a nutty hand and win stacks. You might hit a nutty hand and lose stacks. You might get free cards. You might improve to a stronger draw on the next street. Your draw might decrease in value on the next street. You might hit your draw on the turn, which gives you redraws. You might be forced to fold the next street unimproved. You might be forced to fold the next street even if you do improve.

The turn is the easiest to mathematize this stuff, as the possibilities are limited. The flop is much more difficult to calculate correctly. Preflop is wide open to even more possibilities.

All that said, I think there are some facts I could state without being out of line:

a) if my draw is a fold in equilibrium,* then I must fold unless I'm gaining ev in some way relative to my ev in equilibrium. Some opponents give lots of free future cards(ev gain for my draws), or they might stack off too light on future cards(ev gain for my draws).

b) if my draw is a call in equilibrium, then I must call unless I'm losing ev in some way relative to my ev in equilibrium. Some opponents force me to fold too often(ev loss for my draws), or they don't pay off enough(ev loss for my draws).

*assuming the entire hand up to this point was played in equilibrium by both opponents.

The implied odds concept is, I believe, valid and useful. However, its implementation is often poor. To incorporate all relevant aspects, one has to consider current pot odds, future equity, effective stack size, future bet size and villain call frequency. The future equity should be composed of at least two factors, the probability you will hit your improvement outs and the probability you will win the hand if you do hit. The latter is a way of incorporating reverse implied odds. It is possible to incorporate all these factors into an implied odds EV equation, which can, for example, be used determine the future bet size you need to make the current -EV call a good decision or determine the minimum win probability given a hit.

Consider the following situation (originally posted in 2p2). Hero has a jack high flush draw on the turn. Villain bets 10 into a pot of 20, which hero has to call to stay in the hand. Should hero call? If he does, what should he bet on the river to provide positive EV?

The following is from an Excel VBA program that incorporates all of the above applied to this problem. A flush draw on the turn has a 19.6% chance for a hit. Using that for hero’s win equity, the model shows hero’s immediate EV is -3.35, so on that basis he should fold. Assume one enters the following additional data: Hero win probability given hit: 85% Villain Call Probability Given Hero Future Bet: 70%

The model shows that to achieve +EV with a hit, hero must bet at least 31.25 equivalent to implied odds of 6.1 to 1. If villain call probability = 30%, then a bet of at least 61.5 is required. (See illustration.)

I suppose the part that’s nebulous to me is the data entered for “hero win probability given hit” and “villain call probability given hero future bet.” What is this based on? If it’s based on villain history, you would need an extraordinarily large amount of ACCURATE data on villain’s tendencies as well as on his current mindset (i.e. a player may not be playing his standard style given his level of tilt, if he’s tired, if he’s sick, etc.). Not to mention whether or not villain is shifting gears or changing his overall style of play. I understand the probability accounts for possible atypical play but again, that would require a great deal of personal knowledge and experience with villain and for that knowledge and experience to be accurate. Aside from the nosebleed stakes with a very limited player pool or home games where you frequently play the same people, I don’t understand how anyone can gather this data with any amount of accuracy. This is especially true for live play without the use of software.

Your questioning about estimating two of the inputs – win given hit probability and villain call probability is valid. But, what would you have to estimate for a math-based decision if not using the model – win probability and fold probability. So, in that sense the model presents a challenge similar to that of a less complete assessment. I think assuming a win when you hit as is usually done in implied odds analysis is not very realistic so attempting a realistic estimate is going in the right direction.

To illustrate how one may go about it, assume a set mime with a 6 pair. Trying a number of flops with the equity calculator Equilab against several opponents with ranges of 20% to 40%, I found that with a 6 on the flop giving hero a low set, he has about a 75% winning chance, which can be used for the win estimate. Of course, bets, villain, board, card evaluations will come into play during the game. I also don’t think you need a massive amount of data using opponent history (HUD) for an initial assessment.

Isn't this what makes up most of poker, making estimates for unknown variables?

Obviously, but estimates based on what? And in this case, we’re using those estimates to justify calls that are mathematically -ev in most cases. If we’re basing this variable on little more than a small sample size (at best), why do we put so much weight on it?

Even if you don't know the exact tendencies of a specific player, sometimes you can profile them so that you can estimate that they have the general tendencies of a certain type of player. You may end up with estimates that are reasonably accurate for members of that player pool but won't necessarily be precise.

I don't need tons of data to profile an opponent as someone who double barrel bluffs flop and turn too often. If I'm looking for someone who pays off a flush too often with non-flush hands, I might only need to see them do it once to know that I can victimize them with a river value bet if I hit my draw.

Ok, but after seeing villain make the play one time, you feel you’d be able to assign an implied odds percentage likelihood that he’ll call a river value bet? We can potentially understand basic player styles and tendencies with limited data (possibly) but we’re using that limited data to make mathematically incorrect plays which is wrong. Why do we allow our hunches and assumptions about how villain may or may not play a hand influence our mathematical odds? I know poker is a game of partial information with an element of gambling, but it seems we’re allowing the gambling element override the math involved in our decisions.

I’m just saying we should stop putting so much weight on an abstract concept based on partial information that’s used to sway our decisions against the mathematical odds.

I'm making an estimate with a high margin of error. That doesn't make the math wrong. It just makes it less certain. Do some people do it wrong? Yes. Is it possible that you are too risk-averse? Yes.

How does it do this? If the math works, the math works. Now, if you are saying that people should realize that estimates are just that, estimates, and they should be careful to not weigh them to heavily, to me that's a given.

When the math says it’s -ev but our estimate based on little information and a large margin of error convinces is to make the -ev play, we are distorting the actual math to fit our needs and justify a -ev play. Although it may be a given to you to not weigh them heavily, it’s not a given to everyone and a number of poker strat books teach to do exactly that.

Op what exactly are you hoping to get out of this discussion? We can all agree implied odds are a real thing and are hard to accurately estimate. What more do you want?

Discussion about their utility. I think it’s pretty clear from my responses. What’s wrong with that?

Ok, I agree. Bad play is bad play.

vs players that bet too much, implied odds go up.

vs players that call too much, implied odds go up.

vs players that bet and call too much, implied odds go way up.

vs players that check too much, implied odds go down.

vs players that fold too much, implied odds go down.

vs players that check and fold too much, implied odds go way down.

-----

there's a thread around here somewhere(couldn't find it), in which a heads up high stakes battle between two pros was analyzed. One player was making seemingly loose flop calls, and winning big pots on the river when his backdoors came in. The other player adjusted by checking back some seemingly very strong hands on the river to deny implied odds when dangerous cards fell.

----

I think that the "bet flop, check turn, value bet river" line that I face vs lots of straight forward tags is most vulnerable to hands that wouldn't have enough draw value to call most turn cards; free river cards = profit for draws.

The bet flop bet turn check river line is much better for denying equity, and if you improve to a monster on the river you can get big value or you can check back if you don't improve.

That said, there's a time and place for both lines. Just noting how implied odds can be controlled and denied by the bet/check decision on the turn and river.

There's nothing wrong with it but it just felt like people came in with reasonable and similar responses and you were just like "but but but no the math doesn't work out!" So I was just confused on exactly what you wanted to discuss.

Oh, they did and my apologies if I didn’t come across as appreciative. I have a tendency to ask a lot of follow-up questions and a lot of “why” questions. It helps me gain a greater understanding of things.

No worries you were respectful and the confusion might have been on my end. Thanks!

Is it fair to analyze implied odds after the hand is completed? Can you say, "I guessed correctly because the guy paid me off?"

Not really because you don’t get paid off 100% of the time.