Sunday, December 2, 2012

A Reconciliation of Sorts

I had a sort of epiphany, or at least an "aha! moment", in the shower recently after reading this thread on 2p2.  For those of you who can't go to gambling sites, here's the original post:

60-120 at Commerce, 5 handed late night game. Villain (UTG) is a good Asian TAG, I think he is a pro, might be a poster here, but I don't know who. We don't have a lot of history except a few 40 sessions we played before. I never saw him capping HU OOP before and don't know what his range looks like in this spot. Also, FWIW, I think he hates me. 

He opens, I 3! red sevens next in, blinds fold, he 4 bets and I call
Flop 8J3hhs and I call his c-bet
Turn is another low heart (below 7) he bets quickly and I call quickly, as well
River is the fourth heart, low card again (no straights possible), he now checks.

As an aside the author of the post only played in the game because I told him it was awesome on my way out, and he promptly won a used Toyota Camry, but that's neither here nor their.  The point is that this spot is a super simple one for thinking about the difference between a strategy based on game theory vs one based solely on exploitation, and how you can endeavor to reconcile the differences between them.  Whole books have been written on this subject, but here is my rough take on it, using the river decision on this hand as an example.

I'll probably botch this up pretty badly, but I'll try to wave my hands (vigorously, as my old professor used to's important when hand waving at details to do so with great vigor) enough that it sort of is at least factually accurate.  Suppose we arrive at this river decision, with our opponent finally releasing the initiative on a 4 flush board, and suppose we are endeavoring to use a strategy based entirely on game theory.  What should we do?  Well first we need to know what our range is, then we need to know what his range is, then we need an idea of how he will respond with the various holdings of his range, then we need to take into account the size of the pot and make sure we fire off the right number of value bets (that are still profitable) and bluffs (roughly one bluff for every N value bets, where N is the number of bets in the pot), and after that we're still not even close.  We also have to take into account what hands he's screw playing with, and how often it's for value vs a stone cold bluff, and then figure out how we will respond with the various holdings in our betting range, making sure we don't fold too much (lest he have a "bluff 100% profitably" spot), and also that we have some re-bluffs (probably just from hands that were bluffing in the first place) to balance any remaining value combos (at this point probably just the ace of trump) we have left to three bet.  And you know what, I've probably left a bunch of important stuff out, but you get the idea.  This shit is hard.

Now what if we were just going to try to have a simple exploitative strategy?  Basically all we'd have to do in that case is decide what hands he's going to call with and value bet anything that beats more than half of that range.  Obviously there is a little more to it, since you can't ignore the possibility of a check/raise (so you maybe want to value bet a little more tightly), but that's basically it.  Obviously having a good read on your opponent will help you tremendously here.  Does he have a check/raise range at all?  Is the ace of trump or king of trump in it?  Questions like that are super important in determining how wide you can value bet.

Notice I haven't said anything about bluffing at all, right?  Here's why.  In a vacuum (meaning just for this hand against this guy), your decision on weather or not to bluff can simply be based on weather or not you think your opponent will fold more than 1 in N times (where again N is the pot size) after he checks.  You don't really have to worry (too much) here about how many value bets you have and if you're bluffing at the right frequency or not, because you're playing live 60/120 in a casino you frequent maybe once a month or less.  All you really need to do is decide if you're getting a fold more than one in eight times when you fire the river.  If you think you are, the correct exploitative thing to do is just bet 100% of your range.  If you think you aren't, then you shouldn't bluff with ANYTHING, and all your bets should be for value.  And that's really all there is to it.  Now I'm not saying this is the correct way to play, but it's a methodology that will probably take you pretty far in live mid stakes games.

I thought this post was going to end up being better, but I guess I'm tired and hungry and stressed out from watching Charlie Batch (the man is an American Hero) save the Steeler's season from the bonfire.  Hopefully it at least made a little sense.


Chris Vitch said...

Well I took the time to log in and post a comment when you posted something great so I guess this is fair - this post is just awful and full of stuff that is flat out wrong.

If you are trying to use a GTO approach on the river in this example hand you DO NOT CARE ABOUT VILLAIN's RANGE WHATSOEVER, you do not need to think about what he has, or if he will call, or raise, or any of that, it is all irrelevant. The only question is where are you at in your range and what should you do with that section of your range. It's actually incredibly simple and easy to estimate on such a board. My instinct without doing any counting is you have an easy bet and a slightly harder call if raised.

The exploitative strategy is the harder one IMO, you write "well just figure out what he will call with" and if you can do that its easy to know what to do. Well um, what if you are wrong?! It would be nice if our opponents just told us what they would do with each hand in their range and then actually do it!

The stuff on bluffing is probably the worst and most wrong thing you wrote - firstly because again you make it seem easy to say "well he will fold this hand 1 in 10 times and I only need him to fold 1 in 8 so this is profitable" - except those numbers are completely made up estimates that nobody on earth could ever know with certainty. And even if you somehow COULD know this with certainty that still doesn't make it right to bet 100% of your range! Just because he folds too much doesn't mean he folds every single time, so turning hands with lots of showdown value but not enough to value bet into bluffs is lighting a bet on fire all the times he calls (which he will still do with a high frequency even if he "folds too much").

Jesse Smithnosky said...

My last comment was eaten. This makes me sad.

Your points are obviously all correct DD. Sometimes I practice faith based posting where I start pecking just hoping the thing shapes into something worth saying. My best posts usually come from this method, but it also results in some clunkers such as this post, that just sort of end when I run out of energy and words. I disliked how this one turned out, but your your comments have helped me see what I was hoping to say.

Old school pros are flummoxed by GTO stuff, while the new guard swears it's easy as pie to use once you change (from the ground up) how you think about hands. These same math guys are appalled at some things live pros do, basing decisions on thousands of hours and something they simply know to be true. Or think they know. The two groups have a hard time talking about hands and respecting each other. And while both ways of thinking work well, mixing them within a hand can lead to some hilariously bad plays.

Thanks for the thoughts DD. I'm happy you're even reading let alone commenting.