Friday, April 13, 2012

GTO Musings

OK that's a pretty heavy title for what probably won't be a very long or interesting post, but I've spent a lot of time and energy on GTO (game theory optimal) play recently and it just sorta feels like I should write something about it. So I guess, here it is....

I recently read The Intelligent Poker Player, which spends a good bit of ink explaining the difference between the GTO (math, balance) and exploitative (feel, soul reading) schools of thought. For those who are not, you know, professional poker players, the basic ideas here can be stated rather simply. Limit Hold'em is a solvable game, at least in heads up situations, and there exists a Nash Equilibrium nirvana known as "GTO". If you play this way, or even very close to this way, with a good understanding of your opponents ranges, you are basically unbeatable. No matter what anybody else does, they cannot win money from you in the long run. An example of a GTO concept would be having 1 bluffing combo in your river betting range for every 10 value betting combos in a 10 big bet pot. There are lots of other ideas and concepts, but it shouldn't be that hard to wrap your head around the fact that there is a "correct" way to play that involves balancing your bluffs and value bets and call downs in such a way that you are not exploitable. Another concept is that if your opponent raises you on the river, putting the 7th and 8th bets into the pot, you should not fold more than 1/4th of the time. If you do, he should raise you 100% of the time, since he's risking only 2 bets to steal 6 (I think I did that math right there).

The exploitative approach doesn't really worry about a lot of these concepts, and it's not hard to see why. The concepts I stated above rely on your opponents having half a brain and figuring out that you have a "hole in your swing", as it were. Most of the people I play with day in and day out simply aren't going to get to the river in a 10 bet pot and realize they have a profitable air bluffing opportunity with 100% of their range. Heck, most of them aren't going to bluff the river even close to the correct percentage of the time, and some will literally NEVER do it. So making these sorts of GTO justified calls on the river really can, to quote a bike 20 player, put a lot of wear and tear on the old bankroll. You see, the idea of GTO play is to make sure that nobody can exploit you and ensure that you will win. The idea is not, however, to play in a way to maximize your expected value against weak opposition. Bet/folding the river is one of the bread and butter plays of beating low and mid-stakes limit hold 'em. I'm sure that I do it way more than GTO would ever allow, and I'm also sure that the vast majority of my opponents have no hope of ever taking advantage of it. There are other times when relying on GTO concepts is simply leaving money on the table, for example when an opponent simply is never folding. These times in theory should not exist, but there are opponent and line combinations from which you can sometimes deduce that your bluff with NEVER succeed. So save the bet....

I've spoken with a few friends about the subject, and spent some time thinking about some pros I know that come from each school. Most of the hotshot internet guys turned live pros (DosEquis, Juice, Sailboats, La Peste) come down decidedly on the GTO side of things. Other players, often those who have played live forever (MikeL), are definitely more in the exploitative camp. And there's quite a bit of contention and name calling along these lines, more so than you'd probably expect. The online guys, the ones who have seen the long run and have played millions of hands in pretty big games against some tough opposition and managed to show a positive win rate often label us live guys (and I'm definitely a live guy...I've probably logged less than 300K hands on the internet) as downright fish. And you know what, we don't really have a good response, because in truth what they do IS more sophisticated than what we do and they did beat games that we could never beat, but the fact of the matter is that a lot of them spend way too much time thinking about how they play their entire range in a certain spot than they do just attacking the low hanging fruit. I'm not saying that GTO play is the wrong approach, or even a less than wonderful one. What I'm saying is that if you find yourself playing in games where you need to be considering GTO concepts against more than a few of your opponents, you should probably look for a better game. Sure, there are definitely spots where it's important. If you find yourself playing 200/400 at Commerce with 4 world class pros, 2 live pros and 2 big fish, you're going to need to be pretty close to GTO against the pros in order to avoid losing back your profits from the fish. So yes, guys who have a very strong, almost second nature style understanding of GTO play have a big advantage in that they can safely sit in almost any game.

But are these concepts a big part of beating up the LA mid stakes games? Certainly not. The way to win in the games I play is to study your opponents, figure out what mistakes they are making, and adjust your play in order to exploit them. A guy opens too many hands? 3 bet him light. A guy always seems to be showing down and quietly folding his hand? Value bet him thin. A guy barrels the turn too much? Delay your aggression until the turn with both your value hands and semi-bluffs, and also fold the flop a little more because you won't really ever be seeing the river for free. A guy bet folds the turn a ton? Delay your value raises until the river, and raise the turn almost exclusively as a bluff. A guy literally never value bets the river? Donk when you make your hand (I did this to the biggest fish in our game like 3 or 4 times today). If you find yourself thinking about GTO concepts against more than 1 or perhaps 2 of the players in your 40/80 game, you're probably in a bad game and you should find a better one unless you have an amazing seat. Again, I'm not saying that you shouldn't strive to understand GTO play as you move up the poker ladder. What I am saying is that you should be careful not to see every situation as a nail and GTO as your only hammer. Most of the time the most +EV play in a good mid stakes game will involve deviating from the GTO book, and the pros who understand that concept and apply it consistently correctly are the ones who post big (close to a bet per hour) win rates over reasonable (thousands of hour) samples.

I thought I was done, but I have one more thing to say. In a high volume online environment, it is very easy to have a relatively small win rate and still produce quite a bit of income. For example, back when I was grinding away on the Full Tilt 3/6 games, it was pretty easy for me to play 300 hands/hour. Supposing a win rate of only 1 bet per 100 hands, you could win $18/hour. Throw in rake back and the rewards program and you could basically double that, and all of a sudden you were making something like $35/hour from the comfort of your office without the inconvenience of wearing pants. If you played much higher than that, obviously things got even better. The point here though is that with high volume (and rake back) comes the ability to have a small win rate. Live players with win rates of 1 bet per 100 stand almost no chance of surviving. Go ahead, run it if you want, but things are just brutal when your WR is that low. Remember that most people assume 30 hands per hour (I personally think that 40 is more realistic, but haven't ever been able to count), so 1 bet per 100 hands is literally something like .3 to .4 bets per hour. In order to survive on that rate you need to be playing pretty big; obviously 20/40 isn't really going to cut it. And the downswings you're going to hit...just brutal really.

So what's the point? In order to survive playing poker live for a living, you HAVE to have a very high win rate, one that can perhaps only be generated by playing an exploitative style against relatively weak opposition on a regular basis. If you're trying to eek out .2 bets hour playing in tough 200/400 games, sure you're theoretically making $80/hour, but the bankroll you need to survive the swings that come with that territory is so large that you obviously (existence proof) have the ability to make more than that (or at least close to that) much doing something else. So while I haven't thought this out completely, I think that I might think that playing a maximally exploitative style and making sure you put yourself in good games could be more important in determining your success as a professional poker player than knowing the first thing about GTO play.

At least now that they blew up the internet :)


ExMember said...

Short version: Game-theory optimal poker cannot be beaten. Exploitive strategies make way more money because your opponent's play is nowhere near game-theory optimal. Therefore you should focus on exploitive strategies.

Dustin said...

Good read Jesse. Funny thing about this, before what I really knew what GTO really was (haha still not QUITE sure what it is) I fancied myself an exploitative player. As it turns out my natural style, and WTSDboundness especially, were more GTO based than I thought. It was more of a happy accident than it was by design.

I think all of GTO type guys that you mentioned deviate appropriately when facing bad players. This book really has done wonders for my game though. If you follow his rules for holdem (until you know where to deviate exploitively) it makes playing poker a lot easier, effectively hides information, and most importantly keeps YOU from making mistakes. I think exploitation can sometimes be used as an excuse for having poor fundamentals, and especially as an excuse for FPS.

RE: exploitative folds- I think they are overrated. Definitely live players take this too far. I was thinking when I started playing live, that I would need to start folding a lot more. Thing is people bluff, people misvalue their hands, people do stupid shit. I am surprised over and over again, and I think I am making more money than the exploitative folder on my thin river calls. Of course I do make some folds against some guys that I wouldn't against others, that part is just common sense.

Dustin said...

Oh and forgot to state my point. It's going to very hard to make more money playing FLH than the "GTO" guys you mentioned. Hardly an endorsement for the exploitative player making a lot more money in live Los Angeles retard games. The typical live pro is exploitative, basically because they don't know any better and because they lack good fundamentals. And I think we can agree that the typical live pro is way way lol.

Of couse we know that none of the good players you mention are either style, we each just fit stylistically somewhere in between. And I think we all move from one style to the other when appropriate.

Tyler said...

quit saying 'game theory optimal'. you sound like a yokel. "nash equilibrium" has 1 million hits on google versus "game theory optimal" which has just 70k which all deal with poker virtually.

also, all zero sum games have a nash equilibrium.

bellatrix78 said...

Agree with the sentiment, but disagree that it is an "either/or" matter.

"Most of the time the most +EV play in a good mid stakes game will involve deviating from the GTO book"

For example your sentence above is not true. There are quite a lot of plays and I would even say that most of the plays where GTO and exploitative are the same (after all, there are only 3 options in LHE on each street anyway).

Exploitative play is narrowing your opponents ranges, GTO play is keeping yours as wide as possible. I don't see why you can't do both at the same time...