I am sure that this question will show how little I know about holdem, but I was looking at some of the Game Theory Optimal (GTO) and Nash equilibrium calculators that SunPowerGuru posted, and I am not sure I understand the utility of these tools for actual poker. I understand that they introduce mathematical elements to range building and decision making, and I understand that GTO poker is intended to be unexploitable by making variable decisions based on your range and your opponent’s perceived range and introducing elements of probability into your play based on the value of your particular hand within a particular range and situation. I also know that the balance in poker is between playing exploitatively (taking advantage of opponent weaknesses) versus playing GTO. But what I don’t get is the actual utility of all of these calculations and ranges. It seems like our brains, as players, are already conducting these analyses and trying to add so much additional information can disrupt the thought processes and situation-specific judgment that lead to good decision-making.
For example, if I have AQs preflop and raise 4x from UTG+1 in a high-stakes ring game and get folds around to the small blind who flats, I am already constructing a range for them based on the information available (their rank, their position, their previous play). Outside data about likely ranges and how my hand performs against them would seem superfluous. I already know how my hand is likely to perform against them if they are likely to call with wide ranges (suited-connectors, small pocket pairs, etc.) and if they would 3-bet or flat with premium hands. Board texture postflop will then determine whether they could have draws, draws+pairs, sets only or possible overpairs to the board. I then have to make the exact same decision as I would in any hand, whether or not I think I have the best hand, and whether I can get them call with worse or fold with better. I understand the ways in which game theory decisions are introduced. For example, if I missed the flop on a K high board and my opponent check-raised my CBet, I would fold because an unpaired AQs is one of the weakest parts of my range and they likely have a very strong range given that they are raising when I have shown nothing but aggression, unless I think they are likely to have bluffs (and it would be pretty foolish to have a lot of bluffs in this situation). Or, I might not CBet at all with this weak hand and some very strong hands so that I protect my range when I miss and give my opponent opportunities to try to bluff the turn when I have a very strong hand (plus I give myself a free card). I guess my question is, how does introducing a whole lot of math about specific ranges and hands (which may or may not actually apply to a real-world situation given that the computer cannot possess a robust set of information about your current opponent) add anything beyond what memory and some equity calculators already give you?
Since GTO theory assumes your opponent will always counter with the optimal strategy it can be useful in discovering the types of situations where your opponent deviates from perfect play. Once you find their “leaks.” it’s more profitable to switch to exploitative play.
Some of these deviations in your opponent’s strategy aren’t all that intuitive, so it’s nice to have the statistical data to go by. If you run enough hands, you get a better feel for what’s actually going on. Consider the Monty Hall problem you mentioned in another post… it’s not intuitive until you actually work it out, but once you do, you “own” that principal and it seems obvious.
For, me, GTO and Nash aren’t end goals, since they would ultimately read to a 0EV situation. I use them to find the exploitative optimal sweet spot that will give me the best return.
1 Like
Here’s a simple example from the game of rock, paper, scissors which will illustrate the point…
GTO would suggest that you throw rock 33% of the time, paper 33% of the time, and scissors 33% of the time. If your analysis showed that your opponent throws rock 50% of the time, paper 25%, and scissors 25%, the exploitative optimal solution would be to deviate from GTO and throw paper 100% of the time.
You would then win 50% (when he throws rock), lose 25% (when he throws scissors), and tie 25% (when he throws paper), giving you the best possible EV for his strategy.
1 Like
Well, I try to bet a lot when I have high cards and that way I do not hafta do any arithmatic
And following up with this line, lets assume player 1 is forced to throw rock half the time and its not just a predilection that you picked up on. In this case, SPG’s stated counter-strategy of throwing paper 100% of the time is still valid. However, a good opponent will now counter this move by throwing scissors every time he is not forced to throw rock, resulting in a 0EV situation once again. Move and counter-move goes on ad-infinitum, always resulting in a 0EV proposition. Not a big deal except for the fact that in poker, now everyone is failing to cover the rake.
My understanding of GTO poker is that it was meant to create essentially a perfect defense, or baseline. Someone once described it to me as the way you defend yourself from being run over. By knowing GTO, you become harder to exploit while also making the sub-optimal play of others easier to identify and therefore exploit. Without knowing what the baseline is, we cannot possibly identify our own leaks to plug or the leaks of others to capitalize on.
All of this being said, I am not close to proficient in GTO poker. It is counter-intuitive for me and requires adjustments to my game that are easier thought about than executed. When starting to learn GTO theory, I was told this was “normal” and that the people who had the most trouble adapting tended to be the TAG players. Yeah me. Just so lucky.
Here is a good site to read up on GTO and to go through some “brain-teasers” that show how and why GTO is valuable, and not always intuitive: http://blog.gtorangebuilder.com/
Lastly, from what I understand, GTO is not the most profitable strategy against the types of players you are likely to encounter here, or on micro/low stakes games anywhere. Here is a really good article from back in 2014 that hits on a lot of the questions I think Joe was asking in the original post: GTORangeBuilder Blog: Unexploitably Exploitative: How to (Out)Think Like a Fish
1 Like
There is zero value in playing GTO on a play money site, frankly. The greatest value to be gained is by having a keen understanding of ranges and how their equities change when you add or remove certain combinations on the margin. This will allow you to construct sound preflop open/3-bet/4-bet/5-bet ranges, for instance. Trying to protect your check range by placing your nutted holdings there is losing a ton of value against players who are primarily concerned with their two cards and not your range at all.
Of course, if getting good at poker is the end goal rather than Replay chips, a solid understanding of GTO is necessary. The human brain is computationally speaking a joke. Poker is at its core a mathematical game, so it seems bizarre to suggest that math is somehow unnecessary. Given how complex the NL game tree is, solvers are an absolute must if one wishes to even approximate what the strategy in Nash equilibrium should look like.
2 Likes
Thanks for the responses everyone. I was just trying to understand what information would be gained from these particular programs beyond what equity calculators provide. Now I understand them a little better.
As you said, 2outs4u, on Replay it is more profitable to go for value and exploit the weaknesses of other players who do not adjust. The more I have played at 20k/40k ring, the more I realize how exploitable even the highest ranked players are. So, maybe I just haven’t played in a situation where GTO is necessary.
And after all that Maths , the other guy is having a bad day and has had an argument with his partner and then just bets because he is fed up and don’t really care or he has had one to many beers and just feels like causing trouble and BETS and then catches the River and beats you, !!! So Bang… Math goes out the window along with your chips…!
This video kind of sums up my questions about GTO poker. Game Theory (GTO) vs. Exploitative Styles? The GREATEST Chess Player in the World Weighs-In! - YouTube
It is obviously useful to know the underlying math, but since the human brain is trash compared to a computer, it seems like trying to play a GTO style can take away from good decision making or cause you to lose value by going for balance. It is also difficult to know if you are actually playing in a balanced manner. Another thing that Torelli brings up is that it can be used to justify any decision in a particular case because it is most relevant for tough decisions where any action should be made some percentage of the time. He also says this video is particularly relevant to live poker, but playing on Replay, there is plenty of information given out by opponents about skill, hand ranges, and betting styles that can used exploitatively. GTO definitely seems important, especially when playing against high level opponents over a large number of hands.
Right. It’s good to understand the basic principals, but not a lot of help in day to day play, especially since you need to have as much information as possible on a player.
I stick to the medium MTTs, and see the same people over and over on final tables. I also see the same names on top of the leaderboard month after month. I have fairly extensive hand histories on 6 or 8 of these players, and used a GTO-type analysis to find their weaknesses. This allowed me to develop an exploitative optimal strategy for each of them.
It’s a never-ending struggle though. I’ve noticed that a few of them have already adjusted their play in order to exploit my exploitation! Hahaha, such is poker, huh? Frustrating as this is, it has helped me tweak my game a bit, so worth the effort, I think.
Basically, it comes down to this… you have to know the rules if you want to break them successfully.
1 Like