Hello fellow poker players,

I really want to learn a more general GTO approach to the game of poker eventually learn GTO to the level of unconscious competence. So I decided to look on this forum, youtube and reddit but I could not find more general approaches to learn GTO almost fully heads up in SRP pots. Most videos are about specific situation and not about how to learn GTO in general in my opinion. Now I come with my own approach and I want to know what you guys think of my approach and if there are improvements or flaws or that I am just being an idiot.

My approach:

- Master the 25 flop subset from Simple poker.
- Use GTO+ to play against the solution and learn while playing.
- Play from every position vs the other postions so (aswell as IP as OOP), BB vs SB, BB vs BU.... HJ vs LJ etc.
- Use the preflop GTO ranges from Zenith poker (because they are free and I -do not have alot of money to spent atm)
- Every day I will practice about 4 hours a diffrent scenario untill I do not make more than 2% mistakes, then I will move on to the next scenario. I think it will take me 4 hours (dividend over the day) to master a position down to making only 2% of the time a mistake (Its just an assumption I could be enterilly wrong).

This will approxiametly take me (if I am correct): 25 (flops) x 16 (postitions) x 4 (IP aswell as OOP) = 3200 hours and I will be done after 800 days (around 2.2 years).

Maybe it will get easier over time and it will be maybe possible to learn 2 or maybe even 3 scenario's a day. Which means this can maybe be doable in about a year or so.

Please let me know what you guys think and if its somewhat a reasonable approach to learn GTO.

Just to clarify, since I am getting no replies and after reading my own post again, this is what I mean:

My goal is to master the 25 flop subset from simple poker by exploring every one of those flops for hours with GTO+"play against the solution" in every singe scenario, so IP and OOP and in every position vs position E.g BB vs SB, HJ vs LJ etc. The goal is to think in terms of GTO to the level of unconscious competence, so when a certain scenario appears you know subconsciously what to do. This principle comes from chess where, especially in faster time controls, you know instinctively what to do. In poker this is very important because its not humanly possible to think rationally about every single spot.

I'm doing similar thing, i think your idea is fine. If you want to be practical you shpuld pay more attention to common spots and less to uncommon ones.
Btw playing vs GTO+ will not teach you frequencies but it is really good way to learn what are reasonable lines. I don't know what you mean by mastering one scenario?

Well, with a scenario I mean for instance: HJ vs BB SRP on a 3s3dKs flop (1 of the 25 flop subset from PIO).

I think studying like this will make you pretty amazing at poker as long as you understand why the bot makes the decisions that it does. I would also recommend testing yourself, perhaps master 23 of the flops first then see how well you do on the last 2 flops. If you do well, congratulations, you can probably do well on any flop. If you do poorly, then perhaps you only memorized data rather than concepts.

I've been out of poker for years and I also had this approach in mind to revamp back. I'll be using the pio subsets(idk if there is a diff between simple)

I couldnt really find the subsets of Simple Poker so I decided to use the pio subsets aswell. Only problem I stumbled on is that the subsets represents the 22100 flops pretty well for pre flop solutions, but appereantly not for postflop solutions (thats atleast what I read on this forum in a few diffrent posts). Let's take a look at the first flop from the 25 subset for instance: 3s3dKs. Lets say we are in a SRP LJ vs BB, the LJ will have around 57 combo's of Kx including AK (atleast in my LJ RFI open), now lets say we have a very similair flop of: 3s3dAs. This changes the amount of top pairs in our range drastically, because now we have 99 combo's of Ax and this changes our strategy significantly. So to use the same strategy of Kxx flops on Axx flops or even Qxx flops is not going to be very GTO at all. I wonder how others who study like this do this? Does anyone have a good flop subsets that will be good for general GTO post flop solutions?

You can never really "learn" GTO because each situation is so unique. I get what you're going for though. I think the strategy you described would work, but it's overly complicated.

Here's what I would do:

- Watch this video about how to solve subsets in GTO+: https://www.youtube.com/watch?v=2dMsWUeMc-M&t=1144s. Then solve the 89 flop subset for BU vs BB, SB vs BB, CO vs BU, and UTG vs BB for single raise and 3bet pots. Save those solutions in a separate folder.

- Play online and stop every time you have a question about a spot. Then find a similar spot in your solved subsets and try to figure out why the solver is choosing its strategy.

Example: You're playing 2NL. You open 99 in the CO and BB flats. Flop comes JT2r. You're not sure whether you should cbet. Open your BU vs. BB sims, choose the closest flop you can find (e.g. JT4ss) and see what the solver's doing.

Rinse and repeat. It's not perfect. But until you get to high stakes that's about all the depth you need.

Honestly 2.2 years is a huge commitment. You'll be a much better GTO player regardless after that much practice.

My 2c: Don't try and limit yourself to 25 flops. If you're going to spend this much time practicing you might as well get a feel for a much wider range of flops. Build your intuition rather than trying to memorize the answers. You'll find that many of them play the same way anyway.

The drawback of your strict approach is that you won't really learn WHY certain plays work. You won't develop an exploitative sense. After you've mastered some spot, it can be good practice to change one of the ranges and see how the strategy adjusts.

First of all thank you all very much for your very thoughtfull and insightfull responses! I tried to use each of the advice you guys have given me, and updated my learning plan.

- Here is an update on my plan:

Firstly, instead of the PIO subset I will use the 44 flop subset from GTO+. I chose this subset because it preforms really well in average error in overall EV and average error per hand (source). I will then sort these flops into diffrent groups such as LLL (low low low) flops, LLM etc.

Secondly, I will practice the most commons spots first and more thorough than spots that happen less often. So this means practicing the most commons flops (the most common flops are LMH, LHH, LMM,LLH and MHH, source) , the most common textures such as rainbow flops, etc. , and most common scenario's (which are SB vs BB, BU vs BB etc.).

Lastly, instead of 4 hours I will try to do it in 3 or maybe 2 hours for both IP as OOP to save time. Especially the uncommon spots I will take less time studying them because there is enough to learn in poker...


- Here is my studyplan:

1. At first I will test my knowledge about a certain spot.

1.1 First I will look at how a flop affects the IP and OOP ranges.

1.2 Then I will decide which range has the nut and equity advantages and how this will translate in terms of EV: so does a board overrealise or underrealise IP or OOP, or is roughly the same as the overall equity.

1.3 Then I will look at how these previous thoughts will translate in terms of bet sizing and hand selection: which hands have relevant blockers or front/back door equity and which of those do we want to bluff with? Do we want to bet or check our entire range? Which hands do we want to bet for value and which of the value portion will protect our checking range? Will the OOP player have a 2bet range or only a calling range? Do we want to overfold as the OOP player to make the IP player indifferent? Etc.

1.4 Then I will think about how turns and rivers influence the actions and which of the turns and rivers are good for the OOP range and which are good for the IP range. Then I will think about the bet sizes on turns and river and if my previous thoughts about bet sizes and checks on the flop are in line with the lines I want to take on later streets.

2. After that I will check my results/thoughts with the solver to see where I was right and where I was wrong. Then I will reason why the solver does things diffrently than me, if I do not get why a solver does a certain thing I will post it here on 2+2 and try to get answers from others.

3. Lastly I will practice the "learned" theory by playing against the solution on GTO+.

3.1 First I will try and geuss the frequencies while playing against the solution. When I am wrong I will go back to the solvers answer and see why I was wrong. I will do this untill I make very few mistakes.

3.2 After a while I will come back to this spot and practice it again to see if I learned it to the level of unconcious competence.

What do you guys think of this approach?

Your plan is very good and more structured than the plans of most people that play the game. Careful now not to focus too much on the details and actually start putting in the work. Follow your plan for 2 weeks and then you can iterate on it. Good luck.

Playing against the solution in GTO+ is suboptimal, because when strategies are mixed, the EV will be the same. You often just end up clicking through without learning anything. I would recommend a different approach:

1. Take a larger subset, like 80+ flops, and do a very simple tree with 2-3 flop sizes and just 1 size for every street and 1 raise size for every street. I'd recommend something simple like 30% 67% and 125% for single raised pots. This will allow us to fine tune our flop strategy. Learn the global frequencies, and then look for outliers. Know what types of boards warrant a certain sizing scheme and frequency. Most importantly, think about WHY the solver is using a certain sizing scheme. It has to do with how the ranges interact.
   
2. Once you have the frequencies and sizes down cold, take a smaller subset and build out more complete sims to practice with. You can simplify the flop based on what you learned in step 1. This will make the trees a bit more manageable. I would recommend not using play against the solution initially.
   
3. Really learn what your continuing range looks like vs turn probes, as you will face them often when you check back flop. Come up with some heuristics you can use in game. If you want, you can now use play against the solution to drill this spot.
   
4. Do the same thing for turn barrels after you bet the flop. Learn how the flop strategy changes your turn strategy. For example, if you are range betting small on the flop, you will often polarize your range on the turn. If you are betting bigger less frequently on flop, polarizing your range, then you will continue to polarize the turn and check back your air to give up river.
   
5. Now that you have turn play down in both the flop betting and checking lines, you can start to pick up river play. I would think about how your pool plays and really alter your turn and river play based on this. As an example, the solver may take a line where it bets small on the flop, overbets the turn, and jams the river quite often. In real life, if villains overfold to the turn overbet, their river range is too strong to continue bluffing into.

I hope this helps.

I just want to emphasize what last poster said. Don't play against a solution because a lot of spots are mixed.

For example flop strategy mixes ATo 70/30 bet/check, that means that these 2 options have exact same EV. So every time a hand is mixed you could just choose 1 of the options and play if pure and not lose a single EV versus a gto solution.

Why is it working like this? Because gto doesn't care about your strategy and never adapts.

So why these frequencys matter? They matter in order to stay unexploitable and remember gto is only maximally exploiting the other gto solution and will not exploit such a deviation as above. The only mistake you could make versus a gto is betting for a wrong betsize or choosing an option that is not mixed at any frequency.

In other words after a month you could play solid versus a gto solution and not lose much EV but in real world people could exploit you hard if you, say pure bet every mixed hand.

Solve most common spots, run aggregation analysis, develop strategys, try to find a logical reason-patterns behind a solver and don't forget that a solver cares only about EV and it is not smarter than a chicken.

Noob here, I'm also learning a lot of solver ranges and sizes, but don't we learn it so we know how V deviates and exploit them? Is there a lot of EV to be gained by learning how to stay balance at all times, especially in microstakes/low where V is mostly not?

About your first qeution: I don't think so, we learn so we know what to do against unknown players. GTO sets a baseline so we always know what to do in general. Many of our opponents we dont play more than a thousand hands with, so to actually exploit people we do not have enough hands on is just geuss work. Many of our opponents are unknowns so we should therefore apply a more GTO approach to exploit them, many people seem to forget that GTO is an exploitative strategy.

About your second qeustion: Yes, there is, because when people differentiate from GTO, GTO will exploit them. And if you want to take it a step further you can exploit your field by playing a GTO strategy that is specifically designed to exploit your field even more. Just learn more about nodelocking.


About my studyplan, I updated it again and I implemented MicroDonkYT's response. The focus is now way less on "playing against the solution". I started learning a few days ago with the updated plan. I want to thank all of you who helped me and responded, I really appreciate it!

My 2c is to incorporate nodelocking and Max explo strategies against opponents deviations for instance if we know BvB that BB is floating flop at high freq' we start to check range and play 2b/f. And also spend alot of time developing heuristics 'cos w/ your original approach of grinding play against the solution the numbers and frequencies will converge to static and you'll fail to grasp the core concepts of gto.

Sent from my G8441 using Tapatalk

This is honestly more valuable, as it gives you a good sense of "why" a specific strategy works. Try out different stratgies and see how the solver attacks your weaknesses. IMO, exploring how the strategy changes with nodelocks is a better path to understanding fundamental strategy than rote memorization of GTO sims.

In reality, max explo strategies often carry a significant degree of risk and are fairly obvious to counter. It is important to learn what a reasonable baseline looks like. But beyond that, if you want to understand the "why" rather than the "how", nodelocks are the way to go.

Do you mind updating and sharing your study plan? I'm also looking to seriously up my game and would like to see how your overall plans look like if you don't mind

I honestly think trying to mimick solver through sheer memory/repetition will hurt you a lot.
What actually leads you to learning GTO is learning about the foundations of the game and applying those foundations when you play, it just so happens that looking at solves reveals a lot of foundational aspects of the game if you look critically enough.

At some point you will be able to predict roughly what solver would do by logical deduction + some amount of memory/practice to speed things up. Also, don't play like solver pls. Exploit.

Automating the flop has a lot of benefits though. Just like in chess, you want to have certain things memorized so you can expend your energy in other areas. You can keep it simple and just use 2 or even 1 size in a lot of spots. The solver gives me a good idea how to structure each range and offers a lot of insight as to how your range wants to bet.

So before I studied GTO, I had no idea that on boards like JT2 SB vs BTN in 3bp, we want to have an open rip range with QQ. Now that I know that QQ is an open rip, I can look at the obvious bluffs the solver figures out, and then I can look at the calling range the IP solver has. I can then determine how this works in real life. If villains have some obvious defends they won't have in real life, I shift more bluffs in the range. If I don't really know, I try to play it balanced.

This is a framework I use to study every spot:

One thing to keep in mind is the frequency of different textures also. Two tone flops roughly 55% of the time, rainbow flops 40% of the time and monoton flops 5% of the time. Also boards are most often high-mid-low I think like 25% of the time and only 3% or less low boards so spend minimal time studying monotone low boards and more A92 type things first. I am planning something similar to keep busy and truly master poker.

This, a thousand times this. Which to be fair, is the most difficult part

This process seems like a complete waste of time and effort to me, but if this is what captures your interest, I guess go for it.

It’s unclear to me what this 2% margin of error is or how you will measure it. Is this something that GTO+ provides? If it does provide such a metric, I’d be curious how it measures it given that when you play as the solution, you’re still just playing a single hand, and what will matter most is how you play the entire range. Not to mention that there’s a ton of mixing involved and many actions will have equivalent EV.

Also, it seems to me that you’re significantly underestimating how much memorization will be required and possibly overestimating how useful this will be to the overall profitability of your strategy. If you’re going to master 25 flops, for every formation (SRP, 3bp, 4bp, etc), it would take a massive amount of mental power to do this, and that without even playing the turns and rivers.

Surely you can invest your time more wisely than this, unless you’re gaining a ton of value from the entertainment of it.

This is very good advice

Hey guys,

Thanks for all the responses! I created this post a year ago and have updated the plan severall times because of all the great input from you guys and some other insights.

Rickyt88: Yes I have looked at that and I believe good subsets will make up for that right? I mean they are meant to reflect the 1755 unique flops right?

BigBurge10: I measured this in terms of EV loss, so if I kept 98%+ of the EV in my nodelocked attempt compared to the "GTO" sim, I would continue to the next one. If I scored lower than 98%, I would then try it again later untill I hit 98%.


I don't think I am? I am nodelocking and mixing the entire range, not just 1 hand.

I am not memorizing anymore, I have updated my approach a dozen times now and it more focused on the why (which it already was imo, because if you want to mimic solver play in your nodelocks you have to know why and how a solver plays in similair spots and base your strategy on that). I also trimmed the plan alot, it now takes less times and also focuses on turns and rivers in diffrent common lines (Eg. X/X/X, X/B/C/X/X/ etc.).