A deeper look into the advantages and disadvantages of the 2 approaches towards strategy execution in Poker, and how to decide which is best for you
One of the oldest and most heated debates in the poker community reaches the Poker Made Simple newsletter!
People need to know, after all, what is in fact better: to play a theory oriented style or to try and exploit each and every opponent?
I believe I have a lot of insights to give about this topic. I’ve played a very exploitative style of poker for years with great success. But I’ve also at one point completely shifted my approach, literally overnight, and started randomizing every single decision at the tables – and continued to win a lot. I’ve studied the games of all the best players in the world, their tendencies and their deviations from theory; and I’ve also wrote actual code for GTO bots and simulated hundreds of thousands of hand histories to study GTO behaviour. I have personally taught hundreds of students myself, on both styles, and I have seen their development and their results over time. So now I want to bring all this knowledge together and help you determine, for your specific case, which approach is likely to give you better results. As I said when you signed up, I’m here to make poker strategies simple so that you can make more money. And I take my promises very seriously.
Let’s dive into this.
MY PERSONAL STORY – GTO x Exploitative In practice
It was August, 2019. After crushing 200z for over a year and building a solid bankroll, I started taking some shots at 500z, with the goal of turning it my main game. I remember feeling super excited about that move up, to the point of taking multiple screenshots of me playing against some of my idols, like Otb_RedBaron, ishter11 and king10clubs:
Maybe these names don’t mean much to you if you weren’t playing online poker 5 years ago; but for me, someone who in 2016 was playing 5nl zoom while following and reading every poker journal, blog and PGC in every poker forum in the internet, these guys were the real deal.
Otb_RedBaron was still considered by many the best NLHE player in the world – he made millions at high stakes from 2012 to 2017, then took a bit of a break and would show up occasionally at 500z in 2019; ishter11 was one of the sickest zoom grinders of that time (and perhaps of all time), someone who was printing $200k-$300k a year playing midstakes zoom pre-rakeback; and king10clubs was also a legendary long time grinder, one of the biggest winners at Pokerstars 5/10 games, and also someone who would print $200k+ a year playing online poker.
I “grew up” in poker idolizing those guys and only 3 years after I started, there I was having to battle against them in what was considered by many the toughest online game in the world that would run everyday, 500z on PokerStars. It’s like there was a golden aura about that game, and I wanted to prove myself.
During that period, despite crushing 200z for almost 6bbs in a very large sample, there were a few things about my game that I was unhappy about. I was playing a very exploitative style, which I had built by studying population tendencies, comparing their frequencies to solver frequencies, and building max exploit models to attack the imbalances found. Everything was working great overall, as my results were showing me, but I felt some people were starting to adjust against my strategy.
There was one particular line I was extremely imbalanced at, which was the triple barrel in single raised pots in position. I remember looking at my range composition for that line a few years later, and I had almost 60% bluffs in my betting range LOL. This was of course intended, as that line is one where regs were (and still are) overfolding substantially relative to solver. My approach was then to be very aggressive and overbluff whenever I had the chance. Not only that, I was also bringing more bluffs from the turn to the river just so I could exploitatively over barrel the river.
The problem was that I started to notice some lighter calldowns than usual against my triple barrels. Some people I perceived as weak started showing some resistance. And some hands I clearly expected people to always fold were finding their way to showdown against my bluffs sometimes.
Initially I didn’t give this much attention. I noticed it was happening, but I kept executing my plan. After all, I had confirmation from my results that my approach was working. I had studied millions of hands of data, and I saw with my own eyes that people were overfolding. I was doing the right thing by overbluffing.
However, an increasingly annoying cognitive dissonance started to grow inside my head with every bluff gone wrong. In one hand, I had the belief that I was doing the right thing, that I needed to be aggressive and bluff, because I did the models, I saw that it was the correct exploit, AND I was crushing the games. On the other hand, I was more and more often getting called by super marginal bluffcatchers and seeing real evidence that perhaps my exploit wasn’t working anymore. “Which one is right? Should I change my strategy? But isn’t it working though? But aren’t people calling me?” Confunsion, confusion, confusion.
This confusion started to spread to other exploitative lines I was taking in other spots of the game tree. I started second guessing myself and doubting the effectiveness of my overall gameplan. At the same time, feelings of my strategy being somewhat inadequate also started to grow in my head. For example, I used to cbet entire range on the flop in most textures, under the idea that it was a good exploit vs a pool that doesn’t check-raise enough, and with an added benefit that made my strategy simpler to execute. But I felt like I was playing below my potential and capabilities, because I had a lot of theoretical knowledge and I felt confident enough to pick the best betsizing for a given texture. I was forcing myself to play in an overly simplified way when I was fully confident and capable of going for more complex executions.
On top of these feelings, I had the desire to move up to 500z, a game where I believed my opponents would be substantially stronger than the average 200z reg. While at 200z I could make money with my eyes closed, at 500z I would be playing with many long time high stakes players, people who make a lot of money with poker. With my confidence decreasing in a highly exploitative approach, I felt there was no way I should play the same style at 500z. It had to be different.
I’ve always been someone ready and willing to make drastic changes. I don’t hesitate when I believe a different path is the way to go, even if its a 180 degrees turn. And that time it would be no different. With all those feelings combined – sensing that people were adjusting against me, feeling like I had trapped myself in a box and capped my own potential with a fixed execution, and wanting to move up and play against tougher players – the shift was obvious to me. I had to try and play close to solver.
And that’s exactly what I did. Literally overnight, I went from playing a highly aggressive exploitative style with no randomization to throwing the RNG for every single decision at the tables, while trying to play as balanced as I could.
Take a look at my stats from that time. In August 2019, I was playing 42/35/34 in aggression frequencies postflop:
In September 2019, I shited my approach completely, and my aggression stats went down to 38/30/30:
38/30/30 are not good stats for someone playing pseudo-GTO. A GTO bot will achieve something close to 38/32/39. But the point was that I allowed myself to completely change my approach and my execution at the tables, even though I was winning a lot. And to my pleasant surprise, it fucking worked. I continued to win and have an edge over my opponents, even though I wasn’t trying to exploit any weaknesses in their strategies. I knew that this could be the case theoretically, as a GTO strategy is obviously profitable against imbalanced humans. But to make that shift, in such a drastic way, and immediately have results with it…safe to say, my confidence was at an all time high.
You can see from the print above that I was mixing 200z and 500z in September, the month where I made the shift. I didn’t want to jump straight into 500z with this new approach without some practice. But once I had those results in September, I made the full transition in October and these were the results:
Slightly better pseudo-GTO execution, with 38/30/33 aggression levels, and again a crushing month. I had turned myself into a pseudo-GTO oriented reg overnight, and I was still beating my dream stake, 500z on PokerStars, relatively comfortably against a selection of some of the best players in the world.
I was in the clouds.
Lessons from an Aggressive Semi-spewy Reg turned Passive pseudo-GTO regbot
This experience taught me incredible lessons, even things that I was only able to make sense years later. As I matured as a poker player, gained more and more knowledge, played more hands and taught more students, I was able to see for myself what generates edge, what style is the best, and how to execute both of them. Let’s talk about those lessons.
Lesson #1: Playing Pseudo-GTO is easier than playing Exploitative
This was fucking surprising to me. Prior to my transition, I had this inner belief that playing a theory oriented style would be super difficult, and that everyone that tried to do so were wasting their time. What I saw in practice however was that executing that style was way easier than playing the exploitative style I had been following.
When you play a highly exploitative strategy, one of your biggest concerns is game dynamics/metagame. If you want to be succesful playing an exploitative style of poker, you need to do many things that consume a lot of time and energy:
- You need to keep track of the history between you and everyone else, and use that recent history to make your decision about your next move against that player;
- You have to take notes on people’s showdowns to make sure you know when they are deviating against you;
- You have to actively think about how you’ll adjust your strategy in the case someone is targetting you, which requires reframing your mode of thinking;
- You have to check HUD stats to see if the player you’re facing actually behaves like the average guy in the player pool – then if he doesn’t, you need to come up with another strategy against him on the fly;
When you’re playing pseudo-GTO, all you care about is your own strategy. You make your best approximation of theory for that spot, look at the RNG and execute the option the number tells you to. It’s fucking easy. I still remember to this day how much less tired I felt after an entire day of grinding when I made the transition to the pseudo-GTO style. It was like a night and day difference.
Now, it should be made explicit that despite being easier to execute as a framework, it’s not like you’re not gonna make mistakes. Your execution will be as good as your capacity to approximate the theory for that given spot, so it’s definitely still hard to actually play close to solver. But the process of trying to play close to solver is much easier than the process of trying to exploit everyone.
This is a big pro or point in favor of playing theory style, because it’s (also) when we get tired and overwhelmed that we let our emotions take over and big mistakes happen. By getting less fatigued, you can sustain a high level execution for longer, helping you achieve a higher winrate.
Lesson #2: Your approximation of theory is likely way more passive thant it should be
I was able to see this immediately in my game. When I tried to play close to solver, my game became very passive, way more passive than solver itself.
Years later, I was able to see the same thing happen to my students. When I coded a “Grind Simulator” in 2022, a program where students could play against bots to practice their GTO execution, I witnessed players who in real games were playing super aggressive styles with positive redlines become 47 WWSF nits when trying to play close to solver.
For a bunch of different reasons, technical and psychological, our intuitive perception of a balanced strategy is significantly more passive than the actual balanced strategy. People who try to emulate solver end up bluffing less than they should, raising less than they should, folding more than they should, and betting smaller than they should.
This doesn’t mean that an overly passive gto-oriented style can’t win. In fact, I proved the opposite with my own hands – I played like a pseudo-GTO nit and was still able to beat 500z. And I’ve seen similar regs achieve the same multiple times over the years. But the fact is that theory oriented players are still very exploitable. This is a fact, and in my opinion it will always be the case.
GTO thresholds and strategies are often times way too risky for our human taste. Anyone who has studied with solvers enough has seen a crazy bluffcatch or a crazy bluff in solver land that anyone would call a spewy play if executed by a human. Calling down with super marginal hands against uncapped ranges, going all in for huge bets with pure air, overbetting flops and turns with no equity no blocker – GTO execution often doesn’t fit within the box of a “reasonable strategy”, as perceived by an average player. And worse than that, if the conscious mind caps the execution by not being capable of understanding the merits of certain theoretical plays, the unconscious mind heavily caps the execution by not allowing big risks for seemingly little return. It’s what our primitive brain evolved to do – big risks are only acceptable with big returns. But balanced strategy requires lots of big risks for literally no return. The inevitable consequence is that anyone trying to play really close to solver will end up being a bit too passive and will fold too much.
The lesson here is that if you decide to go with a theory-heavy approach, realize that it’s very likely you’ll make yourself exploitable in very obvious ways. You will try, and you will think that you are doing a good job, but the truth is that you will still have most of the same leaks of the typical weak reg of any giving limit. Someone that understands your mode of thinking will be able to spot several imbalances in your game, and you will bleed EV to them.
This is not necessarily a point to playing exploitative, because you actually become way more exploitable when you’re trying to exploit other people than when you are tring to be balanced. Nonetheless, it needs to be stated that you won’t be able to play completely balanced, and therefore the argument used by many of being theory oriented to avoid getting attacked by others is weak at best.
Lesson #3: Playing a fixed GTO style can be very boring
It’s true that playing a theory-oriented style is less fatiguing than an exploitative style. But a framework of execution that only cares about it’s own actions and only responds to a random number generator can get quite old, quite fast.
After playing a very rigid GTO style for a few months, even though I was winning I started to feel like I wasn’t having much fun anymore. I always enjoyed playing a lot of hands, but once I could make money by just loosely approximating theory with the aid of a random number from 1 to 100, grinding got a bit boring for me.
This take is a very personal one, and I would imagine there are lots of people that can find the experience of trying to emulate solver as best as they can very fun and engaging. But it wasn’t the case for me. I kinda felt like poker lost its color; it became this grey activity that barely resembled a game. It was all about clicking buttons at appropriate frequencies. I didn’t like that at all.
Playing exploitative, on the other hand, creates several layers of engagement and fun in the game through the dynamics with the other players and the constant attempts to outplay one another. Despite tiring, the process of developing edges against other players and seeing those edges materialize when you can sucessfuly exploit someone is very rewarding.
Point to playing exploitative.
Lesson #4: All styles can win
If you put in the effort to be successful, then you can do it with any style of play.
I’ve experienced this myself with my own story; I’ve studied many crushers of the game with significantly different styles of play; I’ve coached highly exploitative, positive redline players that crushed; and I’ve coached theory-oriented players with super losing redlines that crushed.
There are many ways to find edges in poker, and you should never assume there is only one particular way of achieving success. The possibilities are numerous.
CONCLUSION
There are many other topics, pros and cons that I could list here, but regardless of how many I would include, the conclusion would be the same.
In 2024, it’s silly to have a rigid mode of thinking where you have to choose between one of the approaches. Both have it’s merits, and most importantly, both have their place and time.
Perhaps in a super soft game you want to play as exploitative as you can. And perhaps in a 3handed battle with regs you want to play as close to balanced as you can.
Going deeper than that, perhaps it makes sense to play a very common node of the game tree with a balanced approach, to reduce risk of getting exploited. And perhaps in unique spots that won’t repeat themselves in hundreds of thousands of hands, you can allow yourself to play your hand and follow your instincts.
That being said, what I can see very clearly after all these years playing and coaching is that it’s very likely you’ll be drawn towards one approach a bit more than the other, mostly due to your personality and risk aversion profile. Ideally you should develop enough self awareness to understand which approach suits you better. Once you understand that, then you should work on including the opposite approach more often in your thinking process. As with most things in life, I recommend you avoid the extremes – obsessing about being balanced in all nodes is certainly not healthy and won’t help you maximize EV (while also being quite boring imo); at the same time, trying to outplay everyone in every single spot and making yourself vulnerable to counter-exploitation all the time is also not ideal, and can lead to all sorts of trouble for your strategy and your mental game. When a downswing comes you will second guess every single one of your exploits, and it might be hard to sustain a coherent execution long term.
A framework I’ve followed for a long time and works great for me is to make my execution a bit more rigid and theory oriented on early streets (preflop and flop) and then a bit more flexible and exploitative on late streets (turn and river).
By keeping my strategy mostly stable on early streets, not only I avoid getting exploited on very common, recurring spots, but I also make sure to get to the late streets with reasonable ranges that I have a good vision over.
Then, once I reach the late streets, I can start developing deviations from theory to attack opponents imbalances and capture more EV. Late street nodes will usually involve bigger pots and therefore bigger opportunities for significant edges. At the same time, they will happen less often and you won’t be exposed to counter-exploitation as much as early street nodes.
This approach saves a lot of energy because my execution is very low fatigue on early streets, which represent the bulk of decisions in a session. Then I can use this energy saved to apply advanced exploits in nodes where my opponents are leaking the most.
Whatever you decide to follow, remember to have fun in the process. Don’t overthink it. Don’t take everything too seriously. It’s totally okay to experiment different approaches overtime, test new things and make mistakes. From time to time, make a drastic change just to see what happens. Stability is important in life, but so is change. Be creative. Be open minded. And enjoy the journey.
Thanks for reading. See you next week.
Until then – keep it simple.
Saulo