Post #4k: Please Humor Me While I Ramble About the Forum (and Probably Game Theory)

I agreed with just about everything you said but this practical advice is particularly worth highlighting. I agree. I did lots of finance exams a few years ago and gained nothing from skipping straight to the answers. I think the best way to use solvers by far is to think hard about a spot and what you would do and why before ever looking at a sim. Only then should you look at what a solver has done differently and (only if the EV loss is worth the hassle) why. Trying to memorise solver outputs without doing this was just a waste of my time.

I think this is true and probably the golden mean for 99% of us. I enjoy my day job but there's only so many hours I can or want to work/can expect a return on the extra effort. I also have hobbies and poker is the main one. It's not unrealistic to think that like most hobbies I will actually make money from an afternoon of it if I keep studying (something I enjoy doing anyway).

But I disagree with the idea that anyone with a day job who spends hundreds of hours a year on poker should be working in that time instead. The opportunity costs argument doesn't justify that for most people.

Hi Vernon, I can tell you put a lot of thought into your post and you definitely say a lot of things that make sense.

I am definitely a much newer school player. I have played poker since I was a kind, but only started playing in cardrooms, casinos, and online since 2021. I am 33. Last year I quit my job after I made almost a year's salary in 400 hours of live poker and I have been playing poker full time since August 2023.

The ideas of implied odds and reverse implied odds are captured in solvers. Example of reverse implied odds: Not opening lowest pocket pairs from earliest positions, not continuing often with 98s facing a 3bet, not continuing all flush draws, especially unpaired, non-nut, out of position on the turn. Example of implied odds: Calling overbets sometimes with the nutflush draw, recognizing EV can come from hitting flush over flush, sometimes we have the best hand with A high and our opponent gives up, etc.

Solvers when used correctly actually can be extremely helpful to climb up the ranks in poker faster than others, especially in live poker where solver use is not as widespread. I went from being a winning 1/2 player to winning somewhere between 120-180 per hour with my main game being 5/5 usually with 10/20 double straddle on in just 2 years and a lot of that comes from studying with solvers. I am confident I am in the top 10 poker players in my city that had a bit over a million people in it and have the least amount of hours playing of anyone else in the top 10.

Some of being good at poker comes from having a certain natural skill and disposition and figuring things out that you don'tneed a solver to tell you. Even when playing $20 buy in tournaments with friends and from my first days playing 1/2 in cardrooms, some things were obvious. Play tighter and win more. Sometimes your opponents line doesn't make any sense because they are bluffing. Sometimes they are never bluffing. Sometimes your opponent almost never has a hand that can call when you fire a big bluff. You don't need solvers to be able to pick up on any of those thongs.

However, having worked with solvers, I have some skills that give me an edge over opponents. A lot could be said about this, but one of your key points has to do solvers not telling you how to play vs suboptimal opponents. First, you can nodelock solvers to play suboptimal. I won't get much into this, but it can be done and the solver can literally tell you how to play vs suboptimal play.

Second, after studying solvers a lot and observing player tendencies, you pick up on spots where you know that your opponents will deviate from the optimal play. For example, you know how often you should get check raised in flops and how often you should get called on flops - how wide opponents should continue vs you. If your opponent folds too much to cbets and doesn't check raise enough, you can cbet bluff at very high frequency.

Other times, maybe you know opponents overfold to turn overbet as they are folding many top pairs and don't have as many sets or 2pair in their range because of preflop and less aggressive lines on flop. So you bluff turn.

Another example, I may know KTs is supposed to be a pure continue vs a 3bet, but I know a 3bet range is supposed to have hands like A5s, A4s, 65s, which I have no reason to believe villain has. So I overfold.

Other times I have to play more like solvers vs other pros or even adjust to someone who 3bets too much. Maybe KQo and suited connectors are supposed to mix 3bet and fold, but an aggressive villain 3bets far more often than not. So I can 4bet a lot vs this player.

Still other times, when you know villain is underfolding and underbluffing, because you know they won't have the proper ranges from prior streets and aren't finding the unintuitive bluffs, you should be check folding way more and not bluffing/hero calling.

You can pick up on the spots where people are deviating from how a solver would play and change your strategy accordingly. This is playing the player. This is the, "it depends." This is exploiting. Solvers should not replace any of the concepts good players have been implimenting for decades. But they can help us figure out how to play some less intuitive spots. For example, say in a rakeless 100bb cash game CO vs BB single raised pot on JT6r, you can do a ton of overbetting with your overpairs balanced by gutshots and even total trash hands like Q5s with a backdoor flush draw (1 over, 3 to a straight, 3 to a flush). And your opponent is supposed to continue with basically any 6x, even 62s with a backdoor flush draw, even K9s and Q8s with a backdoor flush draw. Bet 150% pot on this flop and you will get a ton of overfolds. That isn't something I would know without solvers.

Another example of how solvers help me to think about boards. If it is a single raised pot and the villain is competent, 3betting AK pre, etc, if there are an A and K on the board by the turn, and villain didn't check raise me on the flop, if I overbet the turn, they are almost always overfolding this spot.

In the two above spots, I might even go so far as to say that I would drastically overbluff those spots and it will print money.

Regarding multiway, there are multiway solvers. But even if we don't study them, it helps just to know some of the theory. Play tighter, check more, continue less, when you bet flops bet smaller.

One last note. I ****ing love poker. I love studying with solvers. I have shoveled out a decent amount of time and money working with solvers. But I have also reading and posting on these forums, reading strategy websites, watching YouTube videos. Not everyone who tries to use solvers is going to have the results I have had. If Solvers are not your thing then don't bother with them. But if you find them super interesting m, you might be like me, really curious and one day you break down and shell out some money and find out you really like them and can learn a lot.

Good post, mlark.

I loved reading the OP; very well written and nice to read 😀

Just my two cents on the discussion on whether or not the average player is that much better now than he was in 2007... Not a native english speaker, so bear with me 😀

I started to play 15 years ago I guess, around ~10 years back it got to a point where I played professionally for a few years, mostly mid-stakes PLO online cash games, and did pretty well in it. Once in a while though, I still played with the same group of friends it all started in a 10c20c cash game. They knew that I played poker for a living and was doing well.

So, I coached two of these guys, free of charge. I was not holding back, I really tried to improve their game.

It was completely hopeless. One of them is actually somewhat smart, but had a huge issue with his ego believing some of the stuff I said to him was probably better than the in my opinion extremely flawed logic he used. He didnt change his game one bit.
The other one was just stupid, there was absolutely no progress at all. Still, both of these guys were somewhat regularly playing both online as well as in the casino. Extremely volatile style, win big with skill, and lose purely by bad luck, obviously, and both obviously big lifetime losing players.

Eventually I quit poker and finished my studies and started to do a regular job, around 2016 or 2017 I guess; then, when Covid came around, I deposited 600 Euros at a very small Austrian poker Site which I ran up to 40k since then as a side hustle besides work, and for fun. My game is obviously rusty, and there are some regs who I try to avoid, but still, the general level of play is still about the same as back then imo. Just less players.

So, to conclude the little rant: From lets say 2012 to now the level hasnt changed much imo, at least in PLO, and from what I can tell, also in live no limit holdem cash games up to NL 5-10. And the reason, in my opinion, is that Poker attracts exactly the two kinds of personalities I was trying to coach back in the day.

1) Poker is a game that requires having money. That usually comes with an older age like 35+, and at that age people usually have a certain set of character traits who are hard to change, e.g. the willingness to accept something you thought to be true to be wrong. That would be friend Nr 1.

2) Poker attracts short term luck seekers, if that is a word. These people are generally not the brighest people and are incapable of applying difficult concepts. They just play and hope for the best, kinda. Probably get a little better over time but are stuck at the losing zone. That would be friend number two.

as a side note, probably the biggest improvement in play was when Doyle published his book. Easy to understand concepts, well written, and working if applied correctly.

Wow. Great post, OP. And congrats on the 4K milestone.

I confess I haven't made direct use of solvers yet, and don't spend much if any time studying charts. But I have spent a lot of time watching YT videos from the likes of Doug Polk and others, where solver outputs and charts are discussed a lot.

I agree that simply memorizing solver outputs isn't as valuable as understanding the WHY behind those outputs. I think if we can understand the logic behind solvers' recommendations, we can rely less on memorization, and more on reasoning in-game.

That's not to say I don't think solvers are valuable tools. I think the best players are those who can combine the best of game theory with well-reasoned exploitative deviations from GTO play. Doing that requires at least understanding GTO, if not investing the time in studying it, whatever form that studying takes, be it getting hands-on with solvers, or simply studying what others have done with them.

But before anyone spends time trying to learn and understand GTO, I have to think it's vital for them to understand and have some minimal competency in, if not mastery of the more fundamental concepts of the game.

Doug Polk has posted videos railing against deliberate GTO deviations, no matter what the situation is, no matter what the reasoning is. There was one video, I think he posted it last year, but has since removed it, highlighting a big hand from the WSOP, where a player holding KK, who bet it the whole way, and who rivered top set, folded (correctly, but not in theory) to his opponent's all-in jam, when that opponent rivered a miraculous inside straight.

Doug's position was that it was a terrible play, according to GTO, and if V rivered a straight, so be it, we can't fold top set in that spot. It's "just your time to die", as Doug put it, "don't fold sets." Obviously he's all-in on GTO. I found myself thinking about that video, and that situation, for a while, contemplating what I would have done there. I came to the conclusion I probably would have also made the theoretically-incorrect yet actually correct fold (forgive me, Doug).

He makes logical and well-reasoned arguments. Playing a 100% GTO style is probably the highest EV over the long haul. But I still believe there are situations in which it's not just fine, but critical to deviate and make an exploitative play. I think there are lots of spots like the one above, in which making the solver-approved play loses more often, and more money, than making an exploitative play.

In most low-stakes live cash games, and most smaller tournaments, the player pool is dominated by terrible players. Even if the best players have gotten a lot better, and even if the average players have gotten somewhat better, we're not making our money by going to war with those players. We make our money exploiting the bad players, and there are still tons of them dumping multiple buy-ins in every card room in the country.

Sorry I let it go so long without responding!

Also, I'm not going to address Mlark's post directly except to say 2 things. First, good read. And second, reading it, I realized, I think the reason I don't want to put the money into a solver can be summed up by 2 things: 1) I already knew most of the examples you're using--only the last one is a new idea to me, which I will think about, and 2) I live in a state where I can play online, and the little time I have for poker is (to my mind) better put to use playing than it is trying to decipher a solver output for a very specific spot I might want to improve on. Maybe before I had kids I would have had time to buy and use one.

So I bolded that part because this is the exact idea that sums up why I've been talking about solvers. Solvers shouldn't be considered "fundamental."

In any other game, you learn "fundamentals" when you start learning. You have certain ideas that guide your play so you can improve, be it a sport like basketball or a board game like chess. In basketball, you can't try to study the triangle offense before you learn how to take a jump shot or throw a bounce pass. In chess, you can't learn 15 moves of theory in the Najdorf Sicilian before you learn how to lock down a winning endgame. These are differences between beginner concepts, and how a beginner should be trying to improve, versus advanced concepts and what you can branch out into AFTER you have learned "fundamentals."

This may be my own bias talking but I believe a lot of what "fundamentally" sound poker is, is knowing how to play tight. Not that you should do it, necessarily--just that you know what being tight actually means, so that when you loosen up, you understand how much you're actually loosening up so you don't go overboard. You should have knowledge of what a "solid" game really means, only so that you understand when you're deviating from that and why. A solver isn't going to teach you that--it's going to play looser than what a fundamentally sound tight player would do, and it's not going to tell you which parts of its play are fundamentally sound and which parts are advanced stuff that you do in response to your opponent trying to exploit you if you're playing tight.

I think a really good example of a concept that shows the difference between theoretical soundness and fundamental soundness is the question of why we include bluffs in our range when we're betting in a spot. In a previous thread of mine I got into a debate with someone (Mlark?) about this, where I remember considering the argument that "the reason we include bluffs is to get our value bets paid off." I showed (or I think I showed) that theoretically this is not true: bluffing more or less doesn't make our value bets more or less profitable (against an opponent who responds perfectly to our strategy) unless we are exactly at the equilibrium point, but nowhere else.

Nevertheless, I agree with this as a fundamental concept. If your opponent calls too much there is no need to bluff (and it is probably -EV), you just stick to value betting strong hands and potentially branching out your thin value range, not your bluffing range. You only "need" to bluff when your opponent is willing to fold to your bets.

Even though I disagree with this concept theoretically I would have no hesitation teaching it to a beginning player (and I have before).

Solvers are impossible to learn. TheyÂ’re tedious, and have way too much information. Also IÂ’m born human to do human things, not to play like a robot.

Solvers are also very counterintuitive, so here kq is a fold to an open, here 23s is a raise to an open.

Solvers are also amazing, phenomenally +ev, theyÂ’re just impossible to learn. Lets say you want to solve your AK preflop, thatÂ’s 9 positions, vs 9 positions, vs 6-7 different actions per 9 positions, relative to each of their other 9 positions and 9 positions actions. ThatÂ’s 400,000 different situations, and you didn't even said what kind of player is doing what yet. Add in 20 kinds of players, that 8 million situations. If you solve that, I guess youÂ’re playing AK like youve been playing AK since the prehistoric era

Using solver in the future will be about extrapolating what the solver shows as being winning priorities, which is in a very gray area right now

You are using the word "learn" to mean something like "blind/unthinking memorization/imitation".
Literally nobody recommends this definition of "learn" wrt solvers[1]. Lots of players, including some high level pros, have argued that this is why GTO is bad/terrible ... so wp on beating that horse even deader.

But as many people in the thread (and other threads) have tried to explain ... just don't use them like that.

Eg. 10 min. video about BTN vs. SB preflop 3bets ... https://www.youtube.com/watch?v=NI9St_H7...
Yes, you might want to remember certain parts of the ranges ... but the concepts you can get are a much smaller thing to remember, and not this super advanced magic only for 100/200+ pros.

[1] The one exception is likely preflop where you just have to memorize a bunch of things. It was kind of amusing to see the recent post in this thread from CallMeVernon which acknowledges you need to "understand" what "tight" looks like.

In your video, it looks like the solver is putting out generally winning plays assuming he’s playing against a solver, which isn’t the case in practice.

I’ve asked around 3-4 years ago what were the concepts that could be sucked out of the solver thinking process and nobody could answer me, if that changed by all means show me those concepts and priorities.

Oh yeah im a whirling death becoming a good person, but I don’t kill horses I swear to you

This read was a trip down memory lane in a few ways.

I was just thinking about your comment about changing to adapt to game conditions. I think this is huge in live poker. Especially in Low and mid stakes where the “discipline” is poor. Players steam, tilt and react personally to the play of their opponents.

Most old time strategists in books (Doyle, Caro, The Boom notables) all used to talk about “switching gears”.

There are various windows of opportunity presented in the course of a typical session. Deviating hard from standard strat is appropriate.

I was also thinking of a poster from 5 or 6 years ago. I think his name was Ben. He woul post these long posts about his strategy which was generally a balanced strat al la Mathew Janda “Applications oh NLH” or Millers 1%. (Pre solver attempts at GTO). I suggested at the time that in 5 years this would become the norm for strat discussion. Largely I think it has. I only recently returned to watching and reading some strat and the difference from 5 or 6 years ago is rather striking.

Along these lines. A huge thank you to whoever started overemphasizing blockers. There are a number of players who misunderstand the use of blockers and actually overbluff in spots. Because they use large bet sizings they are getting folds from nits without nuts but their lines often don’t rep in any way the hands they block.

Regarding solvers… I had started playing around with Card Runners EV back in the day. I recently have been looking at solvers again. Someone mention node locking and I think this video does a nice job of showing how node locking can be used to develop an exploit strat.

I enjoyed the read Vernon.

https://youtu.be/SXOIH8c7Vd0?si=G6c-lRTN...