“It was by far my best learning experience in poker”: an interview with RaulGonzalez
Recorded in 2020, published in 2026. Lightly edited for length and clarity.
I first reached out to RaulGonzalez in 2013. He was one of the best heads-up no-limit players in the world, and we made a deal on my Hold'em solver. Our biggest success, though, came a year later with the draw solver: within six months of getting it, he was among the top three 2-7 Triple Draw players in the world, playing up to 2K/4K. He studied the draw solver for about 1,000 hours, more than anyone else, and I'm grateful for his passion.
In this conversation we talk about what studying looked like before solvers, why a top player can't simply hire a better coach, how he learned a game from scratch with a solver, and what happened when he met another team that had a solver of its own.
Before solvers
Oleg: Since when were you an active high-stakes player, and what were the highest stakes you played before we started working together?
RaulGonzalez: I started playing 100/200 no-limit in 2012, mostly against recreational players. From 2013 I was also playing 100/200 and higher against regulars. The highest I played in no-limit Hold'em was 400/800 full-stack, which was the highest game online at the time, and 500/1K cap with 30 big blinds, also the biggest cap tables back then. In limit games I played up to 2K/4K. About 95 % of my hands were heads-up.
Oleg: You were obviously an amazing player already. How did you learn?
RaulGonzalez: Before solvers, a big part was discussing hands with friends and discussing concepts, like when a smaller or a bigger bet size might make sense. We also ran simple math models in spreadsheets: if we have this much fold equity, how much equity do we need to realise on the next streets, so how wide can we open or c-bet? But it was very rough. It's not enough for a c-bet to be +EV; you have to compare it to checking back. So it was somewhat analytical, but not very structured. Mostly it was discussing hands and what we both felt would be a good exploit.
Oleg: What frustrated you most about studying back then?
RaulGonzalez: That was the time when people started mixing bet sizes. Before, everyone bet around 50 % on the flop and 60–70 % on the turn and river. Then people started playing with different sizes on every street, and that is very, very hard to analyse by yourself. Today everyone has learned concepts from solvers, so they have some intuition about a new spot. Back then it was wild guessing. It was frustrating, and it was very nice to finally see in a solver what's actually correct.
The other thing: without a solver you're estimating a lot, and when you lose against someone it's hard to tell how much is variance and how much is you being wrong.
"We thought it was a scam"
Oleg: How did you first hear about me?
RaulGonzalez: It's actually a funny story. As I remember it, you contacted me or my roommate, I was working very intensely with him back then, through a private message on the PokerStrategy.com community. And we ignored it. At the time, the idea that a solution might exist seemed impossible, so we thought, "This is 100 % some scam."
Oleg: It sounded too good to be true.
RaulGonzalez: Yes. Then you contacted the second one of us, and my thought was, "It's probably not legit, but a Skype call can't hurt." So we did a call where you demonstrated an early version of the solver. I was very skeptical at the beginning and got less and less skeptical as we went. Then you made an offer to me and a group of my friends, and I think it was less than two or three weeks until we bought it.
Oleg: Why did you decide to make the deal and pay the money?
RaulGonzalez: The idea of having a solution was very appealing. It felt like it could be groundbreaking and give a huge edge over the field. And I was stagnating. I had been a top 10–12 player for a while, and I couldn't really beat the players better than me, while the players worse than me didn't give much action. A solver felt like a chance to get better than the next few players and play them for a lot of hands at high stakes. And of course I was impressed by the demonstration. I went from "it's probably a scam" to "it's very unlikely a scam; this looks very legit."
Oleg: And how did it work out in Hold'em? You saw it as a chance to challenge the players better than you.
RaulGonzalez: Shortly afterwards I started playing quite a lot against Durrr, which was very unexpected because he hadn't been online for quite a while. I also started playing Isildur at 500/1K cap and at 200/400 to 400/800 full-stack, which was much bigger than I could play against other regulars or the people I considered better than me. So my focus went to those games rather than to the regulars slightly better than me, and the solver helped against Isildur and Durrr as well. When those games died down, I played two particular players I probably wouldn't have played without a solution, though they weren't keen on giving me much action.
Not long after that I got the draw solver, and the draw games looked much more appealing: higher stakes, softer, and it was fun to start a new game. So for me the draw solver ended up far more valuable than the Hold'em solver. But I would never have had the draw solver without starting to work with you on Hold'em.
Why not just hire a better player?
Oleg: Why did you decide to learn from a solver instead of a human player? In most fields people learn from other humans.
RaulGonzalez: In no-limit Hold'em, when I first talked to you, I was maybe the 10th or 12th best player in the world. You can't really get honest coaching from someone better than that, a top-five player. It doesn't make sense for them to coach you unless they charge a ridiculous amount, and then it doesn't make sense for you.
Oleg: What do you think that price would have been?
RaulGonzalez: I guess something like 4–5K per hour. Later on I coached people myself, people I didn't play against, and I could charge around 1K, and at that point I wasn't a top-10 player anymore. So a top-three player coaching a top-10 player while those games were still running would definitely charge a lot more than 2K. At that price it wouldn't have made sense for me.
Then came the draw games. My impression was that nobody had a solution there. Later it turned out one small group of three players had one, a more basic one. If a game is played at 2K/4K and basically nobody has a solution, the edge you can get from one is very big. And you learn much faster from a solution than from a human.
Oleg: Is that true?
RaulGonzalez: For me, for sure. With a solution, if you have an idea of how some concept might work, you can cross-check it across all kinds of different spots and really understand how the strategy works. A human coach without a solution gives you some ideas, then you say, "This looks inconsistent to me," and he says, "Yeah, it's inconsistent, but I think it works like this." That's not a very nice learning experience.
With a solver you can also study as long as you want, while a strong player might coach you an hour or two a day, probably less. And the ceiling is higher: with a solver you can get much better than the field. If you learn from a top-five player, you don't easily get better than the field.
That said, when I learned the stud games I combined both: the solver plus some coaching. A human can explain the essentials at the start, which speeds things up. Once you have the essentials, working with a solver is more efficient.
Learning 2-7 Triple Draw from scratch
Oleg: What was the impact of working with me, game by game?
RaulGonzalez: In no-limit Hold'em the impact was big at first. From spring 2013 our work shifted from 100 % intuition to maybe 90 % solver and 10 % discussion. The challenge became studying the solutions in a structured way and finding an efficient way to learn from them. I got much better at choosing bet sizes and at playing a more structured strategy.
In 2014 I got access to your draw solver, which I mostly used for 2-7 Triple Draw. That game was played on Full Tilt at 2K/4K almost daily. Unlike no-limit Hold'em, I learned it from scratch. I had no idea of the game before. And that was a very good experience: in Hold'em you have a lot of wrong intuitions and wrong concepts in your head, and then you see the solver doing something different. In 2-7 I could build my intuition from scratch.
It went very well. After about six months I was in the top three in that game, and I think many people considered me number one or two for quite a long time afterwards.
When the 2-7 games started dying, I learned Badugi with it too. It's more of a niche game, but it was still played up to 400/800 and regularly at 100/200. The game is complex and the competition wasn't nearly as strong, and I was basically the only one with a solver, so that was a huge edge. I also played some No-Limit 2-7 Single Draw, which is a game that is almost made to be solved: it's all about finding the right bet size and building your range around it.
Later the biggest projects were Stud and Omaha Hi-Lo, for the mixed games. Same story: I learned the games with the solver. Stud is extremely complex, so just having a solver isn't enough; you have to learn to work with it efficiently.
Oleg: How long did it take before the solver's strategy started to make sense and look consistent?
RaulGonzalez: It's gradual, but I'd say 30–50 hours, both in Hold'em and in draw. I remember starting with draw very well. It was by far my best learning experience in all of poker.
Trusting a strategy nobody else plays
Oleg: The first thing that happens when someone starts using a solver: they see a move no reasonable human would make and they get suspicious. How did you know the solutions were actually correct?
RaulGonzalez: I definitely had that experience a few times. When you find out the solver was right, and then it happens a couple more times, you gradually trust it more, because it has always turned out to be correct in the past. You also look at similar spots to see whether the same effect still shows, you talk to friends, and you check the EVs. And often you just have to think it through until you realise your intuition wasn't quite right.
In draw it was harder, because I was completely clueless myself, so I couldn't check whether a strategy made sense. There were some spots where the solution played differently from every human in the game, which is a mixed sign: it can mean a big edge, or it can mean the solution is simply wrong.
What gave me the most confidence was that if you worked long enough, you could see clear patterns. Then, before opening a new spot, you could predict how the pattern would apply, and it applied in a structured way you could anticipate. It also mattered a lot that I had already worked with your Hold'em solution and it had turned out to be right. It wasn't someone new showing me a draw solver. At first it still felt a little like a gamble, but once I started playing and felt I had a clear edge, that people didn't understand some things, my confidence went from medium to very high.
Oleg: Another customer told me that when you study a game you already played, you can get worse at first, because the solver's strategy interferes with your old one. Did that happen to you?
RaulGonzalez: Only in no-limit Hold'em, because I learned the draw games from scratch, and in Stud I knew so little that the solver was a huge improvement from day one. In Hold'em I'd say I played worse for maybe three or four weeks. It depends on how you apply it. Some people see a surprising overbet in a solution and want to use it immediately without knowing how it fits into their overall strategy, or the solver does something 5 % of the time and they do it far more often. That fancy-play syndrome can make you worse for a few days or weeks. Betting smaller on the flop in general doesn't hurt you; it's the special plays. Being aware of that, and studying more and playing less at the beginning, is enough.
Beating another team that had its own solver
Oleg: You played a lot against the group that had their own draw solver. How did that go?
RaulGonzalez: In 2014 and 2015 we played a lot of heads-up against them, from 1K/2K up to 2K/4K. It went pretty well for me. I think I was winning about 0.5–0.6 big bets per 100 hands. We didn't play anything below 1K/2K, a bit more at 1K/2K than at 2K/4K, so I'd guess around $600–700K in total, but it's hard to say exactly now.
Later I became close friends with one of them, and we compared how our solvers worked. It turned out theirs didn't take blockers into account in Triple Draw, meaning the discarded cards. That taught me a few things. First, I had underestimated how much blockers matter: ignoring them changes the strategy a lot and in a very structured way. Second, your solver accounted for one discarded card, and I should pay more attention to multiple blockers. Third, you and I started working on making the solver consider more of the discarded cards. We added a feature that showed how equity changes depending on which other cards you had seen, and it was surprising how big those effects were.
Oleg: Why do you think they kept playing you?
RaulGonzalez: They were confident in their solution, and it's hard to blame them. I'm sure I would have done the same. They had been crushing the 2-7 games for quite a while as the only ones with a solver. When you're in that position, it's very hard to accept that suddenly someone else has one too. The friend I mentioned told me they were convinced back then that I didn't have a solver. In hindsight it was obvious, he said, but at the time it wasn't.
How he actually studied
Oleg: What was your main workflow with the solver?
RaulGonzalez: It always started with playing against the solver. I'd stumble on a mistake or a surprising play, open the spot, and check how similar hands are played there: is the strategy mixed, or was I way off? If I was way off, I looked at both players' strategies to find what I hadn't understood. If there seemed to be a bigger concept behind the mistake, I'd form a theory, in my head, in a message to someone, or in a document. Then I'd open a different but similar spot and check whether it applied there too. If it didn't, I looked for the reason it applied in one spot and not the other.
Oleg: Did you track your mistakes?
RaulGonzalez: Yes. For a while I kept a spreadsheet: every 100 hands I typed in how many mistakes I'd made, and I watched the number go down. That motivated me a lot. Eventually I hit a floor of about eight to ten mistakes per 100 hands in 2-7 Triple Draw. Beyond that it's very hard, because it's often about mixed spots, like this hand bets 100 %, the next slightly better hand checks 100 %, and the next one bets again. That mostly happens out of position.
Oleg: How could a solver teach the essentials better?
RaulGonzalez: I think the most important thing is being able to play against the solver, and getting feedback on how big your EV mistake was. Real-time feedback is by far the best, but even a report afterwards is very good: you can list where you made big mistakes and where small ones, and you see which areas to improve. In the end, it's not so much about how good the solver is; it's about how much the human takes away from it. Most people enjoy playing against someone, even a solver, much more than studying tables of numbers. So making that experience nice, with a score and a count of your mistakes, gamification basically, is crucial.
One caveat: on the river, EVs alone can fool you. All bluff-catchers are roughly zero EV, so you could fold every jack-high against a bet and the solver would say each fold costs nothing. In a real game you'd be folding far too often, and a good opponent would bluff you to death. So on the river I also tracked frequencies: before checking, I'd predict "this category folds 10 %" or "this category bluffs 10 %", and then compare. For flop and turn, EV feedback is very, very good.
Oleg: Did playing against the solver translate directly to real games?
RaulGonzalez: Mostly, but at the beginning of draw I played much better against the solver than in real games. Against the solver, it showed me which cards I had discarded. In a real game I had to remember them, under time pressure, in a different interface, and I was completely overwhelmed for the first two weeks or so. So making practice as close as possible to the real game is valuable.
Who benefits most
Oleg: Is a solver useful for amateurs or part-time pros, or does it need a huge investment to be worth it?
RaulGonzalez: I think there's a lot of potential for semi-pros and amateurs. The better you are, the harder it is to improve, so in a way it's more valuable for weaker players, as long as you understand what the feedback means. There is a minimum, though: if you spend 15 minutes a day on poker, a solver probably isn't for you; you'd be better off with a human coach teaching you the essentials. From my experience, something like 40–50 hours is a reasonable minimum to really get value from it.
Oleg: What kind of player would benefit the most from my products?
RaulGonzalez: Someone who is already used to working on their game and wants to learn a new one, ideally a game where solvers aren't common yet. For example, a no-limit player who wants to start playing a Vegas mix with PLO and 2-7. That player could learn 2-7 or Omaha Hi-Lo very quickly with a solver.
Oleg: Was there anything non-monetary you got out of it?
RaulGonzalez: Working with a solver gives you a much more structured workday. Before, I'd sit at the tables and discuss hands whenever someone was around. With a solver, when I was preparing for a match, I'd say, "11 to 1, I play against the solver," then review and write down my mistakes, take a break, and repeat. I took that structure into other areas too. And over time I got into the mindset that a lot of things that look impossible at first can actually be solved with the right software.
It also helps your confidence. If you've been losing to someone, you can check your plays and your frequencies in the solver and see that you weren't actually over-bluffing; the opponent's crazy hero call just got lucky. Even today, at a very high level, confidence matters a lot, because when you lose it you play worse.
Looking back
Oleg: How would you describe working with me overall?
RaulGonzalez: The Hold'em solver was a pretty good experience, about a nine out of ten. The draw solver in 2014 and 2015 was a clear ten: a perfect experience, much better than I thought it could be. I'd often ask for a feature I hadn't even fully thought through, and half an hour later you'd send a new version that did it better than I had imagined, with something extra on top. Stud came out of nowhere in 2017. I was surprised it was even possible. In 2019 and 2020 it was back to a nine or ten, and it reminded me a lot of the draw days.
Oleg: What did you find most impressive?
RaulGonzalez: The speed at which you added features, and that even though you weren't a player, you could imagine what would be a great way to implement them.
Oleg: Thank you very much. It was fun talking to you.
RaulGonzalez: You as well.
If you play high-stakes mixed games or big tournaments and want to see the draw, stud or Omaha Hi-Lo solvers, reach out at olegsolvers.com. For Hold'em and Omaha, try Zeno, free in beta, at zenosolver.com.