Solvers & Creators

I'll touch on a few points:

1). Regarding solver outputs not being accurate because opponents are not opening solver ranges. You can set up a solve using the range you estimate your opponent is actually playing. Then you can solve for the GTO response given the constraint that your opponent is opening that range. The solver can "adjust" to the range you assign your opponent.

Afterwards you can re-run the same spot using GTO ranges, and compare how the strategies change. This is a good way to learn to adjust to different opponent types. What are we doing differently? You can also compare to see how much EV their strategy is losing.

One of the issues with this approach is that the solver will assume your opponent is playing well on later streets. It will attempt to craft a balanced flop strategy given the preflop constraints, when in reality your opponent might just play fit or fold on the flop. The solver is still useful though.

2). I think "never solver" and "always imitate solver" approaches are both overly rigid. You can study solver outputs and still deviate in creative ways for various reasons.

Personally I think solver outputs are most helpful on early streets. The reason is that solvers are very good at "setting the table." They take actions so that they arrive on the later streets with the correct numbers of value and bluff combos on various runouts. This makes the turn and river easier to play.

If you deviate on the flop and then a particular river card comes you might realize that you basically always or never have a strong hand on that particular runout. Then you're pretty much screwed and open to being massively exploited.

I will deviate on all streets when I perceive that it is better to do so, but I try to imitate the solver more preflop and on the flop. That's to help me construct a range that makes the turn and river easier to play.

Then on the turn, and even moreso on the river, I deviate a lot. That's based on my opponent's tendencies. If they overfold flop I might underbluff turn because I will assume that when they do call the flop their turn range is strong.

Another example, if our opponent is barreling into us many of our hands turn into bluffcatchers that are indifferent between calling and folding. That means that if our opponent is bluffing the optimal amount then our bluffcatchers will have the same EV whether we call or fold.

On the other hand if they're overbluffing we should always call with those bluffcatchers. If they're underbluffing we should always fold them. It's really hard for a human to be perfectly balanced.

3). As a live player I find pre flop to be the most valuable thing to study. That's mainly because preflop is the one street we play every hand. It can be pretty apparent when we make a mistake on the river, but if we're opening J9o every time UTG and on average losing like .2 BB it can be really difficult to ever even realize it's a problem without a solver or analytics.

PS. Sorry for writing a book, but hopefully it helps someone.

Thanks

This is a great post and the kinda feedback I seek. I like the ‘setting the table’ analogy and the ideas behind it. I like the pre-flop focus, but I’m not thinking of opening or doing anything every time. That’s kinda the point with me: playing solid poker, but occasionally get out of line in ways villain won’t expect.

Making the turn and river easier to play is always a good goal. Please understand, I do think solvers are valuable. The idea to focus on ranges probably came from them. However, I’ve been confused that taking a different line pre-flop causes the solver to shutdown with no more advice for a hand. It seems to me it should adjust to new circumstances.

Your post clicked with me - there’s so many variables post-flop that maybe can’t be plugged into the solver, but on early streets it’s always going to have solid advice. I just feel like some get involved so much in specific situations with solvers that they can’t apply this knowledge in real time.

Thanks again, I feel blessed to live in an age where strategic thinking is king. Those that are best at interpreting these ideas will succeed.

In real games it does seem like there is some potential value in just having a hand that your opponent doesn't expect to be in your range. You might get paid off when you make an unexpected straight or something. It's a slippery slope though because that hand's not in your range for a reason. It's typically -EV.

You can still use a solver with unorthodox ranges, you just need to be able to quantify it with the preflop range. Like if the solver never opens low suited one-gappers in a spot, you can modify the preflop range so that it opens them 1/4 of the time or something like that.

Also I stole the "setting the table" analogy, but I agree that it's a good one. I heard it a couple years ago in a poker strategy video. I think it might have been Jonathan Jaffe that came up with that.