Let's Crush it at Poker!

Day[1]

I know if you’ve read the intro to this and you’re not a poker player, this will sound like gibberish to you. 😄

So with that in mind, I’ll take a bit of time to explain a little bit further about what I’m going for with this idea. My future posts will likely be a bit more technical, but this should be a good starting point.

So my game of choice is 6-handed No Limit Texas Hold’Em, playing online. This is a game format that has evolved in strategy massively over the last 10 - 15 years or so. Firstly there was the progression from this being more of a “tells”-based game to something a bit more mathematical. For example, I don’t just “feel like he has Ace-King”. I just think, “oh he could have this small range of hands, and given the odds I am being laid, I am simply getting a good enough price to call.”

From there the game evolved further into “game theory”. With players thinking about nash equilibriums, and realising you can actually play in such a way, by having the “correct” ratio of bluffs and value bets, that you will win by default over your opponent, by playing in a way that is as “un-exploitable” as possible.

Pretty quickly this thinking evolved into computer programs. With there now being software that can, given certain paramaters, use a brute force method (the computer plays millions of hands against itself) to work out what the absolute best strategy is between two players.

As an example, here is an output showing a flop c-bet strategy on an King-Nine-Three Flop:

Now the game of poker is still far from solved, but these programs provide an excellent baseline as to deciding what our rough strategy should be in different situations. These programs also allow you to adjust the inputs, and see how the strategy changes, which allows you to see how you might play to maximally exploit your opponents if you get an idea of what kind of strategy they are playing.

Now you’ll note that this was only a single hand example. And with there being 1755 strategically different poker flops possible, memorising every strategy is essentially impossible (unless Flash took up Poker).

That’s where this project comes in. I want to create a base set of simulations across many different boards. This will allow me to identify patterns in what the correct strategy looks like. I can then test myself on those outcomes, internalise that knowledge, and use that in my games. here’s an example of what some of these outputs might look like (from an older attempt at a similar project):

So my goal for today is simply to prepare and run a set of simulations overnight. That way I’ll have a nice set of simulations outputs tomorrow that I can start to organise into my spreadsheet.

See you tomorrow!