Poker-AI.org

has anyone read this book,
it basically tells you what to do preflop, flop, turn and river
is it possible to create a bot from everything it states in this book?
like a carbon copy, there has been players following this and went from 2nl to 100nl.
also a bot would follow it 100% as humans could make mistakes.

im obviously new to this site and want to learn as i want to create a auto foldem bot which sticks to the guide lines in the book above but dont have a clue where to start lol.

also have the book if anyone wants to pm me for it to look at

pdf sent to two person.

It isn't great...

I recently made a bot that followed NoBS 6 max to the letter. It ended up being a -70c/hour loser at 888 Zoom 2NL (1 table, ~240 hands/hour). Most mistakes were due to not considering board texture in their rule making, and not considering ranges (flow charts don't care if you're against SB or BTN post-flop).

Yeah it would be impossible to win without ranges and board texture rules. 2NL is probably not the best place to test it anyway due to the massive rake and play money like environment in which your going to see a ton of multiway flops. It is however a good place to test the bot knowing it will lose and adjust its frequencies, bug fix etc before moving it up in limits. I am finding that a micro limit strategy is completely different from a NL100 strategy. In micro limits your going to see 3 and 4+ way flops more often then HU which is the opposite of NL100.

Yeah, may not be the best place but it's better than losing a $10 hourly at higher stakes just to disprove theory!

I'm currently coding an equity based bot, utilising an in-build 7-card hand evaluator and enumerator to predict equity in a hand vs range and range vs range dynamic to get around this.

This is true but sometimes its hard to decipher what the true results would be. Every AI needs to be designed for the game its playing. I am currently having to rework mine to handle different limits at once which is a royal pain.



Sounds good. I wish I had my own in house range evaluator as I am losing speed by using a third party tool. I am def too lazy to figure it out so I am stuck with it..



Yeah..no strategy not considering this would be an auto loser. To put this into perspective on my PLO bot: The BB defense vs a button steal would be around 25% and maybe 8% 3B, vs a SB steal you would be defending at around 60% and 3B at around 12%+ if I recall.