Machine Learning and C-Bet Frequencies

Posted by

Posted by posted in Low Stakes

Machine Learning and C-Bet Frequencies

There's a great post that Reddit user Fossana created yesterday where they tried to figure out when a player in a BTN vs BB SRP situation could c-bet range for a third pot and what factors should most influence that decision. In order to come to those conclusions, they used Random Forrests and Decision Tree algorithms. Here's a quote from the post but you should really check it out if this stuff interests you:

After compiling all of this data in excel, I used machine learning libraries in python to see if machine learning could predict which flops you can cbet 1/3 100% without a large loss in EV (loss in EV < 1% pot). The basic idea behind machine learning is that you give an algorithm some training data (some subset of the 72 flops), and then it will come up with an equation that's as accurate as possible with its predictions in respect to the training data. Then you test the validity of the equation the algorithm came up with by seeing how well it does on unseen data (the flops that weren't used in training the algorithm). For this scenario, I used decision trees and random forests as the chosen algorithms. I won't go too much into what these algorithms do, but I think both of these algorithms are good for analyzing this situation because decision trees can give you heuristics to follow (e.g. cbet range on A8- flops), while random forests can give you a good idea of what predictor variables are most important.

The entire post can be found here.

Please leave your thoughts, should you have any, on their methodology or conclusions.

Loading 6 Comments...

Be the first to add a comment

You must upgrade your account to leave a comment.

This thread has been locked. No further comments can be added.

Runitonce.com uses cookies to give you the best experience. Learn more about our Cookie Policy