UFC AI analysis - Emmett vs Murphy
I'm a machine learning engineer and I spent the last several years perfecting a binary classification model ensemble to predict UFC. I recently released v6 of the ensemble and since then we've had 4 profitable events in a row with an average of like, 30-40%. This last year it's really come together. 23% ROI since v4's release earlier this year. I used to 3rd party track it here (https://www.betmma.tips/flyingtriangl3) but the UI is really annoying and I realized you can just use the Wayback Machine (https://web.archive.org/web/20250000000000*/mma-ai.net) to see I'm not changing prior predictions.
I have detailed guides to the math I do (poisson gamma smoothing, zscore normalization, decayed averages, etc on the blog (https://mma-ai.net/news) to help newcomers to the algorithmic betting space get started. I share everything from all the features, to the evaluations, to the exact Python libraries I use there.
Here's the detailed statistical analysis for this weekend's event:
I have detailed guides to the math I do (poisson gamma smoothing, zscore normalization, decayed averages, etc on the blog (https://mma-ai.net/news) to help newcomers to the algorithmic betting space get started. I share everything from all the features, to the evaluations, to the exact Python libraries I use there.
Here's the detailed statistical analysis for this weekend's event:
