wrcarpenter / MMA-Betting-Model

Data-driven, sports wagering model for MMA (mixed-martial-arts) contests.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Screenshot

Mixed Martial Arts (MMA) Betting Model

"Slow is smooth, smooth is fast" - Conor Mcgregor

Tracking mixed martial arts competitions and betting activity to identify optimal trading strategies. Take a statistical and automated approach to placing wagers on weekly events in an effort to maximize returns and minimize human engagement with detailed analysis.

Objective

Create a betting strategy that outperforms generic approaches (random chance, bookmaker odds, etc) and delivers supierior, and uncorrelated, returns to the broader markests.

Inspiration(s)

Bill Benter, a successful horse gambler active mostly during the 1990s in Hong Kong that one of the first to popularize quantitative betting models in a sports context.

Data Sources

Name Link Description
UFC Stats ufcstats.com Historical UFC fight data and roster
Tapology tapology.com Comprehensive event and figter data across numerous MMA venues
Best Fight Odds bestfightodds.com Historical odds for MMA events from a variety of sportsbook platforms

Some Focus Questions

A few main focus questions:

  • What are the main fighter characterics that influence win/lose probability?
  • How random are fight outcomes?
  • Can public odds markets accurately predict fight outcomes?
  • What types of fights are the most predictable/unpredictable?
  • Is there enough publically available data to make informed decisions about fight outcomes?
  • What is the best model choice for predicting fight outcomes?
  • Can I create a specifc ELO system that seems to reflect other MMA ranking systems well?

Betting Sites

  • FanDuel (NY), BetMGM (NY), Caesars (NY), WynnBET (NY), BetRivers (NY), DraftKings (NY), PointsBet (NY)

About

Data-driven, sports wagering model for MMA (mixed-martial-arts) contests.


Languages

Language:Jupyter Notebook 78.4%Language:Python 20.0%Language:Stata 1.5%