This module explores why MMA is so challenging to measure and predict who will win a fight.
On this Page
Learning Objectives
By the end of this module, you will be able to:
- Understand why measuring and predicting MMA fight winners is a challenging task.
- Discover signals that indicate how strong an MMA fighter is and how this signal represents their overall ability.
- Understand how you can make sense of MMA metrics such that you can measure and predict MMA fight winners.
Module Map
This module contains the following progressive lessons.
- Lesson 1: Framing The Problem
- Introduces the challenge behind measuring and predicting fight winners in MMA.
- Lesson 2: Context and Sample Size
- Explores why a dominant record for a MMA fighter fails to predict future wins.
- Lesson 3: Depth of Experience
- Explores how to measure the quality of a fighter's experience, rather than just the quantity.
- Lesson 4: Styles Make Fights
- Tries to answer why the better fighter doesn't always win the fight.
- Lesson 5: The Big Picture
- Answers the question of how can we actually make sense of MMA data in a way that is meaningful, predictive and useful?
Module Resources
For all of the module resources, such as data, code, analysis/project files, and so on, be sure to check out our GitHub Repository. We reference specific links throughout the module, but there are many more code and data resources we included in the repo for your learning.
Key Takeaways
If you take one thing away from this module, it's that predicting MMA fights is nto easy. Other sports, such as hockey, soccer, baseball, and basketball, have vast amounts of data to draw from, so you're able to get a much richer sample with proper signal, distribution, outliers, and so on. MMA data, however, requires context to really build models that invariably come from small data. However, once you build a model (using context as your guide), you now have both a model to use for your analysis and context around the fight data that can guide your narrative.