Search a fighter
Type any name in the search bar above
Loading card...
Fetching upcoming UFC event from ESPN
| Name | Division | Record | Archetype | Tags |
|---|
| Name | Division | Record | Score | Age | Bracket | Archetype |
|---|
For each tag or archetype, shows historical win rate when a fighter with that attribute faced a fighter without it. Signal > 55% or < 45% indicates genuine predictive power.
No predictions yet — use Card Analysis to generate analysis, then click "Lock prediction" on each fight.
Feature importance analysis — which stats and tags actually predict fight outcomes.
Trained on historical fight data where both fighters are in the roster.
No results yet
Click "Train model" to run the analysis.
Takes about 60 seconds. Requires xgboost and scikit-learn.
pip install xgboost scikit-learn
Enter the sportsbook odds and your model's confidence. The calculator checks all three filters and returns a recommended bet size via Kelly Criterion.
Use + for underdog, − for favourite
Your model's estimated win probability
Quarter Kelly is recommended for most bettors — preserves bankroll through variance while capturing meaningful edge.
Hover over any tag anywhere in the app to see its definition inline. Full reference below.