Analytics Dashboard

Model performance metrics and prediction analytics

0.0%
Overall Accuracy
0
Total Predictions
0
Correct Predictions
6
ML Models
Model Accuracy Comparison
Individual model performance
0%25%50%75%100%XGBoostELOMonteCarloPoissonBayesianLinear
Weekly Accuracy Trend
Prediction accuracy by week
123456789101214161840%55%70%90%
Confidence Distribution
Prediction confidence levels
High (>70%): 35%Medium (55-70%): 45%Low (<55%): 20%
Model Performance Radar
Multi-metric comparison across models
AccuracyPrecisionRecallF1 ScoreAUC-ROC0255075100
  • XGBoost
  • ELO
  • Bayesian
Model Details
Individual model specifications and performance metrics
XGBoost72.0%
256 predictions made
ELO68.0%
256 predictions made
Monte Carlo65.0%
256 predictions made
Poisson63.0%
256 predictions made
Bayesian66.0%
256 predictions made
Linear61.0%
256 predictions made
XGBoost

Gradient boosting ensemble using team stats, historical matchups, and situational factors. Best for capturing complex non-linear relationships.

ELO Rating

Dynamic rating system that updates after each game. Accounts for margin of victory, home field advantage, and opponent strength.

Monte Carlo

Runs 10,000 game simulations using probabilistic distributions. Provides robust win probability estimates with uncertainty quantification.

Poisson Distribution

Models scoring as Poisson processes using attack/defense strength ratings. Excellent for predicting game totals and spreads.

Bayesian Restricted Likelihood

Advanced Bayesian model with restricted likelihood estimation. Handles small samples well and provides calibrated probability estimates.

Linear Regression

Feature-based linear model using team statistics. Provides interpretable coefficients and serves as a baseline model.