Concepts Index¶

concepts/ (statistical models)¶

bayesian-inference-sports — Bayesian Inference in Sports Prediction: method of statistical inference updating probabilities as new evidence becomes available; foundational to modern sports prediction models combining prior beliefs with observed match data. Sources: bayesian-inference-sports, poisson-distribution, dixon-coles-correction, bradley-terry-model, glicko-2, expected-goals-xg, calibration-plots, brier-score, log-loss-cross-entropy, polymarket, multi-sport-prediction-stack, poisson-elo-ensemble.
bradley-terry-model — Bradley–Terry Model: probability model for predicting the outcome of pairwise comparisons between items/teams; foundational to Elo rating system and used to estimate team strength from head-to-head results. Sources: bradley-terry-model, elo-rating-system, glicko-2, massey-ratings.
dixon-coles-correction — Dixon–Coles Correction: statistical model extending independent Poisson distribution to better capture correlations in low-scoring football matches; applies correction to 0-0, 1-0, 0-1, 1-1 scorelines via a tau (ρ) parameter. Sources: dixon-coles-correction, api-football, football-data-co-uk, poisson-distribution, elo-rating-system, expected-goals-xg, poisson-elo-ensemble.
elo-rating-system — ELO Rating System: method for calculating relative skill levels of players/teams in zero-sum games via logistic win-probability function; adapted for football, tennis, and American football. Sources: elo-rating-system, dixon-coles-correction, glicko-2, bradley-terry-model, massey-ratings, poisson-distribution, expected-goals-xg, multi-sport-prediction-stack, poisson-elo-ensemble, sportradar, api-football.
expected-goals-xg — Expected Goals (xG): statistical metric assigning a probability to each shot that it results in a goal; sum of shot probabilities gives expected goals for a team/player. Sources: expected-goals-xg, dixon-coles-correction, poisson-distribution, bayesian-inference-sports, api-football, sportradar, poisson-elo-ensemble.
glicko-2 — Glicko-2 Rating System: improved rating system over Elo introducing rating deviation (RD) to measure uncertainty and rating volatility (σ); used in chess, Dota 2, Counter-Strike. Sources: glicko-2, elo-rating-system, bradley-terry-model, bayesian-inference-sports.
massey-ratings — Massey Ratings: method for measuring past performance of sports teams using a system of linear equations solved simultaneously across all game results; designed for order rather than prediction. Sources: massey-ratings, elo-rating-system, bradley-terry-model, bayesian-inference-sports, sportradar.
poisson-distribution — Poisson Distribution: discrete probability distribution modeling the number of goals scored in a match, given a known mean rate (λ); foundational model for expected goals prediction in football. Sources: poisson-distribution, api-football, bayesian-inference-sports, dixon-coles-correction, elo-rating-system, expected-goals-xg, massey-ratings, football-data-co-uk, multi-sport-prediction-stack, poisson-elo-ensemble.

betting/¶

closing-line-value — Closing Line Value (CLV): measures the difference between odds placed at and the closing odds before the event starts; key metric for validating predictive model quality by checking whether the model beats the most efficient market price. Sources: closing-line-value-clv, backtesting-framework, upwork-intake, expected-value-ev, market-efficiency, line-movement-steam-moves, de-vigging, oddsjam, pinnacle-api, the-odds-api, walk-forward-validation, football-data-co-uk.
de-vigging — De-vigging: process of removing the bookmaker's overround (vigorish) from betting odds to get fair implied probabilities; essential for accurate EV calculations. Sources: de-vigging, expected-value-ev, closing-line-value, market-efficiency, oddsjam, pinnacle-api, the-odds-api.
expected-value-ev — Expected Value (EV): average amount won/lost per bet if the same wager were placed repeatedly; positive EV (+EV) bets are the foundation of profitable betting strategy. Sources: expected-value-ev, kelly-criterion, de-vigging, closing-line-value, value-bet-identification, brier-score, half-kelly, oddsjam, pinnacle-api, the-odds-api, polymarket, multi-sport-prediction-stack.
half-kelly — Half-Kelly: staking strategy using 50% of the full Kelly fraction to reduce volatility and account for model uncertainty while still capturing most of the geometric growth benefit. Sources: half-kelly, kelly-criterion, value-bet-identification, expected-value-ev.
kelly-criterion — Kelly Criterion: formula for sizing bets to maximize long-term geometric growth rate of bankroll; Kelly bet = (bp - q) / b where b=odds-1, p=win probability, q=loss probability. Sources: kelly-criterion, half-kelly, expected-value-ev, value-bet-identification, backtesting-framework, upwork-intake.
line-movement-steam-moves — Line Movement & Steam Moves: tracking how betting odds change over time; steam moves indicate sharp bettor action and can signal market efficiency shifts. Sources: line-movement-steam-moves, market-efficiency, closing-line-value, value-bet-identification.
market-efficiency — Market Efficiency: degree to which bookmaker odds reflect true probabilities; sharp vs. soft bookmakers, favorite-longshot bias, and implications for finding value. Sources: market-efficiency, closing-line-value, de-vigging, value-bet-identification, line-movement-steam-moves, oddsjam, pinnacle-api, the-odds-api, polymarket.
value-bet-identification — Value Bet Identification: finding bookmaker odds higher than the true probability warrants; the core skill in profitable sports betting combining model predictions with market odds. Sources: value-bet-identification, expected-value-ev, closing-line-value, market-efficiency, kelly-criterion, half-kelly, line-movement-steam-moves, oddsjam, upwork-intake.

validation/¶

backtesting-framework — Backtesting Framework: systematic evaluation of a sports betting model by simulating performance on historical data; must mimic real deployment conditions including only information available at prediction time. Sources: backtesting-framework, overfitting-sports-models, walk-forward-validation, closing-line-value, brier-score, kelly-criterion, overfitting, football-data-co-uk.
brier-score — Brier Score: strictly proper scoring rule measuring mean squared error between predicted probabilities and actual outcomes; lower is better, 0 is perfect. Sources: brier-score, log-loss-cross-entropy, calibration-plots, bayesian-inference-sports, expected-value-ev, backtesting-framework, walk-forward-validation, overfitting-sports-models, poisson-distribution.
calibration-plots — Calibration Plots: visual comparison of predicted probabilities vs. actual outcome frequencies; a well-calibrated model shows points on the diagonal. Sources: calibration-plots, brier-score, log-loss-cross-entropy, bayesian-inference-sports, walk-forward-validation, overfitting-sports-models.
log-loss-cross-entropy — Log-Loss (Cross-Entropy): scoring rule heavily penalizing overconfident wrong predictions; used for optimizing probabilistic classifiers and as a training objective. Sources: log-loss-cross-entropy, brier-score, calibration-plots, bayesian-inference-sports.
overfitting-sports-models — Overfitting in Sports Models: when a model learns noise in training data rather than true patterns; prevention requires walk-forward validation, regularization, and small parameter counts. Sources: overfitting-sports-models, walk-forward-validation, backtesting-framework, brier-score, calibration-plots, multi-sport-prediction-stack.
walk-forward-validation — Walk-Forward Validation: chronological train/test split where a model is trained on one period and tested on the next; mimics real deployment and prevents look-ahead bias. Sources: walk-forward-validation, backtesting-framework, overfitting-sports-models, brier-score, closing-line-value, calibration-plots, upwork-intake.

data-sources/¶

api-football — API-Football: comprehensive football REST API covering 1,200+ competitions, fixtures, match statistics, lineups, player data, and odds; plans from $19/mo. Sources: api-football, expected-goals-xg, dixon-coles-correction, poisson-distribution, expected-value-ev, elo-rating-system.
football-data-co-uk — Football-Data.co.uk: free historical football results, match statistics, and closing betting odds for 22+ European leagues spanning 25+ seasons; gold standard for free backtesting data. Sources: football-data-co-uk, poisson, dixon-coles-correction, closing-line-value, market-efficiency, backtesting-framework.
oddsjam — OddsJam: sports betting odds API focused on market efficiency and value identification; $199/mo plan with full historical odds, line movement, and sharp book comparisons. Sources: oddsjam, market-efficiency, closing-line-value, de-vigging, value-bet-identification, expected-value-ev.
pinnacle-api — Pinnacle API: sharp bookmaker API providing efficient odds for major sports; historically the benchmark for market efficiency and CLV validation. Sources: pinnacle-api, market-efficiency, closing-line-value, de-vigging, expected-value-ev, backtesting-framework.
polymarket — Polymarket: decentralized prediction market built on Polygon blockchain offering public API for market prices, volumes, and settlement across sports, politics, and world events. Sources: polymarket, market-efficiency, expected-value-ev, bayesian-inference-sports, api-football.
sportradar — Sportradar: professional-grade sports data provider covering official statistics, live odds, player tracking, and event data for 80+ sports; enterprise pricing typically $1k+/month. Sources: sportradar, expected-goals-xg, elo-rating-system, massey-ratings, expected-value-ev.
the-odds-api — The Odds API: aggregated sports betting odds API covering 80+ bookmakers, multiple sports, free tier with 500 credits/month, paid plans from $99/mo. Sources: the-odds-api, market-efficiency, closing-line-value, expected-value-ev, de-vigging, backtesting-framework.

models/¶

multi-sport-prediction-stack — Multi-Sport Prediction Stack: unified machine learning architecture for predicting matches across football, NFL, and NBA using shared infrastructure with sport-specific model heads. Sources: multi-sport-prediction-stack, poisson-distribution, elo-rating-system, bayesian-inference-sports, expected-value-ev, overfitting-sports-models.
poisson-elo-ensemble — Poisson-ELO Ensemble: ensemble approach combining Poisson goal-scoring models with ELO rating systems via weighted average; captures both team quality and historical performance dynamics. Sources: poisson-elo-ensemble, poisson-distribution, elo-rating-system, dixon-coles-correction, expected-goals-xg, bayesian-inference-sports.