Methodology
About SportsEdge
The math behind the predictions.
The Elo Rating System
Zero-sum competitive tracking
SportsEdge uses a modified Elo rating system to track team strength across different leagues. Originally invented by Arpad Elo for chess, the system is zero-sum: after every game, the winning team takes points from the losing team.
The number of points exchanged depends on the difference in ratings between the two teams. If a highly-rated team beats a lower-rated team, only a few points change hands. However, if an underdog pulls off an upset, a significant number of points are transferred.
Expected Score = 1 / (1 + 10 ^ ((Opponent Rating - Team Rating) / 400))Key Variables & Modifications
K-Factor
The K-factor determines how reactive the system is to recent results. SportsEdge uses a dynamic K-factor of 20.0 for standard games (24.0 for EPL).
Home Field Advantage (HFA)
Playing at home provides a statistical advantage across all sports. We inflate the home team's rating by 55 Elo points when calculating win probabilities.
Margin of Victory (MOV) Multiplier
Not all wins are equal. A blowout transfers more rating points than a one-point finish, adjusted for the fact that a stronger team is already expected to win by a larger margin.
Season Regression
Between seasons, ratings regress by 30% toward the mean (1500) to account for roster turnover.
Soccer 3-Way Probabilities
Unlike most American sports, soccer matches frequently end in a draw. The standard Elo system is designed for binary outcomes (Win/Loss).
For the English Premier League (EPL), we model draw probability as a function of the rating gap. Evenly matched teams have the highest draw chance (~25–30%); as the skill gap widens, draw probability falls and remaining mass is split between home and away.