Elo Rating (Elo)
A rating system that summarizes each team's strength as a single number, updated after every game based on how the actual result compared to the expected result. Higher Elo means a stronger team.
The Elo rating system was originally developed for chess by Arpad Elo and later adapted for sports by FiveThirtyEight and others. The core idea is simple: every team has a single number that represents its current strength. Before each game, the two teams' ratings (plus a home-ice adjustment) produce a win probability. After the game, both ratings move by an amount proportional to how surprising the result was. A favorite winning at home barely shifts the numbers. An underdog winning on the road moves both ratings substantially.
Hockey Alchemy's Elo uses K = 12 (the size of an Elo move per game), a home-ice advantage of +50 points, an initial rating of 1505 for established teams (1490 for expansion), and a 30% regression toward the league mean of 1505 between seasons. These parameters were calibrated on 16 seasons of NHL data (2010-11 through 2025-26). A 100-point Elo gap corresponds to roughly a 64% win probability for the higher-rated team on neutral ice; a 200-point gap is about 76%.
Elo powers Hockey Alchemy's power rankings (the sorted list of current ratings), per-game win probabilities, and Stanley Cup odds (a 10,000-trial Monte Carlo over the remaining playoff bracket). We run three variants in parallel -- a simple binary win/loss Elo (the default), a margin-of-victory variant, and an xG-driven variant -- so we can compare predictive accuracy across approaches.
Formula
Expected: P(home) = 1 / (1 + 10^(-(home_elo - away_elo + 50) / 400)) Update: new_elo = old_elo + K * (actual - expected) K = 12, home ice = +50, initial = 1505, between-season regression = 30% toward 1505
Frequently Asked Questions
What do the Elo numbers actually mean?
The league average is 1505 by design. Teams typically range from about 1400 (bottom of the league) to about 1650 (top). A 100-point Elo gap corresponds to roughly a 64% win probability for the higher-rated team on neutral ice. Differences under ~30 Elo are within noise; differences of 100+ are large.
How is Elo different from a power ranking?
A power ranking is just a sorted list of teams by strength. Hockey Alchemy's power rankings page is a direct read of current Elo ratings -- there's no committee or human override. The two concepts are inseparable for us.
Why doesn't Elo predict playoff results perfectly?
Elo doesn't know about injuries, trade-deadline acquisitions, coaching changes, or goaltender hot streaks. It treats wins and losses as binary, so a regulation win and an OT win count the same. And the playoffs are a small sample -- even an accurate strength estimate can lose to the better team in a 7-game series.
What's the difference between simple Elo and xG Elo?
Simple Elo updates based on actual game outcomes (win or loss). xG Elo updates based on expected goal differential, which strips out finishing luck and goaltending heroics. Over time the two converge, but the gap between them at any moment indicates whether a team is over- or under-performing their underlying play.