Direct Head-to-Head (H2H) Record

Overview

This pillar evaluates the direct head-to-head (H2H) record between two competitors. It provides a historical context for a matchup, revealing psychological edges or stylistic weaknesses that general performance ratings often miss.

What It Does

The pillar aggregates all historical game results between two specific chess players. It then calculates a weighted score that gives more importance to recent games and decisive outcomes (wins vs. draws). This produces a clear metric of which player has historically dominated the other in their personal rivalry.

Why It Matters

A strong H2H record often indicates a persistent tactical or psychological advantage. This pillar provides a focused signal that cuts through the noise of overall player ratings, offering a crucial edge in predicting the outcome of specific, high-stakes matches.

How It Works

First, the system queries chess databases for all recorded games between the two specified players. Each game result is assigned a point value (win, loss, or draw). A time-decay formula is then applied, increasing the weight of games played more recently. Finally, all weighted scores are summed to generate a single H2H dominance score.

Methodology

The core calculation is a time-weighted sum: H2H Score = Σ (Outcome_Points * Time_Decay_Weight). Outcome_Points are assigned as Win=1, Draw=0.5, Loss=0. Time_Decay_Weight = e^(-0.001 * days_ago), heavily favoring games within the last two years. Games played under Classical time controls receive a 1.2x multiplier compared to Rapid or Blitz games.

Edge & Advantage

This pillar isolates the specific player-vs-player dynamic, revealing matchup-specific advantages that broad Elo ratings cannot capture.

Key Indicators

• Recent H2H Score

  high

  The win/loss/draw record over the last 5-10 games, indicating current dominance.

• Lifetime H2H Score

  medium

  The overall record across all games played between the two players.

• Decisive Game Ratio

  medium

  The percentage of games between the two that ended in a win or loss, not a draw.

Data Sources

• Chess.com API

  Provides game archives and live game data for players on the platform.

• Lichess Open Database

  A massive, publicly available database of chess games played on Lichess.org.

• Chessgames.com

  A curated database of historical chess games from major tournaments.

Example Questions This Pillar Answers

• → Will Magnus Carlsen defeat Hikaru Nakamura in their next match?
• → Will the next game between Caruana and Nepomniachtchi end in a draw?
• → Will Gukesh D have a positive score against Praggnanandhaa in the Candidates Tournament?

Tags