skills.rest
sports-data-hq
sports-data-hqOFFICIAL
Explore all skills by this author →
GitHub

elo-engineering

Build and tune multi-variant Elo ratings for hockey team-strength prediction.

1|Updated May 5, 2026
Search Tags:#feature engineering#sports analytics#elo#ratings#hockey#win probability
One-click install
npx skills add https://github.com/sports-data-hq/hockey-skills --skill elo-engineering-sports-data-hq
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: elo-engineering
Source: https://github.com/sports-data-hq/hockey-skills/tree/main/skills/elo-engineering
Command: npx skills add https://github.com/sports-data-hq/hockey-skills --skill elo-engineering-sports-data-hq

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill builds and tunes Elo rating systems for hockey so you can estimate team strength, forecast game outcomes, and create reliable model features without hand-rolling rating logic.

Core Features & Use Cases

  • Multi-Variant Elo Modeling: Construct standard, fading, form, component, and flat-K Elo systems for different prediction needs.
  • Sports Prediction Features: Turn historical results into leakage-safe rating features for downstream win probability models.
  • Practical Hockey Workflows: Handle home-ice advantage, season carryover, margin-of-victory adjustments, and parameter tuning for NHL-style data.
  • Use Case: Analyze a season’s game results, generate current team ratings, and compare how different Elo variants affect prediction quality.

Quick Start

Ask for the elo-engineering skill to build hockey Elo ratings from your historical game results and return the current team rating table.

Frequently Asked Questions about elo-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build predictive Elo ratings for hockey team strength?▼

To build predictive hockey Elo ratings, provide chronological game results to apply multi-variant modeling for team-strength estimation, home-ice adjustments, and win-probability forecasting.

What is the best way to calculate win probability using Elo ratings?▼

Calculate win probability by using leakage-safe Elo rating features generated from historical results, supporting standard, fading, form, component, and flat-K variants for accurate outcome forecasting.

Can I apply home-ice advantage and season carryover adjustments to NHL data?▼

Yes, you can apply home-ice advantage and season carryover adjustments to NHL-style data, tuning parameters across different Elo variants to refine hockey team-strength modeling.

How do I generate leakage-safe features for win probability models?▼

Generate leakage-safe features by processing historical game results chronologically, ensuring Elo ratings and form tracking are updated sequentially without future data contamination for downstream models.

Do I need chronological game results to use fading and form Elo variants?▼

Yes, chronological game results are required to calculate fading, form, component, and flat-K Elo variants, as sequential ordering is necessary for accurate parameter tuning and rating updates.

What are the differences between standard, fading, and form Elo rating systems?▼

Standard Elo updates ratings uniformly, fading reduces the impact of older games, form tracks recent performance momentum, and component separates rating factors, each requiring specific parameter tuning for prediction.

Related Skills

Explore More Data & Analytics Skills→
PuckAPIPuckAPI
2

elo-engineering

Create multi-variant Elo rating systems for sports teams from historical game data.

Official
Advanced
sports-data-hqsports-data-hq
1

totals-modeling

Build calibrated NHL totals prediction models for game goal distributions and over/under probabilities.

Official
Advanced
sports-data-hqsports-data-hq
1

feature-engineering

Transforms raw hockey data into leak-free model features for predictive modeling.

Official
Advanced
89jobrien89jobrien
4

machine-learning

Automate the machine learning lifecycle from experimentation to production deployment.

Community
Advanced