elo-ratings-math

I've been digging into Elo rating implementations lately, and your skill's approach to the math fundamentals is solid—especially how you're breaking down the rating delta calculations. One thing I'm curious about: how do you handle the K-factor tuning for different skill levels, since that's where most implementations get brittle?

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 80/100, solid B-grade territory. This evaluation's based on Anthropic's skill architecture best practices. Your strongest area is Utility (18/20)—the math is comprehensive and addresses a real capability gap. Weakest area is Progressive Disclosure Architecture (18/30)—you've got 229 lines crammed into a single file, which kills your token efficiency and navigation.

What's Working Well

• Comprehensive mathematical coverage: You go from basic rating calculations all the way through multi-player extensions and variants (Glicko, TrueSkill). That's serious depth.
• Excellent worked example: Lines 143-162 show step-by-step calculations with input/output pairs. That's the kind of concrete thing that actually helps people understand.
• Clear metadata and triggers: Your description ("implementing or understanding competitive rating systems") is specific enough that the skill surfaces at the right times. The YAML frontmatter is clean and valid.
• Bonus points for grep-friendliness: Your structure and consistent terminology (E_A, R_A, K-factor) makes it easy to search and reference.

The Big One: Missing Progressive Disclosure Structure

This is what's eating 12 points. You've written a comprehensive guide, but it's all sitting in one 229-line file. For files over 100 lines, Anthropic's best practice is to split content into reference files so developers c...