Methodology

Data collection

Every morning at 07:00 UTC, a scheduled job queries Apple's official iTunes RSS feeds for the US Games genre (newapplications, topfreeapplications, topgrossingapplications) and enriches each entry with the iTunes Lookup API. Raw data is committed to the public source repository.

Classification

Each game is classified into one of ~16 archetypes (match-3, idle, RPG, etc.) using a rule-based detector applied to the description and release notes. Where signal is strong enough (Google + YouTube Suggest combined ≥ 8), an LLM pass refines archetype, core loop, audience, and visual treatment.

Indexing decisions

Pages without organic demand signal AND without substantive metadata are marked noindex,follow to keep the site lean. Pages with no traffic after 30 days are removed (HTTP 410). We index quality, not quantity.

Images

App icons and screenshots are downloaded from Apple's CDN, converted to AVIF + WebP at multiple resolutions, and served from our own static host. Alt text describes the gameplay moment, not a generic "screenshot".

Editorial limits

Observations are derived from Apple's official metadata. We do not claim hands-on play unless explicitly stated. Red-flag warnings reflect patterns common to the mechanic, not personal experience.

Update frequency

Data refreshes once daily. Per-game pages rebuild automatically on each refresh. The full source code is open and the data feed is available as JSON.