Methodology

Where every number comes from.

Everything below runs on public Steam store and community data, collected by our own pipeline, every day. No panels, no estimates, no scraped dashboards. When a number appears anywhere on this site or in the brief, it traces to one of these instruments.

Discovery

Every demo that went live yesterday, found via the demo's own release date — so a demo dropping on a months-old store page is caught the same as one on a brand-new page. The cohort is complete, not curated.

Effort score

How much work the developer put into the store page: description depth, screenshots and trailer beyond the minimum, localization, tagging. A seriousness proxy — deliberately blind to whether the game is good. We never call it quality.

Tag novelty

How rare the game's tag combination is against a frozen Steam-wide benchmark of tag prevalence. Category novelty, not execution novelty — a brilliant but ordinary-looking roguelike scores low here, and that's the score working.

Demo CCU

Concurrent players on the demo itself, polled from day one. The industry's canonical public metric — but nobody else publishes it for demos. Live attention, not accumulated hype.

Demo reviews

The demo's own review count and sentiment — direct player feedback that exists before launch, when the full game has zero reviews. The real "is it good?" read, and the strongest early signal we've measured.

Followers

The game's community-group member count over time — the standard proxy for demand, since Valve exposes neither wishlists nor store follows. We label it a proxy, always.

The honesty rules