Good games can
have quiet launches.

We look for older indie games with small audiences and promising evidence of enjoyment.

What counts as a potential gem?

Games must be at least 60 days old, be classified as Indie in Steam’s store metadata, and have a completed review scan. Our main collection stays at 300 reviews or fewer; the broader collection includes up to 1,000. This classification is imperfect and may miss independent games that lack the Indie label.

Two ways to catch our attention

At least 75% positive reviews, or at least 50% positive reviews alongside sustained reviewer playtime. Our first version uses explicit playtime thresholds: ten hours for Strategy, Simulation, and RPG games; three hours for other genres. These are rough heuristics, not measured genre percentiles. Short games can qualify entirely through review sentiment.

The ranking adjusts approval using a modest prior of eight positive and two negative reviews, rewards small audiences, and considers the median and share of engaged reviewers. We show sample sizes because five happy reviewers and five hundred are different kinds of evidence. High playtime can include idle time.

A fresh selection every day

Each edition is saved for the day in UTC. We favor games not shown in the last 14 days, vary the genres, and leave room for games with very little evidence. If the pool is small, we revisit the least recently featured games. New selections do not imply new releases or new reviews.

How complete is the search?

The initial database is a small live seed sample. The importer can enumerate Steam’s available game catalog when configured with a Steam Web API key, screen reviews, and gradually enrich the low-exposure pool. Private, removed, inaccessible, or unclassified games may be missing. Failed fetches remain unknown and are retried.

What these numbers mean

Review totals include all languages, Steam purchases, and other purchase sources; they can differ from the score Steam chooses to display. Playtime comes from reviewers, not a representative sample of all owners. A full scan means pagination completed, not that every owner reviewed the game. We retain game-level statistics and historical summaries. Individual review inputs are discarded after processing; reviewer profile IDs are not stored.

Potential, not a promise

These selections are evidence-based candidates, not claims that we personally played every game. Read the reviews, watch some gameplay, and decide what appeals to you. Advertising will never determine the discovery ranking.