Why We Built the Discovery Queue

Steam receives over 14,000 new releases per year. The average indie title sells fewer than 1,000 copies in its first month. Not because these games are bad — but because they never get seen. The storefront is so crowded that even great ideas drown in noise before anyone has a chance to evaluate them.

Traditional game jams solve the "build something" problem but not the "would anyone care?" problem. You prototype for 48 hours, post to itch.io, and hope for the best. There's no structured way to test whether your concept has real market appeal before committing months of development.

We asked a different question: What if we could simulate the entire storefront experience — discovery, wishlisting, market validation — at the idea stage?

Valve's Discovery Queue is one of the most effective game recommendation systems ever built. When you open your queue on Steam, you're presented with a curated sequence of games — one at a time, full-screen, with all the context you'd need to make a decision: screenshots, trailers, tags, reviews, and a single decisive action — "Add to Wishlist" or "Next".

This UX pattern works because it solves three problems simultaneously:

Steam's system generates billions of data points about player interest. We don't need billions — we need enough signal to tell which of our ideas has the best shot at market success.

Every game jam submission on Metal Games goes through the same pipeline that a real Steam release would. That means each entry needs the full asset set:

This isn't busywork — it's the point. By requiring Steam-grade assets at the concept stage, we force every idea through the same lens a real customer would use. If you can't make your game look compelling in a store capsule, that's a signal worth catching before you invest months of development.

The queue presents each submission one at a time — header capsule front and center, tags below, screenshots in a gallery, and two buttons: "Add to Wishlist" and "Next in Queue". Wishlists are authenticated and saved to the vote ledger.

The leaderboard is deliberately simple. Rank is determined by a weighted score:

Votes are the primary signal. This is intentional — we want the ranking to reflect genuine interest, not committee score inflation. The QA score measures submission completeness (are assets present? is the manifest valid?), and the fairness score checks for balanced evaluation criteria. Both exist to normalize, not to override.

The leaderboard sits below the discovery queue as a compact table — rank, title, votes, status. Click any row to load that game in the queue. It's a live scoreboard that updates as votes come in.

Here's where it gets interesting. Human votes alone are limited by the size of our community. In the early days, we might have a handful of real users browsing the queue. That's not enough signal to make production decisions.

So we built a wishlist simulation engine — a pool of synthetic users whose voting behavior mirrors real Steam market dynamics.

The key constraint: every tag on every submission is real. We use the actual Steam tag taxonomy — the same tags that real developers apply to real store pages. This is what makes the simulation meaningful: the bot voters are reacting to the same classification system that drives Steam's own recommendation engine.

The discovery queue, leaderboard, and simulation engine aren't standalone tools — they're stages in a pipeline designed to answer one question: Which idea deserves to become a real game?

The beauty of this system is that it de-risks production decisions with data, not gut feel. By the time a concept reaches the top of the leaderboard, it's already proven it can capture attention in a store environment, its tag profile aligns with real market demand, and its asset package is strong enough to compete on a real storefront.

That's not a game jam winner. That's a validated product concept.

If you're submitting to the Metal Games game jam, you're not just entering a contest. You're running your concept through a miniature version of the same market forces that determine success on Steam. That means: