Steam Next Fest 2026
What moves wishlists across 149 roguelike-deckbuilders
Which tags move wishlists?
Median wishlists of games carrying a tag vs. the dataset baseline (median ). Log scale · all games · tags on ≥ games.
▲ green = above baseline ▼ rose = below
Wishlist distribution · all games
A power law: a few breakouts, a long tail.
Followers vs. wishlists
Log–log.
Tag clusters
t-SNE map of tag profiles · click a point for details
126 games
No games match the current filters.
Wishlists by cluster
Which sub-genres pull the most interest.
By genre
Median wishlists (genre is low-variance here).
Localization vs. wishlists
Median wishlists by language support. Mostly reverse causality — broad localization signals budget/ambition, not a guaranteed lever. Log scale.
What sets breakouts apart
Tags over-represented in the top 10% () vs. everyone else (percentage-point gap).
Opportunity map
Competition (cluster size) vs. performance (median wishlists). Upper-left = high interest, less crowded.
Steam features vs. wishlists
Median-wishlist lift for varying features. ● stable / ○ tentative (small n). Log scale.
| Title |
Cluster |
Wishlists ▾ |
Followers |
M1 pred. |
Fol/Wish |
Generated from next-fest-data-consolidated.csv · games · clustering seed 42, k=.
Data caveats: copies sold / revenue / reviews / country split exist for only 1 game and are excluded.
Followers are imputed as wishlists/12 for of games (only measured); M1 is a model output ≈ a fixed fraction of wishlists.
Price is unset for /. Wishlists is the only independent outcome signal — tags & genres are content metadata.