Bullet Heaven

A price band needs 30 launches to be read on this page (the Sample filter; never under 10). Below that the page says "too few" instead of a figure, and under 30 launches or 5 hits a chance is written as a count, such as "10 of 20 launches".

Even at 30 launches, one extra hit moves the chance by over 3 in 100, so treat close figures as equal. Here: $10 to $20 has 20, $20 to $40 has 3. Pairs need 30+ launches to be ranked: 5 of 6 qualify.

Revenue is estimated from reviews times price, so a dearer game reaches $100k with fewer reviews. That makes odds climb with price in every genre. All genres: 1 in 230 under $5, 1 in 28 $5 to $10, 21 in 100 $10 to $20, 44 in 100 $20 to $40.

So compare odds within a price band. "Expected at its prices" is the chance this genre would have if each of its launches did as well as all games at the same price: 1 in 18.

A hit is an estimated first-year revenue of $100k or more: store reviews x a reviews-to-sales multiplier x price x the share of list price actually paid after discounts, with a regional price adjustment (games sold mostly to Chinese-speaking players earn less per sale, because regional prices are lower).

Sales on Steam before Steam's cut; about 58% of it reaches the developer.

Nobody publishes the true multiplier. Across its likely range the all-genre chance moves from 4.8 to 8.4 in 100, and a third of all hits sit within 2x of the $100k line. Read every chance as "about": this genre is about 14 in 100.

Half of the launches earned more, half less. One huge hit cannot move it, which is why this page uses medians, not averages.

The all-genre figure beside it is the cohort median ($562, 26,758 launches). "Typical, last 4 quarters" further down is a different measure, named for what it is.

Typical figures are shown from 100 launches; below that the page shows the chance of clearing $1k, $10k and $100k.

With about 300 genres, chance alone gives the best one a lead near 1.8 times its expected odds (2.05 times for the top 5 in 100).

Bullet Heaven beats its prices about 2.6 times (actual 14 in 100, expected 1 in 18), over 2.05 times, more than luck alone usually produces. There are many more pairs than genres, so the Best pairs card uses higher lines: 2.37 times, and 2.56 for the top 5 in 100.

Each lens reruns the figures on a slice of the launches. Standard is the base: every launch in the cohort. It is not a lens that can hold or move.

A lens moves the answer when its chance of $100k+ differs from the standard by more than a random slice of the same size would, with 5+ hits on both sides, or when its typical earnings move over 25% with 100+ launches in the lens. Without the breakout moves it when one game earned over half the money. With fewer hits it is too few to tell. Pips: green holds, coral moves, grey too few or not run.

Lens figures cover all prices; they are not split by price band yet.

The West-facing view keeps games mostly reviewed outside China: it sets aside games with 35% or more of their reviews in Chinese (2 of 142 here).

Review language is measured only for games with 2,000+ reviews; for smaller games it is guessed from the game's name. The audience split is genre-specific: leaving those games out changes typical earnings 53% in FMV and 19% in Dating Sim, yet barely moves the all-genre chance (6.1 to 6.0 in 100).

Here typical earnings move -8.8%, inside the 25% line; 9.4% of its revenue-weighted reviews are in Chinese.

Last 18 months keeps launches 18 months old or newer (52 of 142). Tastes move: only 9 of the top 15 genres by hit rate stay top 15 on recent launches.

When the two disagree, prefer the recent figure. Here: 12 in 100 against 14 in 100, typical $8,423 against $8,217: it holds.

A genre is one-hit when one game holds over half its money. Across genres the top game holds 29% in the median genre and the top tenth of games 94%; 57 genres have one game over half.

No single game dominates: the biggest, Megabonk, holds 45% of its money. Totals and shares are safe to read here. Without it, 13 in 100 still hit $100k+.

New launches count every store game with the tag released in each 12 months, not only the indie cohort, from the store data of 11 Nov 2026.

459 against 86 is +433.7%: more games now compete for the same players. Earlier years and a life-cycle stage are not computed yet.

834 store pages carry the tag and can be wishlisted. 36 are on Steam's wishlist chart.

A date counts as verified only when the release-calendar job checked it against the live store, for the next 90 days. Other dates are as the store lists them.

Rising demand: games here hold their players better than similar games of the same age (reviews, games launched since 2021), Aug to Oct 2026 against 2025, from 51 games. It describes; the backtest found it does not predict launches.

Ranks are among 116 (Launch success), 135 (Most anticipated), 107 (Rising demand), 126 (Typical earnings) genres. Typical, last 4 quarters ($11,661, 81 launches) uses launches of 2024Q4 to 2025Q3 only, so it differs from the cohort median.

Lens figures by price band: lens figures are not split by price yet.

Life-cycle stage: needs launch counts further back than two years.

Gaining fast: follower readings started 30 Sep; first readings 13 Oct.

An outside check: the backtest scores predictions against the same review-based estimate, so it shows review counts are predictable, not revenue. Fixing that needs outside sales data.