steam-page-stats

Fetch a Steam game's public page stats and apply the Boxleiter rule of thumb to estimate lifetime revenue from the review count. A plain version of a useful heuristic, with its uncertainty shown.

pip install steam-page-stats
steam-page-stats 1145360

Hades  (appid 1145360)
  developer: Supergiant Games
  publisher: Supergiant Games
  genres: Action, Indie, RPG
  released: 17 Sep, 2020
  price: $24.99
  reviews: 205,000 total
  review score: 97.56%  positive

Boxleiter rule-of-thumb revenue estimate:
  low (×30):     $153.7M
  median (×50):  $256.1M
  high (×63):    $322.6M

  ⚠  This is a heuristic with ~24% of games off by >30% per the formula's
     own author. For a calibrated P10–P90 launch cone (~82% nominal coverage,
     81–86% realized per wishlist tier on 6,422 held-out launches), see
     https://steamforecast.app

Why this exists

The Boxleiter method (Mike Boxleiter, 2010s) is the de-facto rule of thumb for estimating lifetime Steam revenue from public page data:

revenue ≈ review_count × multiplier × price

Modern multiplier estimates range 30–63 sales per Steam review, with significant per-genre variance.

This package gives you:

• A multiplier-bracketed estimate (low / median / high) so you see the uncertainty inherent in the heuristic
• A clean Python API for use in your own analysis pipelines
• A CLI for quick one-off lookups

The package is deliberately small. It does not try to be a calibrated forecaster — for that, you need wishlist-tier stratification, conformal prediction intervals, and a labelled corpus to validate against. That's a hard problem; the rule of thumb is not.

If you need calibrated revenue ranges validated on a held-out launch corpus (~82% nominal coverage, 81–86% realized per wishlist tier, n=6,422), plus causal estimates for marketing levers, see the full forecaster at SteamForecast — methodology and per-tier coverage stats are open at /methodology.

Install

pip install steam-page-stats

Or with optional dev dependencies for tests / linting:

pip install "steam-page-stats[dev]"

CLI usage

# human-readable
steam-page-stats 1145360

# machine-readable JSON
steam-page-stats 1145360 --json

# just page stats, skip Boxleiter
steam-page-stats 1145360 --no-boxleiter

Python API

import asyncio
from steam_page_stats import fetch_page_stats, boxleiter_estimate

async def main():
    stats = await fetch_page_stats(1145360)
    print(f"{stats.name}: {stats.review_count_total:,} reviews")

    if stats.review_count_total and stats.price_cents:
        est = boxleiter_estimate(stats.review_count_total, stats.price_cents)
        print(f"Lifetime revenue estimate: ${est.revenue_low_dollars:,.0f} – "
              f"${est.revenue_median_dollars:,.0f} – ${est.revenue_high_dollars:,.0f}")

asyncio.run(main())

For batch use, instantiate the client directly to share the underlying HTTP connection pool:

from steam_page_stats import SteamPageStatsClient

async def fetch_many(appids):
    async with SteamPageStatsClient(throttle_s=1.0) as client:
        for appid in appids:
            stats = await client.fetch(appid)
            yield stats

Rate limits + etiquette

This package uses Steam's public, unauthenticated Storefront and appreviews endpoints. Steam tolerates roughly 200 req/5min from a single IP across these endpoints. The default throttle of 1 req/sec keeps you well below that.

We send a transparent User-Agent header so you're identifiable per RFC etiquette:

steam-page-stats/0.1 (+https://github.com/GC108/steam-page-stats; calibration-sanity-check)

If you'd rather not advertise this package, override it:

async with SteamPageStatsClient(user_agent="my-research-bot/1.0") as c: ...

Limitations

The Boxleiter heuristic is structurally biased on:

• Top-decile breakouts (Peak, Webfishing, etc.) — single-multiplier can't capture viral discovery. Per the formula's own author, ~24% of games are off by more than 30%.
• Free-to-play with cosmetic monetization — the formula assumes revenue ∝ price × sales; F2P breaks that.
• Long-tailed mid-tier where individual conversion ratios diverge by 10–20× at comparable wishlist counts (see steamforecast.app/methodology for the variance data).

For these cases, a calibrated cone with validated coverage beats a single-number heuristic. The rule of thumb is a useful sanity check, not a budget-decision tool.

Development

git clone https://github.com/GC108/steam-page-stats
cd steam-page-stats
pip install -e ".[dev]"
pytest
ruff check .

License

MIT — see LICENSE.

Related

• SteamForecast — calibrated launch revenue cones (~82% nominal coverage, 81–86% realized per wishlist tier on 6,422 held-out launches), causal marketing-lever estimates, and Total Lift Attribution to recover paid campaign wishlists Steam's UTM dashboard misses.
• Steam Web API documentation
• Boxleiter method, Mike Boxleiter (2014, updated 2023) — the original formulation and 2023 retrospective.