Skip to content

Reviews & ratings

PostRoll reviews are verified-attendance only: you can only leave a review for a game you actually paid a seat on. Reviews drive the star rating that shows on every listing card + storefront page.

We use Bayesian smoothing with a prior mean of 4.0 and a prior weight of 10 reviews. In practice:

  • A listing with 5 reviews averaging 5.0 scores about 4.33 — not 5.0, because we don’t trust five reviews to define the truth.
  • A listing with 100 reviews averaging 5.0 scores 4.91 — closer to the observed average as the sample grows.
  • A listing with 5 reviews averaging 1.0 scores 3.0, not 1.0.

This means a single 5★ doesn’t run the chart on a brand-new listing, and a single hit-job doesn’t tank an established GM’s rating. Both effects fade as the review count grows past ~20.

  1. Open My purchases → Library.
  2. Find the game you played. Tap Leave a review.
  3. Pick an overall 1-5 star rating.
  4. Write a short review (optional, up to 4000 chars).
  5. Optionally score Prep quality, Fairness, Energy at the table, and Accessibility on the same 1-5 scale.
  6. Mark Would recommend? (optional).
  7. Post.

You can edit your review for 14 days after posting. After that the review freezes — you can still delete it, but you can’t keep editing a review the GM has already responded to.

If you’re the GM, every review on your listing shows up in the listing editor → Reviews section. Hit Reply to post a single public response. You can update or clear it later.

The reply notifies the reviewer in-app so they see your response without having to revisit the page.

Hit Report on a review or a whole listing if something violates PostRoll’s guidelines (spam, harassment, off-topic, fake review, private info, etc.). Reports go to the outpost’s mod queue and, when needed, escalate to PostRoll staff.

Repeat reports by the same user against the same review update the existing report rather than stacking — be specific so mods have something to act on.

Listings sorted by Top rated on the marketplace use the Bayesian-smoothed average. The simple rating_avg (no smoothing) also surfaces on the listing card but isn’t what the ranker reads — smoothing is what keeps brand-new listings from leapfrogging established GMs on the back of a single five-star.