Scoring rules
Last updated 5 November 2026 · These are the rules the app runs on right now. When the rules change, this page changes with them.
In short
Every challenge has a leaderboard per series, and the leaderboard ranks runs. On a Measured challenge your run is placed by its result — how many, how long or how fast — in the direction the challenge measures. Runs that reach the challenge's target come first, runs that fall short come after them, and runs the community voted down are labelled Disputed and sit at the bottom. Among equal results, the run whose Legit votes are more convincing wins.
On a Showtime challenge there is no counter: the crowd decides. Runs are ranked by their respect rate — how much of the audience in the app feed gives the run Respect, weighted by how much of it they watched. See 5. Showtime challenges.
- The result
- How a leaderboard is ordered
- Ties: why 48 of 50 beats 1 of 1
- Disputed runs
- Showtime challenges
- Votes and Respect
- Series and the All time board
- What moderators can change
- A worked example
- Changes to these rules
1. The result
Challenges are curated by OutplayMe (you can suggest one from the Challenges tab). Sections 1 to 4 describe Measured challenges; Showtime challenges have their own ranking in section 5. Every challenge says what it measures:
| Type | What counts | Better is | Example |
|---|---|---|---|
| Count | Successful repetitions or scoring actions — taps, baskets, flips. The creator can limit counting to the first N seconds of the run ("in 10 s"). | more | "4 taps in 10 s", "10 free throws" |
| Time held | How long a position is held, in seconds. | longer | "Hold a plank for 4:00" |
| Fastest time | How long the run takes, in seconds. | shorter | "Solve the cube under 12 s" |
| Not measurable | A yes-or-no feat with nothing to count. | — | "Land a backflip" |
Every run is watched by our AI verification, which counts or times it against that goal and writes the number down as the run's result. You see it on the run as a chip next to the challenge — "9 taps · target 4" or "0:08 / 4:00" — and as the main number of every leaderboard row.
A moderator can correct a result after watching the run (for example when the AI miscounted a very fast sequence). A corrected result is final and is marked as set by a moderator.
A run that does more than the challenge asks for scores more: the target says what the challenge asks, the result says what you did.
2. How a leaderboard is ordered
1. Band target reached (or challenge not measurable)
→ target not reached
→ Disputed
2. Result count / time held: higher first · fastest time: lower first
(a run without a result yet comes last within its band)
3. Ties more convincing Legit share (see 3.)
→ more Legit votes → more Respect → earlier upload
- One row per run. If you post ten runs, you are on the board ten times; your weaker runs rank lower. Only verified runs are on a board.
- Target not reached runs are still ranked by result among themselves — a 3:50 plank beats a 2:10 plank — but every run that reached 4:00 is above both. They carry a small "Target not reached" label.
- A result of 0 is a result: it ranks among the runs that did not reach the target, by votes.
- Not measurable challenges have no bands except Disputed; their runs are ordered by the tie-break alone — by how convincingly legit they are.
- The Respect toggle on a board re-orders the same rows by Respect only.
3. Ties: why 48 of 50 beats 1 of 1
Among runs with the same result, the Legit / Not legit votes decide. But a raw percentage would make a single friendly vote worth more than fifty honest ones: 1 of 1 is 100 %, 48 of 50 is "only" 96 %. So the board sorts by the lower bound of the Wilson score interval of the Legit share — in plain words, "how high a Legit share can we claim with confidence, given how many people have voted". With few votes that bound stays low; with many it approaches the real share.
| Legit / Not legit | Raw share | Wilson lower bound | Order |
|---|---|---|---|
| 48 / 2 | 96 % | 0.87 | 1st |
| 15 / 1 | 94 % | 0.72 | 2nd |
| 1 / 0 | 100 % | 0.21 | 3rd |
| 0 / 0 | — | 0 | last among equals |
The app shows you the raw share and the number of votes ("94 % · 16 votes"); the Wilson bound only decides the order. If you like formulas:
n = Legit + Not legit, p = Legit / n, z = 1.96 (95 %) lower bound = ( p + z²/2n − z·√( p(1−p)/n + z²/4n² ) ) / ( 1 + z²/n )
4. Disputed runs
Votes are the community's check on a result. A verified run becomes Disputed when fewer than 40 % of its votes are Legit and it has at least 5 votes. (One angry vote on a fresh run does nothing.) The run then:
- stays public, with the Disputed label on the run and on the leaderboard;
- drops to the bottom band of every board it is on — a doubted 12-tap run cannot hold first place while nobody has checked it;
- goes to our moderators, who watch it and either reject it (it leaves the board and the author is told why) or confirm it as legit, after which the label disappears, the run returns to its place by result, and votes can no longer flag it.
The label is recomputed on every vote, so a run can become disputed later, or stop being disputed when more Legit votes come in. Moderators can change the 40 % and 5-vote thresholds; this page states the values in force.
5. Showtime challenges
Showtime runs are up to 30 seconds and are judged by the people who watch them. The board ranks them by respect rate:
score = (counted Respect + 2) / (counted views + 20) highest first
- Only the feed counts. A view counts when the app's feed showed you the run. Watching it on a profile, a challenge page, a leaderboard or through a shared link still adds to the run's public Respect count, but does not move its rank. Your own views never count, and sharing your run is never penalised.
- How much you watched matters. Under 2 seconds counts nothing; from 2 seconds to under half the run counts 0.3; half to under 90 % counts 0.6; 90 % or more counts a full view. Replays never add more; watching further later adds only the difference.
- Respect weighs what your view weighs. A Respect from someone who watched the whole run counts 1, from someone who watched three seconds 0.3.
- Everyone starts fair. The +2 and +20 give every new run the same starting point (a 10 % rate), so one early Respect can't put a run on top. Once a challenge has 20 established runs, the +2 becomes 20 × that challenge's real respect rate, updated every night.
- Fresh. A legit run joins the board right away. Until it has 50 counted views it shows a Fresh label instead of a number. Places 1 to 3 are for runs with at least 50 counted views; a Fresh run can climb as high as 4th. While a board has fewer than three such runs, the best Fresh runs fill the open places.
- Legit first. Legit / Not legit votes use the same weights. A Showtime run becomes Disputed under the same thresholds as in section 4 and drops to the bottom. Ties go to the more convincing weighted Legit share, then Respect, then the earlier upload.
- Series and all time. The score is a share, so time on the board gives no head start: the All time board uses every run's current score, and a run posted today can beat one from day one. Each series crowns its own winner — when it ends, the scores of the runs posted in it are frozen on that series' board.
Scores are updated every minute.
6. Votes and Respect
- Legit / Not legit: one vote per signed-in user per run. You can change or remove it. Only verified runs can be voted on, and you cannot vote on your own run.
- Respect: one per signed-in user per run, can be taken back, never on your own run. On Measured boards Respect never changes the Result order; it is its own board and the last tie-break. On Showtime boards, Respect from the feed is what ranks (section 5).
- Views are counted once per person per day. They rank runs only on Showtime boards, and only views from the feed (section 5).
- Your profile's Score (Legit ÷ all votes over all your runs) is a credibility number, not a ranking. Your Best rank is the best position any of your runs holds on any board.
7. Series and the All time board
Every challenge runs in series (seasons) of 90 days by default; when a series ends, the next one starts a few days later with a fresh board. A run belongs to the series that was live when it was uploaded. Next to its series, every challenge has an All time board that ranks every run from every series by the same rules.
Final boards are frozen. When a series ends, its leaderboard is locked: places and numbers (result, Legit and Not legit votes, Respect, Disputed) stay exactly as they were at the end, so votes cast later can't rewrite a finished season. If a run is removed afterwards (for example rejected by a moderator, or its author's account is deleted), it drops off and the places below move up. A run uploaded during the series but approved after it ended is added with its numbers at approval. The All time board combines the frozen boards of finished series with the live board of the current one.
8. What moderators can change
- Correct a run's result after watching it (final).
- Reject a run, or confirm a disputed run as legit.
- Create challenges, write the rules you read and the separate instructions the AI judges by, and edit what a challenge measures; re-run the AI on its runs afterwards.
- Set the Disputed thresholds.
Moderators never add votes or Respect, and never move a run above another by hand — only the rules above do that.
9. A worked example
Challenge "Tap with the finger", target 4 taps in 10 s. Four runs:
| Run | Result | Legit / Not legit | Place | Why |
|---|---|---|---|---|
| Bor | 9 taps | 30 / 3 | 1 | best result among runs that reached the target |
| Ana #2 | 7 taps | 15 / 1 | 2 | same result as Cene, but 15 of 16 is more convincing than 1 of 1 |
| Cene | 7 taps | 1 / 0 | 3 | one vote is not enough to beat sixteen |
| Dani | 3 taps | 8 / 0 | 4 | target not reached — below every run that reached it |
| Ana #1 | 12 taps | 2 / 8 | 5 | 20 % Legit at 10 votes → Disputed, bottom of the board until a moderator decides |
If a moderator watches Ana's 12-tap run and confirms it, it moves to first place. If the moderator rejects it, it leaves the board.
10. Changes to these rules
- 5 November 2026 — every challenge has an All time board from its first run, not only after two series (it used to be called Overall).
- 30 October 2026 — Showtime challenges rank by respect rate from the feed, weighted by how much of each run was watched, with the Fresh label and places 1 to 3 kept for runs with 50 counted views.
- 23 October 2026 — leaderboards of finished series are frozen at the end of the series.
- 16 October 2026 — the result became the main counter; boards rank runs instead of players; bands, Wilson tie-break and Disputed introduced. Challenges became curated: created by OutplayMe, suggested by you.
- 12 October 2026 — leaderboards moved to per challenge and per series; the global board was retired.
- 2 October 2026 — you can no longer vote on or give Respect to your own run.
Questions about a ranking? Write to hello@outplayme.app with the challenge and your username.