How the cutoff moves during a cup
An hour into a three-hour cup, the leaderboard says the qualifying rank holds some number of points – and that number is nowhere near the final cutoff. How far along it is can be measured, because every finished leaderboard can be replayed game by game. This guide shows what 3,690 replayed open queues say, and how the live forecast uses it.
The board at each point of the session
In an open queue, where every team plays its games at its own pace inside the window, the threshold at a given rank climbs almost steadily through the session. Halfway through, it has typically reached 49% of its final value. At the moment the window closes it has reached 92%: the rest is still being played.
Open queue, by the clockSingle lobby, by games played
The numbers behind the chart
| Share of the session gone | Open queue, by the clock | Single lobby, by games played |
|---|---|---|
| 10% | 1% | 14% |
| 20% | 19% | 22% |
| 30% | 30% | 36% |
| 40% | 40% | 38% |
| 50% | 49% | 52% |
| 60% | 58% | 61% |
| 70% | 67% | 70% |
| 80% | 75% | 84% |
| 90% | 84% | 88% |
| 100% | 92% | 100% |
In a single lobby – a final, where the same teams play the same games – the clock that matters is the number of games played rather than the minutes: after half the games, a rank sits at 52% of its final score in the typical lobby.
Why the early board misleads
The curve above is an average, and at the start of a session the cups are far apart. One game in, some teams have played twice and others not at all; the board is sorted by teams that happened to queue first. The table gives the typical spread from one cup to the next at each point: early on, a reading can be more than half again above or below what the average curve would predict.
| Share of the session gone | Threshold reached, typically | Typical spread from cup to cup |
|---|---|---|
| 20% | 19% | ± 63% |
| 30% | 30% | ± 56% |
| 50% | 49% | ± 13% |
| 70% | 67% | ± 11% |
| 90% | 84% | ± 12% |
| 100% | 92% | ± 6% |
That spread is why the live forecast gives the first readings little weight, and why comparing your score to the early board is misleading in both directions: a threshold that looks easy after forty minutes usually is not, and one that looks out of reach often comes back.
After the window closes
An open queue closes on the clock, but a game that started before the close is still played to the end, and Battle Royale games last up to half an hour. So the final standings keep moving after the window: by the numbers below, the top of the board is complete about twenty minutes after the close, and the ranks just below it take the same time to catch up from further back.
| Minutes after the window closes | Ranks 1–25 | Ranks 26–100 |
|---|---|---|
| 0 | 92% | 86% |
| 5 | 95% | 89% |
| 10 | 97% | 94% |
| 15 | 100% | 97% |
| 20 | 100% | 100% |
Share of the final threshold reached, by minutes after the window closes, measured on each game's own end time. The board published by the leaderboard sites can lag a few minutes behind that.
How the live forecast reads the board
While a cup runs, the site reads its public standings every few minutes and turns each reading into a forecast of the end. A reading at a given rank is divided by the share of the final that rank has typically reached at that minute – the curve above, specific to the kind of cup where enough of them have been replayed – which gives what the board says the final will be.
That estimate is then weighed against the forecast made before the cup, each by how precise it has proven: early in the session the history counts for more, late in the session the board does. Towards the end, and after the close, the history's weight is cut further as the cup advances – by how much is learned again on the cups that have finished, since on precision alone it held on to too much of the answer. By then, too, a rank the board does not show follows the nearest rank it does show, unless ranks read on both sides frame it. Two things keep the answer honest:
- A threshold never falls. Teams only gain points, so the points the board already shows at a rank are a floor under the final cutoff, and the forecast never goes below it.
- The ranges are measured on live answers. How wide the range around a live forecast must be, at each point of the session, is measured on past evenings the feed followed, so a range that says nine finals in ten land inside it is the width that did.
Replayed on a held-out week of September, between a third and two thirds of the session, the live forecast was about 3% off the final threshold in median, and closer still in the last hour.
What to do with it
- Before the cup: note the forecast at your cut and its range. That is your target, not the first numbers on the board.
- Mid-session: compare your total with the share of the target you should have by now. Halfway through a cup, being at half the target is on pace, not behind.
- Last hour: the live forecast is at its sharpest. If you are just under the line, the board will still rise after the close: plan to finish above the forecast, not above the current board.
Updated 22 October 2026