Magic Kingdom
- Our error
- 5.87 min
- Industry feed
- 8.58 min
- Within 10 min
- 81.7%
- Matched forecasts
- 6,514
Loading…
Loading…
Accuracy Report
Graded against what actually happened: 21,989 forecasts over 30 days, September 12, 2026 to October 11, 2026. Recomputed nightly. Nothing here is hand-picked.
Average error
6.82 min
Lower is better
Industry feed, head to head
9.74 min
On the 13,647 forecasts we both made, we average 7.03 min to their 9.74 min — 28% less error
Forecasts graded
21,989
across 30 days
The app adds free real-time wait alerts, plus a Premium day planner.
Download WaitScout| Park | Our error | Industry feed | Within 10 min | Matched forecasts |
|---|---|---|---|---|
| Magic Kingdom | 5.87 min | 8.58 min | 81.7% | 6,514 |
| EPCOT | 6.68 min | 9.1 min | 79.5% | 2,972 |
We do not win every forecast — we win 52.4% of them
That is measured on the 13,647 forecasts where the industry feed also published a number, which is the only set where a head-to-head question has an answer. It covers 62% of what we forecast; the feed does not cover the rest, so nobody can be compared there. We lose 47.6% of individual matchups, and our average error is lower on top of that — when we are wrong, we are wrong by less. A page that only showed the flattering number would not be worth reading, so both are here.
Every forecast is timestamped when we issue it. After the day closes, we compare the predicted wait for each ride and hour against the realized median wait we recorded for that same ride and hour. The gap between the two, in minutes, is the error. We average it. That is the whole method.
We report error in minutes, not as a percentage. A percentage makes short waits look catastrophic and long waits look excellent: being 5 minutes off a 10-minute wait scores far worse than being 20 minutes off a 90-minute wait, even though the second miss is the one that ruins an afternoon. Minutes are what you actually stand in.
The comparison column is the same forecast horizon scored against the public industry feed. It is not the same population as our own error figure, and we are not going to pretend otherwise: the feed publishes a number for about six in ten of the slots we forecast, so every head-to-head figure on this page is computed only on the slots where both of us committed to a prediction — same rides, same hours, same days, both sides present. Our own average error above covers everything we served, including the slots the feed left blank. Rides that were closed are excluded — predicting a closure is a different problem, and we do not get to count it here. Nothing is dropped for making us look bad.
78.8% of WaitScout forecasts land within 10 minutes of the realized wait, graded nightly against 21,989 forecasts over 30 days.
Every forecast is timestamped when issued, then compared after the day closes against the realized median wait for that ride and hour. The gap in minutes is the error, averaged across all graded forecasts. Rides that were closed are excluded.
On the 13,647 forecasts both WaitScout and the industry feed made, WaitScout wins 52.4% of individual matchups.
| Hollywood Studios | 8.38 min | 10.59 min | 72.8% | 2,654 |
| Animal Kingdom | 10.34 min | 14.46 min | 74.7% | 1,507 |
Every figure in this table covers the forecasts we and the industry feed both made, so the two error columns describe the same days and the same rides. Our error across everything we served, including the rides the feed does not forecast, is the average error shown at the top of this page.
Right now Animal Kingdom is our weakest park (10.34 min average error) and Magic Kingdom our strongest (5.87 min). Both are listed above at whatever the data says today.