Prediction Accuracy
We believe in transparency. Here's how our crowd predictions compare to actual crowd levels measured from real wait time data. We publish these stats because we want you to trust our predictions.
Frequently Asked Questions
How do you calculate the actual crowd level?
At the end of each day, we analyze all wait time data collected throughout the day. The average wait times across all attractions are converted to a crowd level score (1-10) using the same methodology as our predictions.
What does "within 1 level" mean?
If we predict a crowd level of 6 and the actual level is 5, 6, or 7, that counts as "within 1 level." This is the most useful accuracy metric because a 1-level difference rarely changes your planning decisions.
Why might predictions be off?
Unexpected events like ride breakdowns, weather changes, or special announcements can affect actual crowd levels. We continuously improve our model by learning from these variations.
How can I help improve predictions?
Simply by visiting the parks! The more real-time data we collect from actual wait times, the better our predictions become. Our model learns and adapts over time.