How our crowd estimates work

This page explains the two calculation methods and what they can and cannot establish.

How we estimate Disneyland crowds

Our original short-range model and separate annual planning rules. No competitor forecast ratings, wording, or formulas are imported.

Three crowd levels across the calendar year

Lighter

Less queue pressure expected. Popular rides can still have long waits.

Busy

Expect queues. Prioritize your attractions and allow more time.

Very busy

Allow substantial extra time, especially around major holidays.

The three inputs for the annual outlook

Day-of-week pattern: provisional points are 20 Monday–Thursday, 35 Friday, 45 Saturday, and 40 Sunday. These are our starting assumptions, not fitted historical findings.
Holidays and published school breaks: add 35 points on a federal holiday/observance, or 15 on an adjacent day; add 20 for a listed school break, with this component capped at 50. School coverage currently uses one LAUSD school’s published calendar and is incomplete.
Dated park events: add 10 for an announced event period, 20 on its opening date, or 15 for a separate-ticket event; overlapping events use the highest of these values. These point effects are assumptions, not measured crowd impacts.

Total annual outlook points are capped at 100: 0–30 = Lighter; 31–65 = Busy; 66–100 = Very busy. They are not the short-range queue index. Unannounced 2027 park event dates are left unknown. The full current and following calendar years are supported for future dates.

Sources: US federal holiday rules, LAUSD Harvard Elementary’s published school calendar, and the official Disney event pages linked from each date. These facts are sourced; their assigned crowd effects are ours and have not been validated.

Start with comparable historical park days

For each completed day, we take each eligible attraction’s minute-weighted median posted standby wait during published park hours. We then take the median across a consistent basket of attractions. This park-day queue index is the target we estimate—not actual attendance or an individual ride wait.

We require at least three completed history days and the same qualifying ride basket across them: at least eight attractions at Disneyland Park or six at California Adventure. Each ride needs 240 operating minutes, no more than 10% unknown scheduled minutes, and a median between 0 and 180 minutes. Missing records are never filled with zero.

Parameter	Our initial rule	Evidence and limitation
Recent wait history	Recency weight = 2 raised to (−age in days / 7).	Real completed-day records from ThemeParks.wiki. The seven-day half-life is our design choice, not a fitted result.
Comparable weekday	2× for the same weekday; otherwise 1.25× for the same weekday/weekend group; otherwise 1×.	Weights increase the influence of more comparable days. Six days cannot establish reliable weekly seasonality.
Ticketed-event status	1.5× similarity weight when historical and target days share party/no-party status.	Uses sourced Oogie Boogie Bash dates for California Adventure. We do not assume a party causes a fixed crowd increase.
Forecast value	Weighted mean of the historical park-day queue indices.	No competitor ratings are inputs. Separate histories and attraction baskets are used for each park.
Crowd score	1: ≤10 min; 2: ≤15; 3: ≤20; 4: ≤25; 5: ≤30; 6: ≤35; 7: ≤40; 8: ≤45; 9: ≤50; 10: >50.	Original fixed queue-index bands. They are not historical attendance percentiles, probability estimates, or calibrated park crowd levels.
Planning range	Base half-width = max(10 min, twice historical median absolute deviation), plus 0.75 min per day ahead, 3 min if no matching weekday, and 5 min if no seasonal match.	A conservative rule, not a statistically calibrated confidence interval. Seasonal transitions widen the range rather than inventing a demand uplift.
Confidence and horizon	Low, or very low if no same weekday/season match or more than seven days ahead. 14-day horizon; snapshot expires after 48 hours.	Confidence labels are qualitative. More history and prospective evaluation are needed before longer-range forecasts.
Not currently modeled	The short-range model does not fit school breaks, weather, ticket prices, pass blockouts, planned closures, seasonal demand, or attendance. The annual outlook uses separate provisional school/holiday/event rules, not fitted effects.	These require reliable inputs and fitted effects. Event listings and operating hours remain context; no unsupported numeric weights are assigned.

Sources, updates, and attribution

Powered by ThemeParks.wiki. Wait records and operating schedules use its public API. Seasonal events link to Disney’s announcements. Refreshing current observations does not retrain forecasts. Each model run is timestamped and logged. Future-date estimates are saved before those dates; today’s estimate is labeled as a same-day estimate, not a forecast made before the day began. The daily local refresh rebuilds both short-range estimates and the annual outlook. The computer and Codex must be available for local scheduled runs; this is not a hosted background service.

A wait observation is not a prediction. Cached current waits older than 24 hours are hidden; forecasts expire after 48 hours. Ride last-changed times are not retrieval times. Recheck Disney’s official schedule before visiting.

Early historical checks, including the baseline

These are expanding-window hindcasts: each held-out day uses only earlier days, including basket selection. They are not forecasts saved before those dates occurred. Three checks per park are too few to claim accuracy or superiority.

Disneyland Park

6 completed history days · 11-attraction sample.

Short-range historical check: 3 held-out days. Mean error 11.3 min; simple prior-day mean baseline 11.2 min.

The annual outlook uses separate calendar rules and has not been backtested. This tiny short-range check does not validate annual predictions.

Attractions in the short-range sample

Big Thunder Mountain Railroad
Haunted Mansion Holiday
Jungle Cruise
Matterhorn Bobsleds
Mickey & Minnie's Runaway Railway
Millennium Falcon: Smugglers Run
Peter Pan's Flight
Pirates of the Caribbean
Space Mountain
Star Wars: Rise of the Resistance
Tiana's Bayou Adventure

Disney California Adventure

6 completed history days · 9-attraction sample.

Short-range historical check: 3 held-out days. Mean error 8.3 min; simple prior-day mean baseline 10 min.

The annual outlook uses separate calendar rules and has not been backtested. This tiny short-range check does not validate annual predictions.

Attractions in the short-range sample

Goofy's Sky School
Grizzly River Run
Guardians of the Galaxy - Mission: BREAKOUT!
Incredicoaster
Monsters, Inc. Mike & Sulley to the Rescue!
Radiator Springs Racers
The Little Mermaid - Ariel's Undersea Adventure
Toy Story Midway Mania!
WEB SLINGERS: A Spider-Man Adventure