First, a request for assistance: We’re in need of a computer programmer familiar with PHP, MySQL, and SQL (the language) to help with an upcoming project. If you’ve got a few free hours available each week for the next couple of months (perhaps 4 to 8 hours per week, as a guess), and know how to handle data files potentially several GB in size, drop me a line with your resume and a couple of contacts. Compensation? Our undying love and respect, plus the fame and glory of possibly seeing your name in the next edition of the book. Okay, maybe we can throw a few bucks your way, too, but you ain’t gonna get rich. Do it for the love and respect.

Updated version of the crowd calendar is now available. I haven’t checked day-for-day, but it appears some of the predictions have decreased by 1 on the scale with the new revision. A couple of factors that may have caused this: (1) crowd levels have been running slightly lower than our predictions since about mid-August, based on the data we’ve collected in the Magic Kingdom. Might be kids going back to school. Might be high gas prices. Disney announced last month that year-over-year park attendance for WDW was down 1% in the latest quarter, but didn’t offer any explanation. I’m going with gas prices for now, just as a hunch.

Second reason for new crowd calendar is that we’re using a slightly different formula to predict the crowd levels. As you might suspect, we periodically go back and check how well our predictions match up with what we actually observed in the parks. As part of that process, we also try to see whether we can come up with another method whose predictions better match what we’ve seen. So if we’ve got some method of predicting crowd levels – let’s call it M1 – and we know its standard error for the observations we have is 17.80, can we come up with another method – call it M2 – with a standard error less than 17.80 for the same observations?

Once again, Unofficial Guide statistician Fred Hazelton comes to the rescue. Fred suggested adding a coefficient to each parameter we use in the crowd calculations, and varying the coefficient depending on how well it correlates to the observed wait times. So the new formula’s a weighted average. Where’d the weights for each parameter come from? We used Excel’s “solver” add-in (gotta love the simplex method!) to find the weights so that the standard error was minimized. On a daily basis, the standard error translates roughly to 8 or 9 minutes either way (if I’m interpreting that correctly). We’ll see how closely the actual performance is.

Next, a couple of Magic Kingdom park observations from recent Saturdays is below. Note that the day before Saturday, Sep 10, Hurricane Ophelia was churning off the coast of central Florida, which may have caused some folks to stay home for the weekend. With these new measurements, we’re currently running an error of just over 7 minutes on average.

Predicted Peak Wait Time at MK Headliner Attractions
Actual Peak Wait Time
Difference from Predicted Peak
Sat Aug 13
75-84 minutes
60 minutes
Sat Sep 10
45-54 minutes
30 minutes

Finally, we started testing the average wait time per attraction on all of the touring plans in the book and website earlier this summer, at all four Disney theme parks. Results are encouraging. For touring plans that begin at opening in any park, average wait times are in the 4 to 12 minute range. This includes touring plans for parents with small children, adults, and seniors. Maximum waits are tough to predict, because of the occasional odd attraction like the Liberty Belle Riverboat and live performances. (For example, if the Liberty Belle Riverboat leaves on the hour and half-hour, and you get there at 2:31 pm, you’ve got a 29 minute wait until the next boarding.) If you’d like to participate in this test, let me know.

P.S. The 2006 Unofficial Guide left off the names of three excellent UG researchers. We’d like to thank Richard and Lillian Macko, and Darcie Vance for their hard work throughout the summer in Walt Disney World. You folks are the best. Remind me to buy you lots of food during the October trip.


Comments are closed.