The Blog

Okay, before we get into the crowd blog, a couple of news items floating around the web: the Pinocchio Village Haus at the Magic Kingdom will apparently stop serving hamburgers on May 8. As one wag on WDWMagic.com notes, this leaves only 22 other places to get burgers.

Also, Disney’s changing the name of “Priority Seating Reservations” to “Advance Dining Arrangements.” Apparently they got tired of explaining how something with the word “reservation” in it isn’t really a reservation. And the legal department must have had issues with my idea: “I-Can’t-Believe-They’re-Not-Reservations.” Thanks to Mike Scopa at Mouseplanet.com for the heads up. Not only does Mike write a Disney column for Mouseplanet, but he was recently named one of Disney’s “50 Happiest Passholders.” As I write this, he’s in WDW with his son, apparently trying to talk his way into driving the monorail. Good luck with that, Mike.

Finally, Disney’s added several sit-down and counter-service restaurants to the Magic Your Way with Dining plan, as well as new snack options. Most of Epcot is now represented, and Coronado Springs gets a boost. Check out our updated description here.

On to the crowd predictions. First, a look back at what we’ve predicted over the last couple of weeks:

Recall that the “modified prediction” uses an average of four things: historical occupancy rate, number of hours the Magic Kingdom is open in a given week, trips booked for that date in the Unofficial Guide’s free Trip Planner, and WDW resort availability on Expedia.com. The original prediction is from the Guide’s “Best Days” page.

M1 predicted the correct crowd level 2 times out of 9. It was +/- 1 four times, and +/- 2 the remaining three times for which we’ve currently got data. The original method was right zero times, +/- 1 three times, +/- 2 four times, and +/- 3 or more twice. Ouch. On the other hand, M1 was +/- 1 on the crowd level 6 of 9 times, which ain’t a bad first step.

It appeared that the modified method was better than the original method we used, so we’re looking now at ways to improve on the formula used. For simplicity, let’s refer to the modified method and its components as “M1.” Our new method, M2, uses the same inputs as M1, plus the actual number of hours the Magic Kingdom is open that day. It also uses a weighted average for each input, instead of a simple average. Those weights are as follows:

Historical Hotel Occupancy Rate: 0.4
Weekly Magic Kingdom Hours: 1.0
Expedia Hotel Availability: 1.05
Unofficial Guide Trip Planner Stats: 1.4
Magic Kingdom Daily Operating Hours: 1.9

M1′s standard deviation using an unweighted average was 1.80. M2′s standard deviation with the weighted average is 1.04, so the hope is that M2 will be both more accurate and less volatile a measurement, and we can more easily make adjustments to the formula. Incidentally, the standard disclaimer applies: I’m not a statistican, and our regular statistican (the sublime Fred Hazelton) is currently booked on our new transportation and restaurant number-crunching projects for the 2006 edition.

That being said, here are M2′s crowd level predictions for May. Remember that we’re going to eventually switch to a new 1-10 scale. For now, I’ve translated to peak wait times at the MK’s headliner attractions (Space Mountain, Splash Mountain, Big Thunder Mountain Railroad):