Here's our crowd report for the week of Dec 7-13, 2014.
Same report for week of Dec 14-20, 2014.
If you want to see how actual waits compared to our predictions, on an attraction-by-attraction basis, we show those numbers on our
crowd calendar every day for the previous day (top of the page, where it says "How we did yesterday").
There's some truth to this, but it's limited to certain scenarios. Our models have shown some problems dealing with combinations of scenarios it hasn't seen before. One example was last November at the MK, where the combination of MVMCP, MVMCP, Evening EMH, and MVMCP on Thu, Fri, Sat, and Sun, was something we hadn't seen before. But after a couple of those, we updated the models and got better results.
Things like free dining and other sales tend to be easier to predict.
I look at crowd calendar predictions and results every day. Neither we nor Josh at ******* are guessing, by any reasonable definition of the word. Both have misses, sometimes large, but on a day to day basis they're far better than random guessing would get you.
The thing you want to look for in a calendar, is whether there's any objective, publicly verifiable definition of what the scale means. So let's say Moe's Tavern puts out a calendar that says the Magic Kingdom is going to be a 6 today and a 4 tomorrow. If Moe tells you, for example, that a '6' means a wait of 60 minutes at Space Mountain at noon, and a '4' means a 40 minute-wait at Space Mountain at noon, then you've got something to go on. AFAIK, only 2 calendars do that.