Hello @len I’m hoping you can address this question.
Many people have been complaining about crowd levels being off since the fall and while I understand some frustration, I have to admit that my head is fairly near to exploding over the fact that people were actually expecting today to be a 5, with a major holiday tomorrow.
Fall includes MNSSHP and MVMCP, and the dates for these are not stagnant. Neither is the date of Thanksgiving (which always happens on the 4th thursday but can happen on a wide range of actual dates). I’m curious as to whether the software TP is based on takes these variances into account; my guess based on the differences we’ve seen between predicted and actual is that it does not and that it looks at date alone. If it currently does not, I’m wondering if it could be tweaked to do so.
I was wondering the same thing myself, and my thinking is that it may have something to do with the switch from a “percentage” based ranking system where there were equal numbers of 1s, 2s, 3s, etc. to a system where the CL numbers had more to do with the actual effect of the crowds on wait times.
Also, when there was all this wailing and gnashing of teeth about the changes, I saw very few people who tried to figure out what this actually meant. If you go to the park without a plan an increased CL will have a huge effect on your day, but if you use Optimized TPs then the total effect on wait times is pretty minimal.
I am wondering if maybe an automatic adjustment based on park hours (for those that miss those things) may help? I mean- if MK is open from 7:00am to 2:00am- I expect crowds (mind blowing crowds).
Well, you can’t change it, so I guess you have to accept it. Although I do recall @LaurelStewart having an “open period” where she would consider tag requests - perhaps she will do so again.
The wait time models do not take date alone into consideration, far from it. We have hundreds of factors, about 20 to 30 main ones depending on the attraction. As far as HOW a particular model predicts a certain value that is too complex to know or explain. A drawback from complex modelling is that you lose any human interpretation. Park hours interact with days of the week which interact with weather which interacts with school schedules, etc. etc.
If park hours are a predictor, why would the prediction not adjust when MK hours are scheduled from 8:00am-2:00 am (EMH- predicted 5 actual 10) or AK from 7:00am-8:00pm (EMH- predicted 6 actual 10). We know that Disney does not add hours like that unless they expect crowds of 9+, so could hours be weighted higher went they extend beyond a certain point (for example MK beyond 15 hours). I understand that at some point AK may routinely be open to 8:00, but currently that tells me it will be a 10.
Also, I asked in another thread. Liners are having a hard time walking from one attraction to another. It is throwing TPs off ( reports are that the plans are spot on it is just that fighting through the crowds the walk time cannot be adjusted slow enough and so at some point they fall behind and it throws the TP off). Any ideas on that one? It may just be a 9+ day issue but with these crowds lately it is throwing everything off.
In some cases, we would have already assumed the park hours would be adjusted, and based the predictions on that assumption. For example, we’re assuming that Disney is going to extend its Christmas hours. In other cases, as Fred says, there are other factors that come into play.
The models can definitely be improved, though. It may be possible to see how something like “MK beyond 15 hours” would work as a rule. If there’s anything else you can think of, let’s get a list together and see what happens.
The walking times thing is interesting. I’d be okay with slowing down the average walking time in the Optimizer. What I’d want to see is how often this happens - we tend not to hear from people whose plans went well.
True dat. However, I will say that the last time I used Optimized plans in parks the times were spooky accurate - pluses and minuses here and there, but overall almost right on.
Although I do recall