These are great! Good peek into their processes. Hopefully they are using this data to improve their forecasting in 2023.
Just realized that the chart in the first article doesn’t jive with the stats in the second one. I thought TP only hit within 1CL 58% of the time, but in the Disney comparison, the pie chart says 75%.
So, the 20th-27th of January was a rough time across the board last year.The deep dive article gives the explanation of MK being closed early one day as a reason for the big miss there, but sort of ignores that week at the other parks even though it was consistently underpredicted at all of them. The thing is, that time was also very underpredicted (most of our days were underpredicted by 2+ CLs… we had an actual CL of 10 at both parks on AK/EP day 
) in 2020, though at the time we blamed that on the ProBowl being in town, even though they’d already revised the CLs up by 2 or so after that was announced. I’m starting to wonder if there isn’t just some late January travel bug.
It doesn’t appear to be that much more than normal on average. Since 2018, late January is constantly 6/7 CL. Before 2018, it was low to medium.
First article was resort crowd level as a whole. Second one is the aggregate of each individual park-day. It’s easier to get within 1 crowd level for the resort as a whole than it is to do it for each park 
Oh interesting! So that could mean a lot of the misses were due to crowds going to a different park than predicted, basically. Like the example of avoiding MK on party days.
Right, but if you look at the color-coded calendars for 2022 that show how well they did in that second article, you can see that each park had at least 4 days of orange/underprediction in the last 9 days of January 2022, and 3 parks had 2 or more days of 2+ CL underprediction in that time. In 2020 the observed CLs were like 7-10 at every park, almost every day that week which was 2+ higher than predicted pretty much across the board. Just looking at the observed vs. predicted numbers from the end of January 2018 shows that they were underpredicted then too, and by nearly the same magnitude.
It seems like they’ve upped the CLs since 2018, just maybe not enough.
One thing worth calling out is that Easter doesn’t seem to affect CL much, compared to the week immediately before or after. I looked at 2023 predictions, and actual for 2022 and 2019.
Can you say more about CL’s and Easter? We’re going Easter weekend and week after this year.
From the article:
Easter is April 9 this year and the calendar is predicting the same pattern that we have seen in recent (pre-pandemic) memory – busy crowds leading up to Easter weekend but moderate crowd levels on the actual day. No major changes to report during Easter as a result of this update. It seems strange but yes, the calendar is suggesting that it is better to visit on Easter Sunday than the week before.
The 2023 crowd calendar is here. If you’re not a subscriber you can still see some dates in March and early April. Looks like a rolling window so if you wait a couple weeks you may be able to see Easter week and week after.
Historical crowd calendar for Easter period (subscription required):
April 2022
April 2019
I’m glad to see that Fred did a complete review of the wait time models.
But I’m still disappointed that he just left the conversation when we were discussing LL RT protections…
Was very confused about how the update link sent to my email was talking about omicron affecting crowds. Finally realized they had sent me the link for 2022, lol!
Most people are off the week before Easter.
I’m not a current subscriber to TP, so I am unable to see the actual crowd calendar projections for last week. But we just returned from being in MK and Epcot on the 14th and 16th. Anecdotally I would say the parks were packed. Epcot in particular was wall-to-wall people. The crowd levels were not what I was expecting. It reminded me of what it was like the 1st time our family went to WDW at Christmas time many years ago. I did check other crowd calendar sources on line before booking this trip and all seemed to indicate last week would be in the “moderate” level. I realize everyone’s perception is different, and there could very well have been something else going on at WDW I was unaware of. Unfortunately we’re kinda sorry we went, it was that crowded to us. I jokingly told my wife that I wonder if today’s moderate level is yesterday’s “crowded” level. Next time I guess we need to re-subscribe to TP’s beforehand. 
2/14/23 TP predicted CL3 for MK, actual was CL7.
2/16/23 TP predicted CL7 for EP, actual was CL5.
Thanks for that info. Comparing those numbers to what we actually experienced is a real reset of expectations for us, which your historical upward trend in data already seems to verify.