Crowd Calendar Review for February 2019

Share This!

We’re more than halfway through the month of February, so let’s take a glance at the Walt Disney World crowds so far this month:

Park Predicted Crowd Level Actual Crowd Level
Magic Kingdom 5.2 5.7 (+0.5)
Epcot 4.1 4.8 (+0.7)
Hollywood Studios 4.0 6.8 (+2.8)
Animal Kingdom 4.6 8.3 (+3.7)

Predictions for Magic Kingdom and Epcot did really well. If you visited those parks in February, you experienced a crowd level within one index point of what the Crowd Calendar predicted on most days.

Hollywood Studios and Animal Kingdom are a different story. In particular, Animal Kingdom’s posted wait times indicate crowd levels nearly double what was expected. Does this mean guests flocked to that park in record numbers during February? We don’t think so, based on three things we’re seeing:

  1. The rides seem to be operating near 100% capacity
  2. The difference between actual time spent in line and posted wait time is larger than it has been in the past
  3. The posted wait times seem to change faster than is physically possible – they’re more “volatile” – than at any time in the past

Ride Capacity

We don’t think the issue is with ride capacity. We spent several days counting the number of guests exiting the rides per hour at Animal Kingdom. For every ride, the number of guests we counted was within a few percent of each ride’s maximum hourly capacity. So we don’t think the issues we’re seeing are due to staff or ride capacity cutbacks.

Actual vs. Posted Wait Time Differences

Recall that the Crowd Calendar represents the average posted time between 11 a.m. and 5 p.m. at a park’s core set of attractions. Usually, this is a good proxy for the number of people in the park. Even when it doesn’t directly correlate to attendance it is still a useful metric because it represents how long (Disney says) a guest has to wait in line. This month at Animal Kingdom we are seeing patterns in posted wait times that we’ve never seen.

An example of this is DINOSAUR on February 10, 2019:

DINOSAUR February 10, 2019

There is a lot of information of this chart, so let’s break it down:

  • The blue-green area with the green border represents the set of posted times observed on that day.
  • The red line represents what the Crowd Calendar predicted and the dotted line represents an average day.
  • The circular markers represent an actual time recorded by a guest in line.

You can see a similar representation of this data on each of the Crowd Calendar Daily pages like this one for February 10, 2019.

Posted wait time curves like this are unusual. DINOSAUR saw six spikes in posted time throughout the day. The largest spike was during the 1 p.m. hour, where the posted time went from 25 minutes to 90 minutes and then immediately back down to 30. Calculating the observed crowd level on a day like this is problematic because these spikes push the average wait time higher than it probably should be. We predicted an average of 30 minutes for DINOSAUR on February 10, which represents a crowd level of ‘4’ for that attraction. But we observed an average posted time of 41 minutes which represents a crowd level of ‘9’. Between 2:30 p.m. and 3:30 p.m. posted times spiked up from 30 minutes to 60 minutes, then back down to 20; however if you lined up at 2:50 p.m. you only actually waited 30 minutes. (We don’t think this is physically possible, so it’s an example of both problems we’re seeing.)

Here is Kilimanjaro Safaris on February 15, 2019.

Kilimanjaro Safaris February 15, 2019

We collected several actual wait times for Kilimanjaro Safaris throughout the day on February 15, and they all match the predicted curve relatively closely. However, the posted times are higher. This is a good example of what seems to be happening at Animal Kingdom in February. We see observed crowd levels at the upper end of our scale; however, those levels are not always reflected in the actual time guests are waiting.

The wait times at Kilimanjaro Safaris were also unusual on February 12:

Kilimanjaro Safaris February 12, 2019

These posted times seem to be legit, but the long waits early in the day are odd. This could be an indication of the attraction operating at a reduced capacity in the morning hours. A 64-minute wait for Safaris at 10 a.m. is unusual – especially for a Tuesday in early February. The capacity seemed to increase in the afternoon, but the wait times never fully subsided until the park approached the closing hour.

Here’s the chart for DINOSAUR on February 7.

DINOSAUR February 7, 2019

We see drastic spikes for DINOSAUR’s posted times on this day, as well. Notice how those spikes represent wait times much higher than an average day. This is surprising considering the other three parks showed normal crowds on February 7 (MK 5, EP 3, HS 4).

DINOSAUR February 6, 2019

We see a similar pattern the day before, on February 6. Neither one of these days has a recorded actual time during the busiest part of day (can’t say I blame our community: who wants to wait 50 minutes for DINOSAUR?!), so we don’t fully know what the wait times were like, but it seems likely that those posted times are inflated. The few actuals we do have show wait times in line with predicted levels and pretty much in line with the average. Let’s look at some other Animal Kingdom attractions.

Expedition Everest saw some strange wait time patterns in the last few weeks. Here is Everest on February 10, 2019.

Expedition Everest February 10, 2019

Wait times throughout the first half of the day were in line with predicted levels, but then at around 3 p.m. the posted time climbed steadily, peaking at 140 minutes! These are wait times we normally see on the busiest days of the year, not on a Sunday in early February. Expedition Everest can have strange wait time patterns because it has a high frequency of temporary breakdowns, but we didn’t see any of those on February 10. It is difficult to find a reason why the wait time would spike so high at the end of the day like this. Did it switch from operating five trains to only two? Clearly, this spike cannot be due to an influx of guests after 4 p.m. – the curve is too steep. Something else is happening here.

Wait Time Volatility

The third unusual thing we’re seeing is large swings in posted wait times, especially large drops within a short time. The chart below shows DINOSAUR’s posted wait times from February 14 as an example:

Dinosaur February 14, 2019

At 1:30 p.m., DINOSAUR’s posted wait time was 90. Fifteen minutes later, it was 20: Person “A” got in line at 1:30 p.m. and got on the ride at 3 p.m. Person “B” got in line at 1:45 p.m. and got on the ride at 2:05 p.m.

It’s not possible for both of those wait times to be true. It means Person B got in line later than Person A, and yet somehow got to ride first – i.e., Person B passed Person A in line. (Remember, we’re pretty sure the the ride is operating at full capacity, because we’ve counted the number of people leaving the ride.) So one of the posted wait times is wrong.

We see a similar drop starting at 4:24 p.m.: the posted wait goes from 70 to 25 in the span of 6 minutes.

We also see large increases in posted wait times, far beyond what can be explained by normal traffic patterns. For example, the posted wait went from 20 minutes at 2:10 p.m. to 65 seven minutes later. For that to happen, around 1,300 people would have had to walk to DINOSAUR in 7 minutes.

It’s highly unlikely that 1,300 people walked to DINOSAUR in 7 minutes at 2:10 p.m. To put that in perspective, the number of people who line up for Seven Dwarfs Mine Train at park opening was around half that the last time we counted. DINOSAUR isn’t as popular as Mine Train, especially not at 2 p.m. Also, a line of 1,300 people walking to DINOSAUR would be somewhere between 0.3 and 0.6 miles long.

We’ve collected wait times continuously for the last 78 months (since August 2012). If you count the number of these unusual wait time changes across all Walt Disney World attractions, the months of January 2019 and December 2019 rank as the #7 and #2 worst out of those 78. January 2019 was the worst January we’ve ever seen, and December 2018 was the worst December we’ve ever seen. (November 2018 was the worst November we’ve ever seen.)

Effects on the Crowd Calendar

When we observe wait time curves like these, it really messes with our calculation of “How did the Crowd Calendar do?”. Despite seeing crowd levels of ‘3’, ‘4’ and ‘5’ at other parks, Animal Kingdom shows ‘9’s and ’10’s. But are they really ‘9’s and ’10’s? The curve for DINOSAUR on February 10 represented a crowd level ‘9’ instead of the ‘4’ we predicted, but was it really a ‘9’? It is fair to say that Animal Kingdom wait times have been higher than predicted so far in February, but it is also fair to say that those higher wait times don’t seem to represent the entire truth either. These wait time curves are so different from what we normally see, it is difficult to compare with days in the past.

We took at look at the volatility of these posted wait time curves (the up and down-ness) over the years and found that Animal Kingdom attractions are more volatile in 2019 than they have ever been (our database goes as far back as 2006). Why that is, is anybody’s guess. But, it is something we will have to deal with if it continues.

Crowd Calendar Updates in March

When we update the Crowd Calendar at the beginning of March we may see some adjustments for Animal Kingdom based on these patterns. Magic Kingdom and Epcot seem stable so we don’t expect any major changes there, just the usual tweaking. At Hollywood Studios we will see some significant changes. Tower of Terror is going to operate at reduced capacity throughout Spring and Summer. This has an impact on the wait times and therefore also impacts the Crowd Calendar. In addition, we are adding the new Toy Story Land attractions to the list of core rides at Hollywood Studios that we use to determine the crowd levels. This will help stabilize the numbers there.

Got any questions for the number gurus? Theories about the odd spikes at Animal Kingdom? Let us know in the comments.

Fred Hazelton

Fred Hazelton maintains the crowd calendar, theme park wait time models and does hotel rate analysis for the Unofficial Guides. He's also done the models for the new mobile wait times product Lines. Fred Hazelton is a professional statistician living in Ontario, Canada. His email address is fred@touringplans.com. You can also follow him on Twitter: @DisneyStatsWhiz.

25 thoughts on “Crowd Calendar Review for February 2019

  • February 20, 2019 at 11:59 am
    Permalink

    Thanks for sharing this!

    One theory I have for this are the large groups that are at WDW in early February for dance and cheerleading competitions. I visited February 2-7 and arrived as the dance competition was wrapping up and leaving as the cheerleading competition was just beginning. Particularly on February 6 and 7, I saw a ton of large groups touring the parks together. I’m wondering if several large groups enter the queue for an attraction at the same time, perhaps that would cause the wait time to spike?

    I’m also wondering if this is an impact of Disney limiting the operating hours for some of the attractions at DHS and DAK. For example, we know the hours of Kilimanjaro Safaris, Dinosaur, and Majarajah Jungle Trek have all been reduced. At DHS, One Man’s Dream and Muppetvision have been reduced. Perhaps these reductions explain some of the overall increases, particularly those in the mornings or evenings.

    Reply
    • February 20, 2019 at 1:54 pm
      Permalink

      I don’t think Cheerleading groups explain it. To generate true wait time spikes like we’ve seen they would have to number in the several thousands. Also, they’ve attended around the same time for many years and we didn’t see this previously. The shortened park hours are an interesting idea. It’s possible.

      Reply
  • February 20, 2019 at 12:13 pm
    Permalink

    Very interesting stuff. I would be curious to see more actual wait times at each attraction in AK and how well they correlate with posted wait times. Also how closely are the drastic fluctuations in posted wait times between attractions and if Disney is using the wait times as a means of crowd control and movement within the park. We already know they do this, but we know that Disneyland is working on Project Stardust, and perhaps AK has become a testing ground for it.

    Reply
  • February 20, 2019 at 12:30 pm
    Permalink

    I was in AK on the 12th and there was a while in the morning (before the rain came) when both Dinosaur and Everest were down, which could account for longer waits at Kilimanjaro. Interestingly, the app showed Dinosaur open with a 15-20 minute wait, and they were letting people into the ride, but telling them on the way in that the ride wasn’t running. This was true for at least 30 minutes. Afterwards, we walked in standby directly to the Fastpass merge point and the pre-ride movie, even though the posted wait was still 15-20 minutes.

    Reply
  • February 20, 2019 at 12:46 pm
    Permalink

    Thanks again, such interesting stuff. Wondering: do any of those jumps or drops in wait times correlate with the emptying/filling of show buildings after/before shows (Esp Nemo>Dino or FoLK>KS but another example maybe B&tB>ToT/RnR, or Indy/JTA>ST)? Not to say they are real, just wonder if CMs are inflating posted waits at those times of day. I am, though, surprised to see so few Actuals posted for these attractions.

    Reply
  • February 20, 2019 at 12:58 pm
    Permalink

    Fascinating stuff. Looks like we need more actual wait data points… if only I didn’t feel the need to nurse my battery.

    Reply
    • February 21, 2019 at 1:05 pm
      Permalink

      We certainly do need more actual times. They are very powerful in the data analysis.

      Reply
  • February 20, 2019 at 2:02 pm
    Permalink

    Okay this is a random thought. I believe Len said on Twitter that wait times are most often updated by cast members, but that Disney is experimenting with automated times.

    What if the computers are changing wait times based on projections or algorithms or whatever…and then someone comes through the line with a red tag and the CMs adjust back to ‘normal’?

    Reply
    • February 21, 2019 at 1:07 pm
      Permalink

      That’s a new one to me Rick. That could explain it. It is kind of a rule in statistics that when you see a spike in a time series, something human is intervening. “Nature” tends to look smooth on a data plot. Clearly, humans are involved with these significant changes in posted time.

      Reply
  • February 20, 2019 at 2:55 pm
    Permalink

    Call me a conspiracy theorist, but could Disney just be jacking up numbers throughout the day in order to manipulate crowds? In other words if they post 100 minutes when it’s really only 20, fewer people will get in the line (and theoretically they’ll be satisfied because they ‘avoided’ a 100 minute wait), while the people who actually do get in the line will also be satisfied because they were expecting a 100 minute wait and got 20…?? Disney is obviously looking a lot at crowds and trying different things, including tiered ticket prices, etc., so why not mess with wait times, too?

    Reply
    • February 20, 2019 at 3:36 pm
      Permalink

      This was my initial thought as well.

      Reply
    • February 21, 2019 at 10:08 am
      Permalink

      I think they do, FWIW. On our last trip we were just a little late for rope drop at EPCOT. When we arrived at Soarin’ the posted wait time was 50 minutes. I read it out loud and the CM leaned in and whispered “Its only 20”. And she was right. I know there is an element of “under-promise/ over-deliver” in all the posted times, but 30 minutes seems pretty extreme. Even my 10 year-old said “they probably are trying make people go to other rides.”

      Reply
    • February 21, 2019 at 1:09 pm
      Permalink

      This is a theory we have considered as well. Along this line of thinking, Disney may be experimenting with guests reaction to very long waits in preparation for Star Wars Galaxy’s Edge which we believe will be the most popular Disney World expansion in a ling time. It’s just a theory, we don’t know for sure.

      Reply
  • February 20, 2019 at 4:46 pm
    Permalink

    The counting system on Disney’s part seems grossly inaccurate at HS an AK. I imagine the correction from spikes is from an actual guest going through with a lanyard.

    Reply
  • February 20, 2019 at 8:19 pm
    Permalink

    Could it be that all the people in the Nemo show all walked over to Dinosaur? Or everyone from FOTLK went to Safari? Or maybe there was a problem that let too many people pick a FP+ time?

    Reply
    • February 21, 2019 at 1:11 pm
      Permalink

      The flow of guests out of Finding Nemo The Musical can’t explain these jumps because we have not seen in it previous years.

      Reply
  • February 21, 2019 at 10:28 am
    Permalink

    I always love your statistics posts, but this one is probably my favorite. Thanks for the update! Are you planning to do more field research to see what’s going on with the Animal Kingdom wait times?

    Reply
    • February 21, 2019 at 1:13 pm
      Permalink

      I enjoyed researching this one too, Ryan. Glad, you enjoyed it. Yes, we will be tracking this trend and reporting on it in the coming months.

      Reply
  • February 21, 2019 at 5:55 pm
    Permalink

    Thank you for analyzing this. As my trip draws even closer, I’m interested to see how this trend may change…or if you’ll somehow learn Disney’s inflation motives. I’ll do my best to remember to post my times on the apps when I’m there in early March!

    Reply
  • February 23, 2019 at 7:44 am
    Permalink

    Love this deep drill-down and up close look at this kind of data. So nerdy! 🙂

    I particularly like that this seems to demonstrate TP as valid and accurate much of the time. I’m a longtime member and proponent of the product so this is satisfying to me, particularly in light of all of the recent debate over whether it’s still working. As the predicteds seem to fall closer to the actuals and averages for the day in the situations presented above, it seems to me it still is 😉

    Keep up the great work!

    Reply
  • February 24, 2019 at 4:04 pm
    Permalink

    I was at Disney last week. MK on Presidents Day (was absolute mayhem), HS on 2/20 and AK on 2/22. I would say MK and AK were the busiest days. Frankly, I think AK appeals to a more mainstream audience in terms of ages. I’m not surprised it is busy. Looking at all 4 parks objectively. I think MK and AK are head and shoulders above HS and Epcot at the moment. I don’t have the historical perspective that TouringPlans has but looking at what the parks offer from year to year might show that if you can only 2 or 3 parks during your stay MK and AK are must sees.

    All that being said Touring Plans is a huge help. It saved our vacation even with the parks being busy, having a ‘touring’ strategy is vital to success. I saw a ton of people using the Disney app but that is only part of the solution if you don’t know the order to meet your goals, the disney app is useless.

    Reply
  • February 24, 2019 at 4:59 pm
    Permalink

    Do you think the announcement of the new Harry Potter coaster opening at Islands of Adventure in June will impact crowds at Disney this summer? We were hoping the crowds wouldn’t be as bad with people holding out for Star Wars in the fall, but now with the free dining and Universal attracting more people to Orlando, looks like that probably won’t be the case.

    Reply
  • February 27, 2019 at 9:12 pm
    Permalink

    We were at AK on 02/11, and went to the EMH since we were on property. I was on MDE on the bus, and noticed that there was a wait time listed for FOP (this was the only attraction with one) at 60 minutes and this was about 15-20 minutes before the park opened for EMH (I was still on the bus, so they may have let people get in line early, but not sure). I figured this was an automatic thing since there is probably a minimum of 60 minutes for that ride anytime, and I’m sure people rush to that ride first on any day or event.

    Or it could have been a glitch, lol.

    Thought that was interesting.

    Reply
  • March 3, 2019 at 9:18 pm
    Permalink

    I know this post is a bit old but I wanted to chime in here with my experiences as a former cast member at WDW. For context I worked The Disney-MGM Studios before its present day DHS incarnation, at several attractions:
    Light, Motors, Action Stunt show
    Backlot Tour (specifically the first part, the Water Tank)
    Toy Story Mania

    By the time I left I had worked both as a trainer and a lead at times.

    First to address the comments about Disney intentionally misrepresenting the wait times. I will say that at least at the time I worked there and the attractions I worked, we never ever did this, and I never heard anyone talking about doing so.

    That said, what does happen is a few unscientific things:
    1) Experienced cast members after a while can roughly tell a wait time by where the end of the queue is, and they will just post that. This can be INCREDIBLY accurate if done by an exprienced hand, but incredible INACCURATE if done by someone how thinks they are experience and really aren’t. Often its the latter.
    2) Wait times are often just bumped at certain attractions in anticipation of show dumps. You just know, roughly, how a show dump affects your attraction and change the wait time accordingly.

    However I will also say that unless things have changed, Disney really doesn’t have a very good idea of what wait times actually are. The only tool we had for measuring wait times was to send a lanyard through. I’m sure most everyone here is familiar with them. The greeter hands a guest entering the line a red card on a lanyard and asks them to hand it to the cast member when they get on the attraction.

    The lanyard system is frankly hugely flawed and mostly useless for various reasons:
    1) Often times the guest forgets to hand in the lanyard
    2) Some guests like to “mess with Disney” and pass the lanyard off to other people, ruining the measurement
    3) The system requires a known start and stop time, but often one side or the other (greet or loader) loses track of things or is too busy to note times immediately, thus they are often incorrect

    Often time unusual spikes in wait time happen because a lanyard goes through then is forgotten or in some cases one goes through and is lost then another goes through and is recorded as if it was the first one. So the timing is doubled what it should be (roughly).

    Also lanyards are supposed to go through regularly. If I recall correclty (and again, MGM, so its been awhile!) they were supposed to go through every hour. In reality they go through far less than that and someone just guesses instead.

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *