Have you guys had your people out in the parks testing the new software since the launch last month? If so, what do the results of the G+ return time forecasting/testing look like?
I’m not sure if this helps, but I was running into similar issues with making my plans.
I just swapped to evaluate like you said and made it up myself for the times I expected to get and entered them as DAS times to have a better idea of expected waits and how to order things.
Yes. We run daily error reports on all of our forecasts (which are not easily sharable here). We see that our forecasts for G+ return times seem to be either good or way off. Which tells me that there’s a lot of inconsistency in the data. I think the approach expressed in this thread is a good one. Use the forecasts to add G+ selections whenever possible but know that you might need to make adjustments for situations that look like outliers.
I would also add that our in-park testing has shown that in practice, our forecasts for standby and G+ return times are actually pretty good in most cases. Which is to say that when you use them they can work really well together.
To this point, I noticed a very different pattern of return times for the first half of last year compared to the last half of the year. Even accounting for busy vs. slow. LL lasted much longer and stayed earlier, longer the 2nd half of the year.
At a very specific point I think Disney must of increased the number of LL per ride. I am sure that skews the predicted return times later for now, until you have more data. The next five months should make a big differnece as the switch seems to have been sometime between May and July.
That came to my mind too. Shorter term LL data (3-6 months) may actually be more beneficial than long term data.
This. There was a sea change in LL availability 2nd half of year, to the point that I would disregard 1st half of year entirely. Analysis here:
@fred my sense is that recent LL patterns are the most valuable from a prediction standpoint. There have been so many changes in G+: LL availability, ride lineup, rules, user knowledge, dynamic pricing, park hours. My go-to for LL return time prediction is to look at recent thrilldata for a day with similar CL, day-of-week, price, and hours.
I’m not sure if they added more availability or if removing G+ purchases from packages resulted in fewer people buying it daily.
I second this Fred.
Oooh, good point.
It makes sense.
I haven’t checked, but since the data used is 2022, makes sense the wait times seem so off. The TP predicted times are similar to the wait times observed on same day last year in Trill data.
I hope things go back to normal this year. I am not expecting CL10 in freakin February.
I just re-evaluated and it changed, but not necessarily for the better. I dont think you can enter Genie right now when you optimize.
It changed when I could book my next one (instead of right after I ride) and the return time for the LL changed by 5 mins. Still not accurate based off of the past two weeks with similar CLs.
I’m beginning to think that the cost of Genie impacts the LL return times more than the crowd levels.
Just throwing a note out there to anyone who has very recently used a Touring Plan with G+ and how it turned out, accuracy wise. Please share your experience here. Thanks
Can I clarify- you are asking for people that used a touring plans in the park with genie+ and optimized their day?
Anything. Whether optimizing in parks or not. Any TP predicted RT and the outcome.
I asked because a written TP is a prediction and an optimized TP is one that responds to the current conditions (ride breakdowns). I think the first is good for general planning. The second is what I want to work - in park.
The main desire is to be able to create an accurate plan and stick with it. Meaning, not having to optimize the day of so other plans in your TP won’t get turned upside down. It used to work, with pretty good accuracy.
Right. And optimizing here can be good or bad depending on how future plans are affected. (Especially if your plan was never optimized from the start out of preference)
I guess I never looked at a preprepared TP as anything other than a rough estimate. Interesting. I do know that Len suggested optimizing after your first attraction so that your plan would adjust.