Park Touring and the Genie+ & ILL option

I’m only using evaluate.

1 Like

But yeah, good point, it shouldn’t happen - it might have been from previous optimizations.

1 Like

Oh, I didn’t realize the obtain step was a physical step that had to be in alignment with the other times. Ok - let me try that.

2 Likes

That would make more sense.

1 Like

I put the obtain LL step right after JC, but it is giving me late morning return times - which is not what I expected based on thrill data for 1/16 (I will be there 2/5). At 9:15 the return time for POC is 9:20 not 11:30?

The return time for BTM was 1:30 and even on thrill data it was 10:30 - neither of those times will work for me. I’ll try to “rope drop” BTM

At least now I know what needs to be done, so that’s progress :smile:

2 Likes

Oof. Yeah that return time is not realistic. Not even on CL10 days. Len, the predicted LL return times are still too far away than what the data says. This still needs more work yet:

3 Likes

Do you think the fact that Splash will be closed has anything to do with it…?

2 Likes

No.

2 Likes

I did put in a ticket with this question/issue with my plans as well. I noticed the same thing regarding the inaccurate return times. I did not use Thrill Data as the comparison - just Becky’s blog post, but still, my results did not seem accurate for planning.

I got this response:
“The Genie+ return time predictions that our Personalized Touring Plan software is using are date-specific and based on a statistical analysis performed last month. We regularly analyze new data and update our forecasts, so I do not expect them to be the same as whatever was used in an October blog post. It’s also not the case that we assume the same return time forecasts for every crowd level 6 day – dozens of factors are taken into account in the date- and attraction-specific forecasts. If you optimize this plan again in a few weeks, you might again see different results.”

5 Likes

Yeah this can be a small problem foe longer rides. Like if I have an 8:30 TOT I would book the next LL before going in the hotel. Which would produce different times than 15 mins later when I am off the ride.

4 Likes

Yeah that’s another big flaw that probably can’t be fixed.

3 Likes

Thanks for sharing, but tbh that’s not a sufficient answer from TP when you look into the data.

This is the past month of data. The very latest RT was 11:00 at 9:20 on a CL9 day:

So this 11:35 RT would be almost impossible:

Also, if the TP predicted RT’s are all off in the same direction (over estimating), that’s a red flag.

7 Likes

Yeah- I figure it is pretty impossible!

4 Likes

I agree with that. I emailed them initially because of this issue. On a CL 6 day, obtaining MFSR at 9:15am gives me a return time of 6pm. I’ve been watching HS all week and that is very off from what I’ve seen for actuals.

4 Likes

I haven’t figured out yet how to get value out of the LL forecasting capability in the TP tool.

3 Likes

This is great feedback. I’m going to bring Fred, our statistician, into the convo. Thank you.

3 Likes

Great, thank you. We understand how difficult it is to get the G+ predictions,etc. mastered, but also hate to see frustrated customers of yours potentially giving up on Touring Plans all together.

6 Likes

There’s no doubt that some of the forecasts for Return Time will prove to be inaccurate. Accuracy improves once we have at least one full year of “typical” data. Just to be clear, 2022 was not typical.

Looking at specific examples like this is very helpful. There may be attractions that could use adjustments to the forecasting models to make them more accurate in the short term.

5 Likes

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?

3 Likes

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.

2 Likes