So…something to remember is that optimize is trying to minimize your WAITING IN LINE time. It is NOT trying to “make best use of your day”. Those are two very different (albeit related) goals.
If your plan has a lot of extra time for the day than the number of attractions you plan in for the day, it will try to place your times of waiting in lines seemingly randomly which minimizes your wait time. It does NOT consider the “free time” as “waiting”.
So, in your example, you can watch Reflections of China with no wait by goin there first. Sending you to Remy first would probably mean waiting LONGER IN LINE.
One way to help minimize this effect when using optimize is to ensure you have enough attractions in the plan to fill up the time you are planning to be in the parks. Or, alternatively, reduce the number of hours you will be in the parks (which comes to the same end). In this way, it will be forced to squeeze your rides and wait time into the otherwise “free time” slots.
It isn’t immediately intuitive. But if you just always remember that it is trying to minimize your time waiting in lines, it makes more sense.
I get your overall point, but I don’t think that completely explains what is happening here. Did you look at the plan? (hopefully I shared it correctly)
Just in the first two steps it shows a wait of 28 minutes for Reflections of China and then 59 minutes for Remy. It is including that initial 28 minutes in the total wait time. There’s no way reversing those two steps wouldn’t decrease the wait time for each (and obviously the total wait time).
The plan also shows two steps starting AFTER 11:00 (the end of EEH). So there are more than enough attractions to fill the time. I had even more originally but removed some it wasn’t getting to (including Test Track and Frozen) based on the message from our optimizer overlords that a “solution may be to select fewer attractions”.
I didn’t realize it was linked! I just took a look!
So…I noticed something important. MOST of the attractions, it has moved to the end of the day, when your wait times are very low. It was able to get through most of those with very little wait. Then, with the remaining attractions, it kind of picks the times you wait the least overall.
The algorithm doesn’t necessarily start at the beginning of the day and move forward…rather, it optimizes so that most of your waits are very low. It is actually doing that.
Remy will always have a longer wait, no matter what time of day, or it appears to be choosing a time within the big block of free time that gives you the lowest (or lower-est) wait. China…just kind of plugs that in.
I’m not saying it is perfect…I’m just saying that some of it makes sense.
But your SECOND plan (or, rather, the most recent one) looks completely messed up. It has you spending 1289 minutes in line??? That’s over 21 hours…more time than you have available in the parks! So something is DEFINITELY messed up there.
I think @len will need to have their team look at what is causing that to mess up so badly!!!
I don’t know! But I’ve had issues too - both with waiting time displaying in the 1000s and with long “waiting” periods (like 3-4 hours at a time) and 5 attractions left at the end of the day that we “didn’t have time for.” What worked for me was breaking our day up into chunks instead of asking the algorithm to figure out the entire day. I think I also switched to evaluate once I got a good sense of what the algorithm was suggesting - to me a minute or two longer of waiting for a ride order that I like better is worth it.
In this case, the first thing I’d do is start removing attractions until the optimizer thinks you can get to everything in one day. That will fix this issue.
I think in this case, the optimizer isn’t able to find a plan that does that. So everything looks equally bad, and the optimizer can’t figure out which way to go.
Let me know if you’d like me to adjust this plan on your behalf. Happy to do it.
Ok, I made a copy of the AK plan and started removing attractions (side note: when an attraction is removed, the remove button goes away for the other steps, forcing an optimize or evaluate step. Seems more difficult than necessary but I’m probably missing something.). Once I got it down to 15 and changed the FOP ILL time, it seemed to work mostly ok.
Wait! It looks like the park hours changed sometime after I created the plan today. It’s not updated at the top of the plan, but is updated elsewhere on the site. That explains why the hours started at 7:00 but the first step started after 8:30.
Another side note: should TPs ignore the very conservative hours official hours that Disney sets 70 days out? I feel like the estimated hours are ultimately more accurate. 9am-6pm for the week of Memorial Day seems very unlikely to stick.
Is the Optimize function still wonky? My plans for late July are suggesting odd selections for G+ as well as, I think , incorrect wait times. For example HS on 7-28 is suggesting the following at EE - MMRR , then ASwS (5 min wait) then SDD with 7 minute wait. I am pretty sure that won’t be accurate. ToT was suggested ad the first G+ Choice and then Star Tours. Just doesnt seem right so what am i doing wrong?
Similar results for Epcot the next day with Soarin as first G+ and then a later 83 minute wait for Remy.
Agreed, it does sound odd…although hard to know without seeing your plan, in case you have some gotchas in there impacting selections. Can you share the links to your plans?
I do believe the G+ recommendations part isn’t working well yet, based on what I keep reading, so I am not sure I would use those recommendations.
Thanks all for any help!
Other question - if optimize won’t suggest the proper G+ selections is there a way I can try to guess them and place them and evaluate the rest>
Just taking a quick glance. The times do seem unlikely…but the low wait time for SDD is likely because it is getting you to the line before the park opens officially. So, RD crowd might head there immediately, but toward the end of ETPE, the RD crowd will have ridden, so the wait time will definitely be lower than if you go there first.
I think it is work changing your walking speed to relaxed if you want to better account for delays in getting from one place to another.
So…is it way off? Maybe…but maybe not as much as it seems.
Having said that, always build into a plan breaks for eating, restrooms, etc…
Probably. But I am mostly against G+ on principle, so I just haven’t invested much time into figuring out how to best utilize it since I will go out of my way to avoid using it entirely.
thrill-data.com – if you find a time frame in the past with similar crowd levels, you can use it to see LL return times. this site has a ton of different ways to look at all kinds of data. I like looking at the ride-specific 8-day history of LL returns (click the tiny button underneath the LL return times for your date). the URL naming convention for TD is standardized so it’s easy to jump from attraction to attraction if you know the full name of the ride - just swap out the attraction name in the URL.
Thank you! I have time to dig into this! We have a group going that includes a 2 and 4 year old as well as 88 year old great grandma, so I am feeling the pressure to give everyone a trip they will love.
Another question - When looking at similar crowd levels, does the day of the week play into this at all? Or would a CL 6 on a Sunday in July be equivalent to a CL 6 on a Thursday?
I plan on plugging this into my touring plan to evaluate then rather than optimize since the G+ choices via optimize are not too useful based on what i am used to seeing - is this an okay strategy?
