Airship Guide

[Aeyis]

To really answer the retrieval question, there are some tests we could do. The problem is isolating variables.

I think the problem is we have too many variables now, too much data.

Few, if any, people will send an airship with 100+ retrieval into SC01 or SC02 - there's simply no reason to do it. If we choose an item that's distributed among many sectors, like Birch Branches or Water Crystals, we run the risk of making invalid assumptions if the sector rating affects the retrieval amounts (like, if higher rated sectors require more retrieval for the same effect).

Likewise few send an airship into a higher sector with low retrieval. However, that's part of what all of this is about. Figuring out how it does work. As far as items across sectors, if you measure both sectors you can actually test those at the same time (whether sector effects things as well).

Regarding sector ratings and retrieval ratings, have you actually done that analysis? When I did, I found it rather inconclusive. But, it's a hard test to do, because higher retrieval ratings often also mean higher favor ratings, and that's what I believe actually affect the sector ratings.

Early on in the theorycrafting, yes.
Recently the only test I did was send an airship with super low retrieval and the rest of the stats high, to higher sectors (14,15,16 iirc) and it did come back with lowest retrieval, or normal for all items obtained. Abyssal or sub for all ratings. More tests would have to be done with that, but since my ship is still a month behind on discovering due to RNG I'm not in the position to do the testing myself unfortunately.
As for disapproving favor effecting it: I've always focused on range/surv/favor builds, and those rarely gave me high ratings. (A or S)

I think the most important thing to keep in mind in testing tho, is that everything seems to be effected by some RNG. It could be possible that even with super low retrieval you get optimal on a high level item...just at a super low/RNG chance.
So we have to keep in mind that we might have to ''ignore'' points of data that are too far out of the expected area.

However if we keep in mind they do not have to be the only contributors to an effect, it isnt actually that hard to test the various stats at this point. All thats pretty much needed is test both extremes of the stats, in various (high and low) sectors. IE: min surv and max surv in sector 1, sector 10, sector 20, etc. It would take some time, but thats all there is to it.


That said, once people get their airships all maxed out we will have more room for making such tests, so a better understanding will come with time, I think.