Same items but different math according to what I have in bags, which one do I trust?

Hello, I have a strange situation.

Using my current gearset, I do /amr show, go best in bags, and it tells me to switch 2 items and some gems for 0.39%

then, I go buy a conquest neck, with that in bags I go /amr show again, copy the new string, paste in website and it gives me same exact items, but different gems and 0.91%

I’m kind of confused, please help, I will attach a google drive video demonstrating my question here

https://drive.google.com/file/d/1BwkfNLtvd7eh6-0t4yn5eL6AmVElzxOV/view?usp=sharing

Thanks!

The best way to get help is with a snapshot, since that gives us all your settings, etc. Instructions on how to create and post a snapshot here:

Fine, here is snapshot with 0.39
aa583f4c7e064eed80fb11a811c24b5a

and here is snapshot with the exact same items but I bought a necklace (which it doesnt suggest equipping) for gear 0.91
d1afde83758a4e8cbdc9d65443a003ca

You can clearly see from the video which one is which.

Thanks for the snapshots, we’ll take a look.

Videos and screenshots aren’t that helpful for debugging issues – we need the snapshots so that we can load your exact settings and inventory and replicate your case on our test servers.

So in this case… there are a lot of gem combinations, especially once you add in meta gem requirements and unique gem constraints. It’s simply not practical to test them all in combination with all of the gear combinations available as well.

When you get different gear, it is trying slightly different combinations in different orders, and thus arrives at a slightly different solution. This can sometimes happen as well if you e.g. lock in a gem or two – it will arrive at a slightly different solution from time to time.

This case is a little larger than we usually see, but still not that big of a difference: 0.52%. In the overall scheme of things that is insignificant, but we’ll see if we can narrow it down a bit more in this case.

Thanks yellow,

Although, one could argue 0.52% is very important when trying to parse higher :smiley:

Hope you can narrow it down soon

No model or simulation of the game is accurate enough to really know if one set of gear is 0.5% better than another. The variance from attempt to attempt is also very high. You would have to do the exact same fight in the exact same way (length, movement, etc) hundreds (probably thousands) of times to know for sure which set was better in-game. That is why we say anything under 1% is good for an optimizer. In-game, you can’t really “feel” the difference until you get in the neighborhood of a 5% dps difference.

What i commented was said as a joke. You make a good point but still, i like to see numbers :smiley:
Cheers