Mistweaver - lastest updated - Difficulty settings?

Hello everyone !

Short version of this post :
Does the upgrade finder bonus roll change depending on current gear or BiB gear ?
What is the current difficulty setup AMR gearing is targetting for an healer ? is it constant ? is it depending on iLvl (i supposed the second one) ?

Long version:
I dont have time to do a in depth work about what i’ve enconter for the last update.
So i was a bit surprised about the missing “Temple of Sethraliss” in the Bonus roll upgrade finder section.
The trinket was a very good upgrade from AMR’s point of view and mine since then.
So here is the past gear i had :
Here is the BiB from AMR :
(So loss in HPS but gain in less death in the fight; so an upgrade from my point of view no problomo here)

Here is with the trinket “Fangs of Intertwined Essence” in 400 'so MM+10 without procs

Here is my current gear with the trinket

=> improvement from BiB in terme of dead and HPS
=> so does the upgrade finder scale on current gear or on BiB gear ? That may explain why the trinket isn’t flag as upgrade in one case.
=> anyway, this trinket seems to be an upgrade for both gear setup (AMR and the previous one) so may it be the difficulty a bit low ? (i’ve put +7 (+47%) on each of this simulation
=> maybe the difficulty isn’t hight enough to offer thinket to shine ?
=> question for the community; do any of healers have somethings similar in terme of gear suggestion/difficulty?

héhéhé i found another things :
Comparaison between two polearme (the one that AMR suggest to me before; an intellect one; i have nothings more than a 385 in agility so that doesn’t bother me to get an 425 intellect to replace it).
And the polearme on Blockade Mythic.
Improvement on death chance and TUF stats.
Maybe it’s just Upgrade Finder that need a bit of tweaks :thinking:

And last thing, my string if you want to redo this :


I see the fangs of intertwined essence estimated at -0.06% in the upgrade finder. These simulations show that it might give a slight HPS increase.

It is very difficult to rank trinkets with special effects across all item levels in such a way that they will always match any spot simulation you do. Especially with healing simulations, which take by far the longest to do, we need to allow for a little extra “fuzz” or margin of error in order to generate enough data to find the trends.

The Upgrade Finder compares each item to the BiB setup. It swaps in that one item for an item in the BiB setup. So, it is definitely possible that the Upgrade Finder will estimate an item as a downgrade, but if you were to obtain that item, it might end up being BiB. We don’t do a full-on BiB for every single item in the Upgrade Finder - that ends up being too expensive. It would also be confusing because some people might not expect to swap around other gear too in order for an item to be an upgrade.

We are continuing to work on improving the gear rankings. Tanks and Healers are definitely a little looser than DPS as far as exactly matching the simulations. But, we’d also argue that Tank and Healing simulations are a bit of a looser model of the game as well. What we are looking for are general trends that can lead us to gear sets that are good.

I do see the Upgrade Finder recommending the Docksplitter Siege Hook as an upgrade for your tank set. I think in your simulation for that one you forgot to put a weapon enchant on it. I ran the simulation again with a Versatility enchant and got this:

So it looks like that case might be working pretty well in the Upgrade Finder.

But, overall, the Upgrade Finder is only going to be able to give an estimate - we think it is a pretty good estimate, but there is definitely going to be some error. I don’t think it is off so far that it would lead you to farm or not farm significant gear upgrades/downgrades, which I think is the most important measure of its usefulness.

There might be a chance that mana-saving trinkets are giving the optimizer some trouble, I could look into that. Those are particularly hard to rank.

1 Like

Thanks for the work you’ve put into the response and in AMR web site =D
Well i’ve nothings more to say, I agree all the way.
I didn’t expect an enchant on it would up that mutch ! but hey, you are working with this everyday so you know Mr Robot very well now :wink:

For the mana trinket, i dont understand the “across all item levels” like having a 370 fang and a 375 and a 400 and 405 and 415 and a 425 etc… ? I guess it make a lots of data…
Maybe some trinket may not have “short cuts” to have get an estimated value for it.

I personnally like this trinket more than a passif one, i feel like when pops at the end of the fight you can heal for 10s more and could help get a kill where mana is lacqing.

Thank for the repply ! And good luck !

Yeah, we can’t run data for every item level of a trinket. We have to create data at a couple of item levels and then generate a function to estimate it for any version that you might have. In general it works well, but you can see how error creeps in.

1 Like