In addition to the new customization features (described more in this post), this post fixes many bugs reported on the forum over the past couple of weeks. If you were having any issues, please give it another try and let us know if anything is still a problem.
As always, give us as much detail as possible so that we can quickly identify and solve your issues.
First of all, i really like the improved customization features - especially how you chose to display secondary stats.
I think I know how I’ll be spending most of my time when looking for ways to optimise my equipment. 
But I still can’t get some older features to work correctly (especially regarding the on-use limit and the mechagon trinket) and noticed a weird problem with the tertiary stats:
The option to limit on-use items still doesn’t seem to work correctly though. Regardless of the exact setting AMR almost always recommends at least one on-use trinket.
I’ve generated a new export string and tried using a private window in a new browser to eliminate problems caused by caching.
Settings (as Imgur album to avoid multiple images):
https://imgur.com/a/yFypihB
Also I’m unable to lock in Punchcards for the Mechagon trinket. Currently AMR recommends using the Dodge-Punchcard (Trajectory Analysis) on a “Max Defense” setting (which would obviously be correct).
When trying to lock in Cyclotronic Blast, the socket just stays empty and reverts back to Trajectory Analysis when optimizing while still showing the orange lock icon.
In the past I’ve prioritized Versatily using the old customization options but seeing the new options I’ve tried using those - e.g. to set Versatility to 2000, Mastery to 1200, Haste to 600 and Crit to 200. This does work as expected, no issue here.
Then, I’ve also changed tertiary stats to exclude Leech (i.e. lowered Leech to 0), as I find that to be negligible as a tank - a couple percent of damage done will amount to almost nothing compared to the already huge health pools. Again, no issue there.
After that I’ve thought about setting all tertiary stats to 0 just to see if something changed. And of course, it did: When customizing both the secondary and tertiary (minor) stats I can’t set all 3 minor stats to 0, as that will throw an arithmetic error when optimising:
Ticket Number for easy copying: 41132159b6bc40e8b2159fb97f82e1d0
That’s just from a quick glance at the new options, I’ll probably take another look later today or tomorrow to see if I’m able to find other issues. 
Love the direction for the new customization features, but i have a use-case for which they don’t go quite far enough. I have long kept a ‘speed set’ for old content, and was hoping that the new tertiary stat and azerite power custom valuations would make managing it as simple as managing my current content gear with AMR.
To try out the new options, i copied my main DPS setup, relabeled it “Speed set”, set speed all the way up in the tertiary slider, and set the azerite power values for longstrider and vampiric speed to ‘sacrifice 45 levels’.
Unfortunately, the new options don’t go far enough for my edge case needs. Reweighting speed to the max, it only adds one of my +speed items to my best in bags. Ten item levels is simply not enough weight to get my older items to the top of the list. I would like to be able to give all the weight to speed, but a slider that goes from zero to 100 is going to be poor UI, and there are going to be users who slide it to 100 without realizing the effects, or who forget to put it back, and are frustrated with their strange results.
My suggestion/request is, to the right of the slider, just next to the max setting, put a checkbox labeled “Top Priority,” and when checked, weight it such that all +speed items outweigh all items without the stat. This keeps the slider in the normal ranges, and would satisfy my slightly different needs.
For the azerite power valuations, i think the same could be accomplished by having the last dropdown menu entry be “unlimited levels,” So that if an item with that power exists, it will be selected, period.
Thanks for your consideration!
I’ll take a look at the on-use thing… thought I had it fixed up with this update, but I guess that there’s still a case slipping through. And yeah punchcards… how I hate you, punchcards! I’ll look at that too.
I found the error in the log with setting all minor stats to 0… seems like some kind of divide by zero happening somewhere, should be easy enough to track down.
@paspoulet I’ll put it on my list to revisit extreme cases like yours, see if we can do something without too much messing around.
I posted an update that I think should fix these issues – give it another go and let me know if something still isn’t working.
Thanks yellowfive. I do think the existing logic will work perfectly. It’s the constraints chosen for the current UI implementation that hold me back. You’re beta is very well done and pretty, but imagining that you had an alpha that just had a text entry box where i could assign any value to speed that i wanted, i could have easily found a value that worked for me. The UI solution you selected for Secondary stats allows for maximum flexibility (within the constraints of the game), the slider for Tertiaries does not.
I’ll stop crusading here, but i am happy to discuss it more if deemed worthy of discussion. I will be a happy user whatever the outcome. Thanks again.
Looking good so far. Tried on a couple different toons and it seems to be working correctly.
Haven’t been able to find anything else I’d instantly see as a bug or issue with the new system in general. Soo… good work! 
That fix (or, more likely, other changes implemented at the same time) has changed around some gear recommendations regarding custom stat distributions pretty drastically though.
I’ve generally prioritized Versatility over Mastery on my blood death knight and almost completely neglected haste and crit while using a “Max Defense” strategy, as that seems to work best combined with the disc priest in my mythic+ group.
With the old customization system I’ve simply prioritized Versatility and results were as expected. Using the new customization “wheel”, I’ve build my own stat priority of 2k Versatility, 1.4k Mastery, 400 Haste and 200 Crit. Results looked pretty much exactly identical to what I had before as I’m relatively close to those stats.
Since yesterday this isn’t the case any more and AMR wants to make some weird changes and deviates further from the aforementioned custom stat distribution:
| Crit | Haste | Mastery | Vers | |
|---|---|---|---|---|
| Targeted Stats: | 200 | 400 | 1400 | 2000 |
| Current Stats: | 456 | 414 | 1399 | 1935 |
| Recommendation: | 659 | 643 | 1353 | 1429 |
While I definitely see the value of some of the changes recommended (e.g. using 445 feet with Crit/Mastery instead of 430 with Versatility/Haste & socket) because of stats not included in the stat distribution (like Strength, Stamina and Armor), I’m not so sure about others (e.g. changing my crafted 440 legs with Mastery/Versatility & socket to 445 legs with Haste/Mastery).
Overall I’d be losing three sockets while also shuffling about 500 points from the most important stat (Versatility) over to the least important ones (Crit and Haste) - and gaining about 150 Strength, 560 Stamina and 170 Armor in the progress.
The main question isn’t why this is happening at all (could still be correct advice, even if it looks weird) but rather why this started with the fixes you implemented yesterday.
Maybe it has something to do with tweaking the extreme increases reported on another topic (Wrong DPS increase with customize stats in best in bags)? Or it could just be correct advice even if it looks weird and I just don’t like it?
Yeah it’s a rather delicate balancing act: on the one hand you have a custom secondary stat distribution that a user wants to reach. On the other, you have to weigh how valuable that is compared to strength, stamina, armor, and other non-secondary stat sources of power for your character.
We want to keep the customization tool as simple to use as possible while still giving enough control to satisfy the advanced user. With the secondary stat pie, there is no explicit information or way to “score” how much to emphasize moving towards the secondary stat goal vs. getting more strength, stamina, etc., so we use several heuristics to strike the balance automatically. This might require a bit of tweaking over the next couple weeks.
Before the update yesterday, there was a pretty insane amount of emphasis on reaching the secondary stat goal – to the point that it was heavily skewing the results towards favoring secondary stats across the board. Perhaps the balance shifted a bit too far the other way with this update… I’ll try out your case and a few others and see if we need to dial the knob the other way a bit. (If it were as easy as just a single “knob” I’d probably expose it directly to users… but it’s actually quite a few things working together. Maybe one day I can come up with a setting for this though…)
I messed around with it a bit more… I think that the current version is optimizing better than the older version, even though it’s less “literal” about getting the exact secondary stats specified in the pie. The old version ends up compromising more than it should in order to get that stat distribution – in other words, you have to believe that secondary stats are way more valuable than they actually are (compared to non-secondary stats and azerite powers and procs) to force the optimizer to get closer to the specific allocation provided.
In your particular case, the new version has more of pretty much everything, better azerite traits, and better trinkets choices, at the expense of not going as aggressively for versatility. You can nudge it more towards versatility by making that piece of the pie bigger.
I’ll try a few more cases to see how it looks, and keep an eye on how people feel about the secondary stat allocation being more like a “hint” or “inclination” than an absolute requirement.
Oh, I completely agree with the recommendations for azerite traits, trinkets and so on. That’s why I specifically mentioned secondary stats and nothing else. Outside of specific effects those items shouldn’t affect secondary stats that much though as azerite items don’t have any and most trinkets only provide primary stats. Of course, there are some pretty strong procs but relying on those as a tank often looks like a coinflip - at least in my opinion.
Even if I’m increasing the allocation of versatility to the maximum possible (i.e. 3.4k) i’m only getting one less change: the aforementioned legs (440 socketed mastery/versatility vs 445 mastery/haste). I’d at least expect the cloak to also stay on my current one (430 socketed versatility/mastery/speed vs. 440 socketed haste/crit) although that’s something I could “fix” myself by increasing the value of movement speed. You’re probably right and I’m just putting too much emphasis on versatility though.
I’m aware what I’m doing isn’t necessarily optimal, it’s just what I found to be working best for my group (disc priest and several dps who sometimes like to take lots of avoidable damage). Previously I could just blindly use AMR to get the results I expected but maybe I just need to be a bit more open to differing suggestions.
Thanks for taking a closer look!
While looking around more and comparing the recommended stats I noticed another thing which could explain parts of my confusion. Generally speaking the defensive strategy for blood death knights seems to favor versatility which matches what’s recommended by other resources.
At about item level 438 this starts to change and the strategy seems to value mastery above everything else while the value of versatility drops pretty fast on the stat graphs. With higher item level this gets more extreme and at something ridiculous like 455 versatility looks pretty bad.
I’ve been at item level 435-440 for quite some time and only recently got some titanforged upgrades on multiple slots which caused me to be able to reach higher item levels. AMR now wants me to use those higher items which in turn obviously causes me to move further up on the mentioned switch from versatility to mastery - which in turn causes AMR to recommend mastery gems at about item level 442 or 443.
I’d guess this effect was suppressed by the old system and the rather strict following of stat allocations and can only now really take effect which is why I suddenly noticed it.
Interesting – yeah I’d have to do a few simulations to see why mastery starts to creep up at higher gear levels.
The tank gearing strategies are always a bit of a challenge, since valuing toughness has a “sweet spot” where it matters: if you are way undergeared, toughness doesn’t matter… you’ll die anyway. If you are way overgeared, toughness doesn’t matter… you’ll never die anyway. So as your gear gets closer to one of those two edges, things get harder to rank for pure toughness. (Thus why we added the toughness slider, so you can put a cutoff where it transitions to DPS gear, which generally works a lot better as you get to those edges.)
We have some code that should theoretically try to scale things up/down to keep you in that “sweet spot” over a wide range of gear, and it works pretty well generally… but maybe it’s not scaling up towards the limit of current gear aggressively enough to keep versatility higher. I’d expect mastery to jump ahead a bit if you are trying to increase your toughness on a fight that already isn’t that hard – the absorb shield will be better on low amounts of damage.
