Bug. AMR rates higher item level lower for 2 otherwise identical trinkets when using stat weighting

I have a BiS trinket from gul’dan

I have 2 different item levels of it. Naturally there is no way that one could confuse the lower version as being “better” since the trinket text is identical and the one with higher item level shows the higher main stat by a few hundred.

However I have seen in at least 2 cases AskMrRobot rates the lower item level higher in value in its ranking system.

This is seen when you click on the item and see the pulldown and read the + and - numbers to see how your current item compares.

I am not confused about this. If I equip the lower item the higher one shows as -198. If I equip the higher one the lower one shows as +198. Clearly AMR is bugging out here.

The catch seems to be that I am using a Gearing Strategy that is not machine-learning. I have put in stats based on the agreed upon “best” stats in the Druid Boomkin discord (by Cyous whose guide is well known and respected and up to date as of March 31 2017). But even with your own input stats, that should have no bearing and not be the reason for the buggy results.
Because even though one puts numbers into the stat weights, the “weight” for int is the same as you swap the 2 trinkets.
The main stat of int is higher on the higher ilevel therefore even with stat weighting the rating of that item should be higher value if it has higher int. But it’s lower as the ilevel goes up.

I had a lot more detail in this but it was not relevant. One item I saw this on is Whispers in the Dark trinket. Since I saw it on multiple items I know that the item is not the problem and the stat weighting is not the problem. It’s AMR’s incorrect judgement that is somehow getting reversed in the scaling based on item level

Where are you looking at the ranking? Are you looking in the Upgrade Finder? BiS? Just clicking on the item in your gear list?

Also, give us a link to your character and your custom stat weights and I’ll take a look. (Once you input your new weights and click “save and use” - if you view the weights, the URL will point us directly to what you are using.)

I’m looking at the ranking in the pulldown after running Best in Bags and then clicking on the trinket that it chose.

I’m not looking at upgrade finder. Not BiS either.

  • Character: AMR Zoopagoo
  • URL from clicking on “Edit Strategy” for the custom stat weights I am using: v1.2 Cyous weights Spellblade
  • Yes these are after “Save and use” I have tried switching strategies to one that uses “machine learning” instead of stat weights. After doing that I did not observe the bug in rating scaling values on pulldown for the trinket. But when switching back to my favored strategy I did

Note: Why am I using custom stat weights? I tried using gearing strategy from AMR based on Mythic+ then based on Ursoc/Krosus and now Spellblade. I stay in AoE spec through heroic single target bosses so I need a bit more haste than the aoe gearing strategy gives me. So I have settled on using pawn strings for raiding.
Note that that is normalized. I think AMR advises against this. I guess I will have to bite the bullet and run simulationcraft and try Panw addon. I pay for AMR Best in Slot to do the job for me but for Heroic/Mythic raid content it feels like I’m making more choices myself for which legendary, trinket and set pieces to use.

Ah, you are using BiB. I’ll need your addon export to be able to see what you are seeing.

Just looking at the weights you are using… I can tell you what the problem is.

The first thing to understand is that stat weights cannot rank special effects like trinket procs, legendary effects, set bonuses, etc. To rank those, we use actual simulation results. We do this by doing a representative set of simulations and then using some statistics to approximate a ranking. In order for this process to work, the stat weights need to be a predictor of DPS.

Once you normalize the stat weights, they are no longer a predictor of DPS and they become incompatible with the special effect rankings stored with the gearing strategy. What has happened is that the stat weights you entered actually result in negative DPS, causing lower item level trinkets to rank higher, because you have inverted the slope!

If you want to use those weights, you should convert them to numbers that are closer to actual DPS. The easiest way to do this is to multiply the normalized weights by the weight for intellect that the gearing strategy calculated, which is 11.25. Then, you need to manually tinker with the “Offset”. You can do this by clicking the “Benchmark” button and see how far off the DPS is on average. Keep changing it until that number is as low as you can get. I did this for these weights and ended up with this:

Using those, the rankings for the trinkets aren’t all wacky - the different item level versions of items all rank in order. The actual ranking of special effects will not closely match simulation results, though. As you can see, the suggestions from the machine learning result will get you about 10x closer to the simulated DPS than these custom stat weights. But, you did say that you wanted to sacrifice a bit of DPS in this situation to get more haste.

The problem you are going to run into here is that stat weights have no way to tell the optimizer: get just a little more haste, unless you also set a cap on haste. These weights are going to dump a ton of mastery from your gear and replace it all with haste.

Oh, and also - that “KynBalSimComp” rotation is some old debug rotation. No idea why that is in the public options. You shouldn’t use that for gearing strategies - it is bad. Use “Default” and then the machine learning will work waaaay better. I’ll get that deleted out of the public options.

1 Like

Thanks so much Swol. Guess I have some sims to rerun!
This should help. Also got some better non-normalized numbers from my own simulationcraft runs. I’ll tinker with this.

Well I can’t rerun any simulations. The amr desktop client logs in and returns an error. I created a ticket :frowning:

Fixed: I moved the AskMrRobotClient folder out of the Program Files (x86) folder (where I have previously successuflly run it for months) into a folder within my users area.

Cool – we did upgrade to a newer version of .net core recently, maybe there are some changes to the security settings that caused your issue. Glad you got it fixed up.

1 Like

Ok after simming, reasonably happy ending.

I redid Spellblade-based gearing strategy with machine learning (good I saw you removed the debug Rotation option)

BiB put me in an iLevel 903 with set bonus that went pretty well last night in Heroic Nightholds.
Trinket choices were solid, at least this time it chose 895 Whispers in the dark, the other trinket, 905 Pharamer’s Forbidden Grimoire was the only one of mine it ranked above 860 Unstable Arcanocrystal (finally). 890 Fury of the Burning Sky and 885 Star Gate and Erratic Metronome ranked significantly lower in the popup trinket selection within Best in Bags.
That selection is backed up by https://docs.google.com/spreadsheets/d/12_FLVPAbGQTxSg42uC1N9r4wvtNF3swW7L3Z-aVHLgQ/htmlview which put Pharamere’s and Fury of the Burning Sky neck and neck at their item levels. Add in the fact that that spreadsheet is loosey goosey based on stats you input and its likely simplistic formulas and it seems like AMR made a great choice.

Good to hear it’s all working out for you now.

1 Like