I’m using both Mr. Robot and Raidbots to find my best in bag gear. Usually they have very similar results, at most maybe 1% difference. But with my BM hunter I’m getting unexpected results now. Two examples:
1: MRR values 465 version of Epaulettes of Arcing Power about 2.5% more then 480
Amalgamated Abomination Spaulders. Probably because the 480 one doesn’t have Primal Instincts but Haze of Raze as 2nd trait besides Dance of Death that both have. SimulationCraft thinks using the 465 shoulders is a 1.6% overall dps loss. See Raidbots for example.
2: An even bigger difference are the neck essences. I’m using visions as major and MR. Robot suggest I use Breath of Undying, Unbound Force and Ripple in Space. Raidbots thinks Focussing Iris, Lucid Dreams and Breath of Undying are the minors to use. I don’t even get why I should use Ripple, does MRR really want me to run around all the time for the essence to proc? According to Raidbots the difference with just those essences is over 5% overal dps!
If I equip the top gear raidbots suggests then MR. Robot thinks I can gain over 8% overal. But when I run both sets/essences combo’s in SimC I get around a 8% loss! I’m completely unsure which to trust now. I’m ok with just a 1% difference between the two. But 8%? Either MRR or SimC must be wrong. But who?
My MRR addon export with Raidbots suggested gear set equipped:
I can’t speak to the specifics of whether mmr is wrong or raidbots is wrong, but if you are using patchwerk on raid bots, realize that patchwerk assumes you literally don’t move or stop dps on the boss at all.
as far as using Reckless force, I do feel that perhaps reckless force is rated higher for high haste builds. not sure why its ranked so high though. both my demon hunter and bm hunter are told by amr to use it, but all the logs and raid bots says to use conflict. I know someone mentioned a bug about minor essences not all being tried out after you pick your major essence. I’m going to try to delete my preset profiles and set them up again to maybe see if that changes them.
For ripple in space, we tell you how good it would be if you can reliably activate the stat proc. If you can’t, then exclude it from the optimizer and it will pick a different option.
Unbound Force is one of those minors that generally ranks pretty well if you’re willing to use it, but people just don’t like it because the buff is so short that it is subject to being wasted due to any little bit of bad luck. Sometimes it’s hard to even know when it is up during a real fight. Once again, we give the theoretical numbers and encourage you to customize if you don’t want to use it.
In general, you’ll see differences between the two simulators for a few reasons:
1.) They are slightly different models of the game and our default script includes some movement - use the target dummy script for a closer comparison.
2.) Especially for DH, the two models of the game have some slight differences. I’ve gone over them in extreme detail in the past - I’m confident their aren’t bugs - there are just differences in how we’ve chosen to model the game.
3.) SimC devs spend considerably less time on optimizing rotations for less popular builds, so you’ll generally see us saying there is less of a difference between the best build and worst build.
Treat the two simulators as two data points for your use. And make use of the customize features that we provide to push the optimizer towards what you prefer.
“Wrong” is highly subjective & ideologically misplaced - only Blizzard has any claim on right & wrong when it comes to gearing/etc… accepting that other sources have different recommendations is a better way to approach results.
Again; ‘told’ isn’t the way to approach a gearing result…a source will only recommend what you use/equip… if you see results from anywhere as ‘telling’ then communication or interpretation of said results is flawed.
Yeah, maybe “Wrong” was wrong
But something is definitely not logical at least. If I load the gear raidbots thinks is best into Mr. Robot then MRR things with switching gear/essences I can get an 11,03% upgrade. However. If I run the MRR Simulator on both setups the Raidbots set actually comes better out of it!
Those two setups are very close – 130 DPS difference, with a +/- 400 DPS margin of error on each run. Because the difference is smaller than the margin of error, you cannot say with confidence that one is better than the other. If you were to repeat the experiment, there is a high probability that the results would flip.
I think there might be some confusion going on here…
This is the gear you say that you found with raidbots: 81,931 DPS
This is the gear that our optimizer recommends you use: 82,533 DPS
The optimizer’s suggestion is just about equal to what you tested in raidbots. Of course what we talked about before regarding the minor essences still applies - exclude the ones you don’t want to use - the minor essences have no bearing here on total output if you can make use of them.
If SimC is saying you get 11% more damage by changing from unbound/ripple to formless void/focusing iris minor essences… I don’t really think that’s going to be the case in-game. They are probably just simulating those differently than we are. We might be using a more conservative estimate for formless void minor and a more optimistic implementation of ripple, or something.
I see - the estimated DPS improvement is larger than what you will actually get.
Yeah, that’s a tough one. We’re estimating the simulated DPS without actually doing a simulation. One of the effects in there is probably getting an estimate that is too high - but at least it picks the right gear!
We’ll see what we can do about adjusting it… although in this case we might not do much because we are trying to focus on more general improvements to carry over into shadowlands and it is picking the correct gear in this case.
It’s not picking the right gear either in my opinion. I think MrR overvalues the Primal Instincts Azerite trait or undervalues Haze of Rage. According to MrR the 465 shoulders are far superior over the 480 ones without it:
Yeah, the estimate is off from the actual simulated DPS a bit there… but the set of gear chosen still falls almost within margin of error of the simulator. So, when I say “it picks the right gear” - what I really mean is that it finds one of the sets of gear that is well within what it would be possible to distinguish between in-game if you were to play perfectly.
A set of gear that simulates at 81.3k DPS and one that simulates at 82k DPS are functionally equivalent.
Like I said, we’ll see if we can tweak some of the estimates a bit on the azerite to get the displayed margin to track the data closer… but the actual gear suggestions are still very good.
