[Shadow] AskMrRobot - are the stat weights correct?

[Deleted]

It's telling me that intellect is worth 9.9
haste 8.8
crit 6.56
mastery 6.47
versa 5.61

Where does it get those numbers from? I thought that with s2m haste is worth nearly x2 intellect, and certainly alot more than the difference it has with crit atm?

[Swol]

Those weights were made for using Mind Spike.

I am going to be making a set of weights for Surrender to Madness soon (maybe today!). I just updated the rotation today to handle Surrender to Madness a bit better than it was before.

The relative power of Surrender to Madness is highly dependent on the length of fight you are doing, as you'd probably imagine. The longer the fight, the less powerful it is. My initial weights were based on a fairly long raid boss encounter, so, Mind Spike is just as good as Surrender to Madness in that situation.

Haste will certainly be very strong for Surrender, but probably not 2x Intellect (unless looking at a really short fight). I'll have some numbers available soon that I can share with folks, along with some of my other findings.

[Deleted]

Those weights were made for using Mind Spike.

I am going to be making a set of weights for Surrender to Madness soon (maybe today!). I just updated the rotation today to handle Surrender to Madness a bit better than it was before.

The relative power of Surrender to Madness is highly dependent on the length of fight you are doing, as you'd probably imagine. The longer the fight, the less powerful it is. My initial weights were based on a fairly long raid boss encounter, so, Mind Spike is just as good as Surrender to Madness in that situation.

Haste will certainly be very strong for Surrender, but probably not 2x Intellect (unless looking at a really short fight). I'll have some numbers available soon that I can share with folks, along with some of my other findings.

Oh thanks for the reply!

What kind of boss duration do you think you will do with S2M? Most of the sims are all around 9-10 minutes which overall lowers our DPS due to s2m not being out for big% of the fight, which ultimately lowers our haste weight - are you going with that number or something more average, along the nature of 6-7 minutes? I guess mythic fights and 9-10 minutes is fine.

[Deleted]

Oh thanks for the reply!

What kind of boss duration do you think you will do with S2M? Most of the sims are all around 9-10 minutes which overall lowers our DPS due to s2m not being out for big% of the fight, which ultimately lowers our haste weight - are you going with that number or something more average, along the nature of 6-7 minutes? I guess mythic fights and 9-10 minutes is fine.

Go on the H2P forums for stat weights which have been calculated correctly.

[GT4]

Those weights were made for using Mind Spike.

I am going to be making a set of weights for Surrender to Madness soon (maybe today!). I just updated the rotation today to handle Surrender to Madness a bit better than it was before.

The relative power of Surrender to Madness is highly dependent on the length of fight you are doing, as you'd probably imagine. The longer the fight, the less powerful it is. My initial weights were based on a fairly long raid boss encounter, so, Mind Spike is just as good as Surrender to Madness in that situation.

Haste will certainly be very strong for Surrender, but probably not 2x Intellect (unless looking at a really short fight). I'll have some numbers available soon that I can share with folks, along with some of my other findings.

If you're implying that mind spike simmed "just as good as surrender to madness" for any reasonable fight length, either your rotation or your simulation is significantly off.

Mind spike can trivially be proven to result in a negligible DPS increase whereas S2M results in a very significant one even for long (e.g. 10 minute) fights given how your DPS multiplies at the end of its duration.

[Swol]

Oh thanks for the reply!

What kind of boss duration do you think you will do with S2M? Most of the sims are all around 9-10 minutes which overall lowers our DPS due to s2m not being out for big% of the fight, which ultimately lowers our haste weight - are you going with that number or something more average, along the nature of 6-7 minutes? I guess mythic fights and 9-10 minutes is fine.

I am using a 450 sec +/- 20% fight length for my generic single target weights. So, 6 minutes to 9 minutes. We are still working on the back-end code that will automatically pick the weights appropriate for people's talents on our site, but, here are the weights I will be using:

Haste     12.86
Intellect 10.31
Crit      9.80
Vers      7.09
Mastery   6.93

As far as calculating a relative value between Surrender and Mind Spike... that's a tough question to answer. Against a single target, if you can get a long void form at the end of the fight to keep surrender going for 90+ seconds, it will do more damage. But, if anything happens that makes it so you can't get that full benefit, you will, of course, lose the edge. If there is the possibility to cleave for a substantial portion of the fight, Mind Spike will be competitive as well.

Part of the reason that you will see other sources of information regard Mind Spike as WAY behind Surrender to Madness is because they have been exclusively using SimC to do their testing. SimC's shadow priest module has been very buggy all through beta and into the start of Legion. For example, Fortress of the Mind is not buffing Mind Spike damage in SimC. Still. Up until two days ago, it was using the 6 second in-voidform cooldown for Mind Blast for the entire fight. The APL in SimC is sub-optimal for all talent builds and will often crash when simulating Surrender to Madness.

It might sound like I'm just ragging on a "competitor" - but really one of the main reasons I made my own simulator was specifically for Shadow Priest simulation and theorycraft. The Shadow Priest module in SimC has been buggy and under-served for a long time. Example: For all of WoD, and actually even before that, it was impossible to channel mind flay for just one GCD(2 ticks) in SimC. It was a known bug that the developers deemed not worth fixing. They were totally fine with just not simulating shadow priests accurately. That bug was in there for YEARS. They fixed it in Legion... at exactly the same time we put up a beta of the AMR simulator and advertised that you could actually simulate shadow priest mind flay clipping correctly. And now, I'm seeing the pattern continue. I told them about the Mind Blast bug a while ago. It took a post in this forum somewhere by Djriff a couple days ago to get them to actually fix it. And now basic talents are still broken.

Shadow priests are unique in that they can get insanely low global cooldowns and they also have a lot of spells with cooldowns which end up sync'ing closely with the GCD. This presents one of the most challenging cases to simulate. SimC did not account for the fact that the game essentially lets you "queue" spells a little bit early - even if they are not quite cooled down yet. As Shadow Priests know, you can press the button to cast your Void Bolt a bit before it is technically "ready" and the game will then chain it as soon as it is ready. Guess what? That wasn't programmed into SimC at all. That is why the shadow APL has those silly "wait" actions. They were there to make sure the simulator wouldn't pick a different action when something like Void Bolt or Mind Blast was really close to being ready. Once I explained how this all works to the main developer of SimC's guts, he added it in. I guess no one ever told the people working on the shadow priest stuff, though... you can take out those wait actions now and it will work just fine.

I know the AMR simulator is new so trust must be earned. Hit me up anytime if you have specific questions about the results or accuracy of our simulations.

And then there is the whole issue of how to calculate stat weights. Before you calculate weights, you have to first decide what you want to use them for. Do you want general advice on the relative values of stats across a wide range of gear? Those weights would be good for finding a BiS set of gear. Or, do you want weights that are going to tell you, based on your current gear, what is the next item you should get?

The weights I create are the general advice weights that can be used to calculate BiS sets of gear and generally rank gear well over a wide range.

The weights that SimC creates by default with its stat scaling (and the weights you see on almost every other site) are weights which are very specific. And, they are specific to people who already have BiS gear. They shouldn't be posted as "general" advice because they are NOT general advice.

Very soon (hopefully next two weeks) I will also have weights and gear rankings available that are specific to Emerald Nightmare boss fights for multiple talent combinations. I'll let folks know when they are ready.

[Kilee25]

As someone who has had to write their own APLs for shadow for three expacs because of the issues you've just brought up, I am very happy to hear all of this. Competition and second opinions are extremely important in the TC community. I hope that people see that for what it's worth. Seek out multiple answers as often as you can.