Are AMR Discipline raid healing gear recommendations correct?


When I use AMR BiB feature as Discipline Priest with default “Raid healing” gearing strategy it gives me following stat priorities (it doesn’t change for higher item levels):
AMR(Raid): Versatility >> Critical Strike >> Mastery > Haste

This is almost complete opposite to recommendations from other sites, where everybody values Haste above other secondary stats and crit is mostly second:
Warcraft Priests (Raid): Int > Haste > Crit => Versatiliy = Mastery
Icy-veins: Intellect > Haste > Critical Strike > Mastery > Versatility
AutomaticJak (Raid): Intellect > Haste > Critical Strike > Mastery (flat) > Versatility > Mastery (procced)
Noxxic: Intellect > Haste > Mastery > Versatility >= Crit

Is this a bug?

This isn’t a bug. We include the value of all healing effects, including effects like pain suppression, leniency, damage reduction from versatility, etc. That causes the value of crit/haste to go down a bit since they don’t scale with those stats and the value of versatility to go up. We are the only people using simulation to value gear for healers as well, which often leads to different suggestions than other sources of information because we are able to use a more complete model of the game.

All the healers I’ve talked to who use the advice from our site have successful results in-game. I also main a priest and use the suggestions from the site and do extremely well.

Think of our advice as another viable way to play the game that has solid data to back it up. There is more than one way to play a healer and be good!


Does the simulation for discipline priest do the atonement ramp up + evangelism + all-out-dps prior to a big hit mechanics (does such mechanics exist in the sim?).

What would really help is some more in-depth description of how simulation works for tanks and healers with specifics to each spec individually. Even if this is true sometimes it’s hard to accept and ppl thinking that “oh they must have disregarded this and that in their model”

From my point of view:
You could do a boss script that allow the discipline to shine better.
The point isn’t to prouve for the 1% best healer the gear they should use but allow a gear choice for most situation.
So base on that, the best choice for disci could be haste if you have the situation for it; other heal integreated your role and the boss allowing it. It’s probably the gear (according to AMR) that avoid the least error with a basic healing rotation.

Also if you are looking to the stats distribution provided by AMR you will see that “haste” as the worst case 10% HPS under the rest. So it isn’t that bad.

Then what would you prefer ? Give you the best scenario and the gear with it where you could do very good sometimes and very bad other time ? Or doing average healing often ?

Sienss =D

We use the Vectus encounter for our raid healing simulations, so you are basically doing constant raid healing, with a periodic bursts that have to be healed up. The phase 2 gives you some time to heal a bit lighter.

Hi, came to say the same thing.

What if i want to sim for actual throughput ? Any way i could do that ?

There really isn’t a way to remove them from the simulator. You could change the rotation to not use pain suppression.

I find it odd that you’re kinda forcing your way of thinking while the whole community is saying the exact opposite of what you’re doing. Clearly, you’re free to think otherwise but from a business POV wouldnt that be better for you to at least have the option to sim like all the other guide are recommanding to do ?

Anyway, that’s a shame. I will resub when/if there is a way to have this feature in the future.

Thanks for the quick reply though.

Maybe we are talking about different things…

Are you talking about the actual simulator? Or are you talking about the gear optimizer?

The simulator is at and does not require a subscription to use. I though you were asking if there was as way to do a simulation that doesn’t include the damage reduction from versatility, leniency, etc. That wouldn’t be possible because the simulator is set up to simulate what actually happens in a fight in the game. And in the game, your versatility does reduce the damage you take - so I did not build in a way to ignore that. You could set your versatility to zero when you do the simulation.

If you are talking about just the optimizer - you could use the customize tools to push the optimizer away from the stats and items that you do not prefer and towards stats and items that you view as more popular.

In general on our site, the DPS specs track closely with what “they” say to do because we’re all using simulators and suggesting stuff that simulates highest. For healers, our advice often does not track with what “they” say because we are the only people using simulation. Our model is the most robust model of the game available with which to actually put some data behind a gear recommendation for healers. I have found that when I use the suggested gear and play similarly to how we recommend to play, the results are very good. There are a lot of viable ways to play a healer. What we suggest is one viable way to do it, backed up by good data that you can look at for yourself if you choose.

I’m ok with my advice being different than what is most popular. The way I look at it: I’m giving you good advice that will work. On top of that, I’m the only person out there exploring a simulation-based gearing strategy for healers. I have found that people tend to do sub-optimal things in-game while healing in order to make the meters look better. I base everything around the pure goal of keeping everyone alive in the most efficient way possible.

If you prefer to follow the popular way to gear… you don’t really need my theorycraft to do that. Click the customize tab and tell it to do what you have already decided to do. In the near future we are going to “can” some of these more popular builds as a one-click customization to save people the time of customizing themselves, as a convenience. We’re not oblivious to the other suggestions out there, but I do find them artificially limiting for healers. There is no reason healers should all feel like they need to gear and play the same way - there is “fuzz” built into the role, so we should embrace different strategies.

The only down side to this logic is that Pawn uses these weights and its promoting an alternative thinking. With everything being 8 times more worthy than haste (according to their numbers haste is 0.26 to everything being 2.0+, it swings items to be WAY off. I know this isnt your problem, I get that.

But youre actively going against everyone’s findings just to downplay haste. Disc is unlike any other class so it cant be treated like a typical healer.

I’d like to chime in again here, because my last thread on me explaining the math and data that goes against your “simulation” wasnt responded to and this relates to it as well.

What is the rotation you are inputting into your simulator? If you are treating Disc like you are treating Holy Paladins, then it explains why you are valuing mastery over haste like you do with Holy Paladins. You guys say you have the data and use the most robust simulators, please show us the data and the reasoning behind it so we can understand where it is you are coming from to suggest strategies that go against the rest of the world.

Otherwise…what you get is a lot of frustrated paying customers who are wondering why their rotations and gameplay isnt going as well as they expect it to be when using your strategy.

The simulator isn’t playing a discipline priest different than people are playing it in-game. You can run a simulation and see what it is doing.

Here’s just a random sim with random gear on:

You can look and see that it casts spells how you would expect for a discipline priest.

The reason haste is not favored is very simple: haste is only good if you don’t run out of mana. These tests are on very difficult fights where mana is a limiting factor. If mana is a limiting factor… haste has very little value when you look at how much total healing you can do over the course of a fight. Sure, your burst healing will be better, but I’m not looking at that as my measure. I’m looking at how much total healing you can do over the course of a fight. Most players aren’t top end mythic raiders. Most players are on teams with a range of skills and progression fights wind up with healers running on fumes for mana. In that case, what I am suggesting is a good idea for which stats are best.

Once you are on farm, you can change it up and stack haste to make those burst combos bigger before spirit shell or whatever. But until then, I think that maximizing total potential output is a very solid strategy for gearing. The only way to actually quantify how good haste is, in my opinion, is to have some sort of model that can track your mana. That is why I like simulation for healers instead of other methods. Ignoring mana is taking out a huge part of the equation. Mana and Haste are linked in a complex manner when it comes to how good Haste is as a stat.

Please explain then why you use this same logic for mythic + and keep talking about mana? You tried to use the same logic and argument before and again it makes no sense when talking about mythic +, especially when you can drink before each pull and get mana back often. Disc is actually pretty good at that with both the talent mana regen options as well as the ability to put shields on and then drink before combat starts.

First thing that came to mind also was Mythic+ but I also question the validity of that logic in a raid setting. MANY MANY MANY Disc Priests roll with the 27% haste recommendation, and seem to excel in raid settings without mana issues.

You dont need to look far in logs to see this. Warcraft Logs - Combat Analysis for Warcraft

Its about using the right abilities at the right time. Not non stop, straight throughput.
A couple points to be argued against youre logic for both Disc and Holy Pally (which I know you know)

  • Healing is very hard to sim. Its reactive in nature and a “rotation” will only get you so far.
  • Haste will get off a heal faster but its up to the user to build in their own ebb and flow for mana issues. Intervate, Mana Tide Totem and Mana Pots are huge tools for this as well.
  • Mana SHOULD be 0% by the end of a fight. Wasted mana is wasted healing.

I honestly think that AMR tries to hold all of the healing classes hand and prevent those deep cuts into mana for good reasons. This is good for things like Resto shaman (my main) which honestly, I dont argue the logic of AMR for this that much. Deep Healing is SUPER powerful in raid. The risk of your playstyle is that is will enforce some bad habits.

I respect your attempt to say its “an alternative playstyle” but I honestly think people come to sites like this for the best playstyle. Having little to no haste will find a tank dead before a Shadow Mend can go off, or the dots and penance going off way to slow to catch up to the tank being chunked. A more “effective” heal might get them a little more health in there HP bar, but quicker heals will get them through some serious damage phases.

Unless you just cant do it, give people an option for Holy Pally and Disc Priest to just go “Haste Build” on your site. Fiddling with the customization can only do so much when youre fighting the authors instructions.

The fact that people are doing well with builds that have haste in no way proves that a build without haste wouldn’t work. I want to at least point that out.

Healing is very hard to sim, I agree. But, I’ve done it. It took a lot of work and I took it on as a challenge to the idea that you can’t sim healers.

In the simulator, we call the “rotation” a rotation… but it isn’t that for healers. It is exactly what you described: a set of rules for the healer that reacts to the current state of the simulation. For the healing simulator, I had to build in a lot of new functionality that is not needed for DPS. Each action in the priority queue has the ability to search through all the allies in the simulation and check their state (health, what hots/buffs they have, etc) and pick actions accordingly. It is set up to play like a person. There is even some fancy stuff built in to account for the fact that people aren’t perfect - like if 5 allies are below 75% health and there is a rule to rejuv the most injured ally below 75% health… a real person won’t always choose the most injured player. If one person is at 70% and one at 67%… they will probably pick the player at 70% just as often as the one at 67%. We have though about things even down to that minute a level to make it as realistic as possible.

The simulation has ally healers, allies, real damage patterns. On top of that, I take the suggestions the simulator creates and I try them in the game. You absolutely can play a healer with low haste and be fine. Maybe that feels more difficult for a lot of people. I get that. I am not at all convinced by the argument that casting Shadow Mend in 1.2 vs 1.5 seconds is the difference between success and failure.

I agree that you can gear for haste and focus more on burst healing as a play style and be successful. I can also say with a high degree of confidence that you could do more total healing with your available mana if you use the default gearing strategy, and you’d also be successful.

And again we run into the problem where you are trying to force heal a group instead of using the classes/specs toolkit to its advantage. Like I said in my previous thread, this is why you come to the conclusion of needing more mastery.

With Holy Paladins you suggest spamming Holy Light as an actual heal. Now you are suggesting Shadow Mend. Im guessing for Holy Priest you are suggesting Heal spam. This is an attempt to force all healers into a generic basic spam healer instead of using each healers unique play style available to them.

The reason guides suggest haste is not so that you can spam your slow heals faster. The reason guides suggest haste is so you can play the unique playstyle faster. Using the built in bonuses of your talents and the spec itself. For Disc priests, its setting up burst windows of opportunity where time is a limiting factor. You only have a certain amount of time to get things done, so if you speed things up the better it come out. And its not just the generic heal you get from your spells, its the atonement or the shield you are spreading to another person.

This same thing goes back to holy paladins. You are forcing mastery to account for slow heals and making sure you get something for how long you are spending healing. You are thinking of all healers as the same, which yea when you take a step back and look at the spells they are. We all have a fast heal, an aoe heal, and a slow heal, and something to protect ourselves with, and a cd to speed things up or make it more powerful. Problem is, in reality they are not, and if you play them all in a generic way like that it hurts and you miss out on the core abilities that particular brings to the group.

I keep asking for the data to back these suggestions up as you continuously keep saying you have data, math, and the most robust simulator, and thats why you go against the world and you say they dont know what they are talking about. Please show it because right now, it seems like there is no data and its just your opinions being fed into a program and thats false advertising.

Maybe you dont play the game that often or check the forums…but I have been given 3 warnings this week alone for just mentioning AMR in discord forums. I guess you upset a lot of the other theorycrafters when they tried to talk to you about the issues as well so now they refuse to even allow it mentioned in the discords or on the guides. It actually goes worst than that as Guild Leaders go onto those discords to ask about their healers in raids and the advice is if they are using AMR to get them to stop or kick them.

So when we get into that situation where we are asked why we are doing things so differently and not following the rest of the world…we need to have a better response than we are just following the opinions of AMR developers…we need to show the data and reasoning behind it.

1 Like

Edit: retract my statements. I get where youre coming from but hey, maybe im just addicted to the smoother play of having high haste.

I never suggested Shadow Mend spam - what are you talking about? I did suggest that using Holy Light is worthwhile for holy paladin output if possible.

The discords not allowing people to talk about AMR goes back years. They don’t like us - we do our own thing. It is not worth the time to talk about. The dislike of AMR has very little to do with actual theorycraft.

I linked a simulation up there for a discipline priest. If you click on the “log” section you can see line by line every single thing that happens in that simulation. You could run a simulation for Holy Paladin as well (set output to log) and go through it to see what actions are being taken. I think you will find that there isn’t anything weird or off-meta going on in the actual simulation. The only thing “weird” is the conclusions.

I can use the simulator to illustrate all the things I am saying. Swap the versatility and haste stat in the discipline sim I linked above - output goes down. Shorten the fight length and do the same test: the haste build sims better because mana is a non-issue.

You are telling me that your experience is not lining up with the script I am basing my default gearing strategies on. I am saying: ok, no problem. We built customization tools for exactly this purpose, and we are working on making them even easier to use and more apparent. The experience of healers is highly variable depending on the group they play in. I cater the defaults to what I think is the most common experience for players, which is heroic raids with a mix of skill levels in which progression fights tend to be healing intensive.

I don’t need to be allowed to do my own theorycraft by some discord server that won’t even let me talk on it. For healing, there are many way to play the game that are all good. That has been my opinion for years and I haven’t seen anyone able to prove me wrong. I use the advice I give, and it works. I feel that it is important that I provide a different set of advice than the other theorycrafters because no one else is going to do it. This is a video game that we should feel comfortable experimenting in.

Was just looking through those logs and a couple things pop out:

  1. Its having me generate holy power when im already at 5. At one point generating 7 holy power instead of spending to fish for Awakening procs or at the least dps with SoTR
  2. It never uses a spender in the Beacon of Virtue window
  3. It is not trying to spread holy shock for legendary effect
  4. Boss is hitting non tank players for 3-5% health on a +10 which even lower keys definitely hits for far more damage
  5. its popping beacon without having much holy power and then generating with holy light
  6. Im not seeing any mana regen at all
  7. Its popping Divine Toll at 5 Holy Power…you dont ever do that as Divine Toll Generates 5 Holy Power itself so that goes back to 1) and generating 10 holy power

This is why im trying to point out that the simulator is not based on data but based on opinions and thus the output is opinion based as well…not data driven. The rotation itself is in question. Since thats in question the whole output of the simulator is in question and thus not providing the best data and options to the customers as possible which is why we use the tool and pay for it. You can say the output is logs and data…but if the input is opinions and the inner workings is opinions…then it doesnt matter what changes you make to stats or gear…the output will be incorrect. Do you see what im trying to say?