Holy Paladin Haste weight

According the Pawn string (which, with the removal of the simulator and related information, is the only clue as to what the gear optimizer thinks it’s doing), a point of Haste is worth between 1/500 and 1/600 of a point of the other secondary stats for my Holy Paladin.

( Pawn: v1: “Dagoreth - Holy”: Class=Paladin, Spec=Holy, Avoidance=0.02, CritRating=6.28, HasteRating=0.01, Indestructible=0.01, Intellect=15.89, Leech=0.04, MasteryRating=5.64, MovementSpeed=0.01, Versatility=6.36 )

Avoidance is supposedly twice as good?

This seems like sheer nonsense, since Haste directly increases Holy Power generation.

Why should I believe this?

b975423535214b8c8992bfc6214b646c

Mm. The pawn string is supposed to update with your current gear. So all I can guess here is that AMR thinks you have enough haste, and any other stat is just way more value.

But that makes no sense.

Stat weights aren’t supposed to go to zero when there’s a better stat. The better stat just gets a higher weight.

And Haste Rating is my lowest secondary stat.

If your theory is correct, then Mr. Robot is saying Heroism won’t increase my healing…

It’s obviously nonsense.

Eh I think it’s really just looking at gear upgrades. Obviously heroism is going to increase your HPS. That being said, it was really nice to see some of the behind the scenes stuff to actually get some clarification.

Heroism = Haste.

If Heroism increases my HPS, so should Haste Rating on gear.

Okay, but heroism is always going to be that flat 30% haste. There is a point where haste isn’t going to be your best stat. Is the optimizer accounting for heroism? I don’t know. You can’t just stack infinite haste and assume you’re going to do more hps than having a good mix of stats. You’ll need Swol or yellowfive to respond for more detail.

Happy Holidays.

Whether haste is the best stat, or the worst, isn’t the issue.

The black box gear optimizer is saying it gives no benefit whatsoever.

Playing around, if I customize the fight length from the default of 150 to the minimum of 30, the haste weight becomes large.

It steadily decreases as the fight lengthens, which I suppose means it thinks I’ll run out of mana in fights longer than 30 seconds.

In two weeks of raiding and four weeks of mythics, I think I’ve run out of mana once - on a particularly messed-up Kurog pull.

I’m having the same with my Mistweaver monk. If I run a best in slot sim on M+ 15 keys alone it literally wants me to have 0% haste. I’m under the understanding that Haste is my most important secondary so am unsure how to build my gear set right now with AMR.

Be great to know if this is intended or a bug

Going into Customize, setting the fight length to 30, and unchecking “Bloodlust”, may give something more sensible.

I hate to say it, but I think blind faith is the only reason to treat AMR’s recommendations as more than random guesses at this point.

There’s nothing to back up any of it.

Haste! My nemesis.

For the past couple expansions we have been using the simulator to generate data for our gearing strategy. The results always favored very low haste builds because haste tends to make you run out of mana early. Haste is strong until you run out of mana, and then it is basically worthless. Your healing output will only increase with haste if you don’t run out of mana.

For this expansion I’m doing a calculation similar to what the simulator did, but without an actual simulation. This is for speed purposes. And I get the same result.

If you want to be constantly healing, you can’t go that long without running out of mana unless you remove mana-intensive spells from your rotation. Or, you have to stop healing for significant portions of time.

When optimizing for m+, you should take into account that you regen mana between pulls, so maybe 30 seconds is the appropriate length.

I’m going to revisit healers in the next couple weeks and set up defaults based on what I’m seeing in logs. If people are not running out of mana, that means they aren’t healing at 100% output. That’s fine, but it is hard to predict. You can do more total healing with your mana if you do not gear for haste, that is a fact. But, if you don’t need to, and don’t have time to use up all your mana, then haste becomes a good stat.

You can adjust the optimizer right now to favor haste. Lowering the fight length is one way. Shifting the gearing strategy more towards DPS is another way. Using the “healing activity ratio” customization is another way. Add in more idle time, more DPS time, and reduce mana-intensive time (aoe healing mostly).

FWIW, Driney, the Holy Paladin from Liquid, is running with 33% Haste; he has almost twice as many points in Haste as the next highest secondary:

Since they’re an extremely well-coordinated team, he and the other healers no doubt get Innervates to help with mana, but still.

The way healers play in high-end raid groups and super high m+ keys is not how the average player experiences the game.

They are extremely coordinated and all the players are very good at avoiding most avoidable damage. This makes the damage their groups take much more predictable. Burst healing during periods of heavy, unavoidable damage is what is the most important. In between they aren’t taxed very heavily and can throw in some DPS/maintenance healing.

I don’t think it makes sense for general healing advice to follow what players at that level are doing. I have healed a lot of normal and heroic content with teams that have poor coordination who struggle to progress even at those levels. As a healer, the amount of healing I have to do to keep everyone alive is ridiculous, and constant. Getting every last ounce of healing out of my mana pool leads to the most success (and highest numbers on healing meters). That almost always translates to low haste builds.

I am going to add a preset build for holy paladins and mistweavers which follows these popular high haste build strategies, but I don’t recommend it for average players. I have to finish up some of the code to make the custom strategy presets, but it will be ready soon.

Mistweavers are a bit of a different case. The popular strategy out there to get a lot of haste is based around using spells in a particular prescribed order to get max renewing mists out, and thus bigger vivify healing. I find the whole theory behind that to be unrealistically fussy and rigid for average players, but I will make the preset anyway!

Of course the RWF contenders are outliers.

But however good their team is at avoiding damage, the healers are still putting up 2-3 times the HPS that mere mortals could achieve, while seamlessly switching to pump very respectable damage whenever possible.

Conversely, in my experience, struggling non-hardcore groups usually wipe to unhealable mechanic failures rather than lack of steady-state healing throughput.

Further, if you will OOM 100% of the time with 10% haste, as the model seems to assume, you are probably going to OOM a large fraction of the time even with 0% haste. So the model seems questionable.

It is not hard to calculate when you will run out of mana if you keep casting spells.

The part that is hard to predict is when players will stop casting, so they don’t run out of mana. And when you stop casting, what spells don’t get cast? Usually you will strip out filler spells, like holy light or flash of light.

You will always use holy shock on cooldown and use your holy power. But you can’t cast filler spells for the entire fight without running out of mana. That is easy math. Holy Paladins have the added complexity of being able to become gcd-locked such that you can’t actually use all of your infusion of light procs in some builds. Most players avoid those builds, but I still have to model them anyway.

Healers are designed to be mana-negative. You will run out of mana if you keep casting spells non-stop. The only way to avoid being mana-negative is to reduce your healing output. Cramming as much healing as possible into huge burst healing windows will often yield maximum results, especially on the meters. Doing that will rarely yield the maximum amount of healing you could possibly do with your entire mana pool, though. Players at the high end do not have an opportunity to constantly heal at any time, so they can’t gear that way to get the best output.

Over the years, I have modeled healers in many different ways. I’ve also set up elaborate experiments in-game in the past where I had my entire raid team help me test it out.

The haste/mana relationship is always going to be something that can’t be perfectly quantified. You definitely can play low haste healer builds that work well - top tier parses can be had. In a more competitive play environment, you run into issues where healing gets sniped too fast sometimes, leaving you lower on the meters, which adds to the unpopularity of low haste, especially at those levels.

I’m going to add higher haste pre-set builds to the optimizer, like I mentioned. It is absolutely viable to play with low haste, though. And, mathematically, you can probably do more total output with lower haste if you have the opportunity to spend all your mana. The math checks out on that. The simulator backed that up too, but no one believed the simulator, either! Doesn’t matter how many ways you model it, people just like to have haste. Doesn’t really matter what the math says.

A priori, the math says a balanced distribution of stats should be optimal, since 4 multiplicative 10% buffs are better than a single 40% buff.

Back in Mists/Warlords, I wrote and maintained an addon to calculate my Restoration Shaman’s stat weights on the fly, using calculus on the healing formula:

https://www.curseforge.com/wow/addons/restoshamanstats
(page includes detailed documentation of the math)

At the time, this had to be done while in combat, since the combat log didn’t have enough information to determine each target’s health after the fact (for the Deep Healing mastery effect).

For haste effects on direct heals, I scaled the nominal, full effect based on the healer’s percentage activity in an 8(?) second window around the cast. If you were chain casting, it would count haste fully. In low activity periods, it would devalue haste accordingly. This also automatically devalued haste if you went OOM.

With the additional info now available in logs, computing healer stat weights offline using their actual casts (and the actual damage-taken pattern, healing comp, fight length, etc of the group) seems like it could be more reliable than generic sims or abstract formulation. This method can’t compare different talent builds, but it should be very accurate in quantifying stat effects.

Swol is more of the healer guy for us, but I have written a lot of code over the last 10 years for running simulations, calculating stat weights, optimizing, etc., and I can tell you that this isn’t always true. There are specs where a pretty uneven distribution of stats performs significantly better.

You can do a simple example to see how even the slightest shift in the relative value of your 4 secondary stats would make this untrue.

With a given budget, say you could get 20% crit, 20% haste, 20% mastery, and 20% versatility. And to make this somewhat realistic, we’ll say you have at least 5% base crit and 8% base mastery, so your actual values for those will be 25% and 28%. This gives you a multiplier of:

1.25 * 1.2 * 1.28 * 1.2 = 2.304

Now let’s say you redistribute your stats and dump all your haste, spread it equally over the other three stats. Your multiplier would become:

1.31667 * 1.34667 * 1.2667 = 2.246

The difference here is pretty small – you’re talking ~2.5% output difference. We ignored a lot of nuance, but this gives you the right ballpark, enough to see that even small imbalances in stat value might make completely ignoring a given stat optimal.


Calculating based on a log of player activity is actually pretty difficult and presents many challenges, though it would be cool. The first challenge is getting enough aggregate data that is sufficiently similar with which to calculate. Basing it on a single log isn’t that useful… there is too much variation from fight to fight. And most groups don’t do enough trials under similar enough conditions to make aggregating easy.

The second challenge is that each fight, and even each phase of a fight, can be pretty unique. It’s probable that a user would not want to optimize for the whole fight but only for parts of it. This becomes a difficult user interface problem, and also a set of difficult choices for most users. It starts to slip into that category we try to avoid, where a person needs to know the answer in order to use the tool that gets the answer!

Thirdly, the log-based approach runs into a huge confirmation bias problem. You’re only going to find enough data with logs that have popular stats builds, which will influence how people play, which will influence your results. It will be difficult to find alternative ways to play. This is not to say it would be impossible… just very hard. Related, the way people play is going to be based around the popular builds… which may not be the proper way to play with a completely different set of stats. It is difficult to extrapolate from logs of everyone playing high haste builds to how well a low haste build would perform.

Something abstract like a simulator is better for exploring more theoretical possibilities. A calculation model is even better in my opinion, since it is faster and even less influenced by the current meta – writing rotations for simulators is very difficult and will usually only be optimized for the more popular approaches. A calculation model can be written in a more objective manner and changed much more quickly as the game evolves.

Basing it on a single log isn’t that useful… there is too much variation from fight to fight.

Not so. The stat weights my old addon calculated from the combat log were very stable from week to week.

Also, if things fluctuated that wildly, it would make no sense to think you can analyze stats meaningfully in the first place.

I don’t understand how you think a generic sim or a completely abstract formula could possibly be more reliable than analyzing the spells a player actually casts, on the targets they were actually cast on, in real fights, with real groups.

Talk about “confirmation bias”…

It’s probable that a user would not want to optimize for the whole fight but only for parts of it.

And a generic sim or abstract formula will allow them to do this how, exactly?

With enough pulls, for example on a progression boss, a log-based approach is actually the only way to do what you describe, unless the player wants to write a bespoke simulation or (even more improbably) formulation.

Thirdly, the log-based approach runs into a huge confirmation bias problem. You’re only going to find enough data with logs that have popular stats builds, which will influence how people play, which will influence your results.

Maybe I didn’t explain the idea clearly enough. The only data any player would need are his/her own logs.

I get the approach – I just think it’s really difficult to make it work well. I’m sure you could come up with some OK local stat weights… but anything else would be tough.

For stat weights in particular, we don’t really use them anymore… but I have always preferred using a large data set to generate them and smooth out noise.

I think you are both touching on some valid points here, so i wont really put too much here, but as someone who has healed since classic, and has oushed cutting edge and world firsts over the year there is definitely a significant difference in maximum output and effective healing.
From my understanding amr focuses on maximum output, whereas a lot of players want effective healing.
Generally speaking if you arent in a mythic raid team and arent running less healers then pugs(our mythic team runs 3 healers with an off spec healer that can change should the healing be partifularly tough on a fight). In such a situation maximum output can be great. But if you are running a heroic with 3-4 other geared healers haste is going to be “more” lucrative because your healers are fighting for effective heals. There simply isnt enough damage going out in most cases for healers to get there heals in on the meter.
This is why haste is favored by most guides because they are looking for how to look good, but that doesnt mean you are optimally healing.

2 Likes