Resto Stat weight?

[Deleted]

I know the basic Stat priority is Int>Mastery>Crit>Haste>Vers

but im trying to get some values to setup my pawn.

noxxic says: Intellect [9.06] > Mastery [7.56] > Crit [6.06] > Haste [4.56] > Versatility [3.06]

how viable are these numbers considering they are coming from noxxic.

edit: tried to simcraft statweights simulationcraft keeps crashing when i use healing template

[Shawken]

I'd like to know this, too. I have two 840 rings and an 840 neck and want to know if I can swap to my 850's without mastery.

[Deleted]

I'd like to know this, too. I have two 840 rings and an 840 neck and want to know if I can swap to my 850's without mastery.

It's always been the same for resto shaman. In progression, Mastery is king due to people learning the fight and your mastery triggering since people are learning the fights and taking unnecessary damage. The more people know the fight and take the damage they only need to take, your mastery will lower. The higher health people have, the more value crits get since mastery loses it's potential. Haste is merely an output stat which increases the amount of HPS you can dish out, but with a lack of crit you could be struggling on the mana front.
Personally i'm sitting on 10k mastery with 20% crit. Unless you're doing mythic dungeons and mythic+ then i'd suggest you use the 850 rings/neck for the intellect, but for anything like mythic/+ and raids, mastery will be better than intellect most of the times due to the progression

[Shawken]

Rings and necks no longer have main stat, hence my concern on which one to wear.

[Axelond]

I would suggest something like this

Int 1.0
Mastery 0.65
Haste 0.60
Crit 0.60
vers 0.55

[Koor]

I wrote a spreadsheet which implements the formulas I detailed in http://www.mmo-champion.com/threads/...y-stat-weights.

1. Copy your intellect, crit, haste, mastery and versatility into the left column cells B4-B8 in:

https://docs.google.com/spreadsheets...ub?output=xlsx

(We need to enter the raw values, not percentages -- percentages are calculated automatically in column C).

2. Set the assumed target HP in cell B10 for mastery calculation, for example 0.25 for a target with 25% HP.

3. Then the stat weights will appear in the right column in cells E4-E8, and can be copied into an addon such as pawn (https://mods.curse.com/addons/wow/pawn) to show the gear score in the tooltip, and whether it is an upgrade.

[Gardiff]

Hey Koor I made something similar a while back similar to this.

http://www.mmo-champion.com/threads/...Stat-Simulator

I didn't include intellect but it was more for the gearing up process (i.e. buying obliterum gear to 850 ilvl and choosing your own stats) and so that is why I did not include intellect (since we don't really get to 'choose' intellect, rather it just comes from ilvl which usually is a no-brainer). I also didn't go the last step and actually come out with the weights like you did but I wanted to compare. It is a bit different in that I believe mine shows what you should currently change to optimize, whereas yours is an actual thing you can use for upgrades.

That being said I have two questions.

1) Does this take into account baseline 5% crit, baseline 8 points (24%) of our mastery?
2) Does this take into account extra weights for haste/crit because of things like haste double dipping (riptide, hst, htt, HR - you get more ticks and a lower cast time / gcd) and crit's bonus effects (mainly just QA)?

I have been using my own but it doesn't really help for the gearing up process, it is more about optimizing what you already have on.

Anyhow my main question is how are you valuing crit the way you are for things like QA/resurgence. I estimated it (and gave some reasoning to why I said crit should be valued at 20% higher value), but left it open ended as to what the user would value it as if things were normalized. Similar thing with haste for the aforementioned things.

[Koor]

Hey Gardiff, it's similar, main difference is I didn't take into account talents or artifact traits (so the model ignores QA and resurgence). It's only based on the basic heal equation, so in that sense it only gives a ballpark number.

The base 5% crit and 24% mastery are taken into account (in the formulas for column C that convert secondary stats to percentage).

I think that the lower cast time and gcd of haste should already be taken into account by assuming linear benefit to haste, and we shouldn't add an extra gain on top of that. The reason is that we assume in the equations that 1% haste increases throughput by 1% regardless of the spell (if we start with 0%). With hots it happens by having more frequent ticks, and with direct spells by allowing them to be cast more often.

Now with spells that contain both a direct part and a hot part, like riptide, since the 1% haste is applied to both, it also assumes that the direct part is 1% stronger. But of course it isn't, it heals for the same amount, but what it actually means is that the GCD is faster by 1% and that is equivalent to a 1% increase in the direct healing spells (could be any other direct healing spell that we cast in that time we saved).

In a sense what I mean is that I think we don't double dip on haste for riptide/HST/HTT since they have cooldowns. It's a bit different with a spell such as the druid's rejuvenation since it doesn't have a cooldown / charge time. Our riptide still has 6 seconds recharge time regardless of haste (I believe it isn't reduced), so if we have more haste we decrease the GCD but we can't have move riptides going around since we're still limited by the 6 seconds recharge, where a rejuv will be stronger, but the druid can also have more rejuvs active with more haste, so there it double dips.

[Gardiff]

Hey Koor, thanks for the reply.

I understand what you are saying here, but I am going to use the rejuvenation example and will later explain why it makes sense in the context of high HPS even for shaman.

A druid who stacks 20% haste (6500 stats) for pure rejuv spam for 30 seconds would get 20 casts off (300% spellpower spell) which would do 6000% total healing before other considerations.

With 20% haste, they now get off 24 casts which is clearly a 20% increase in output. However, because rejuvenation also benefits from extra ticks of 20% haste, in reality each of the 24 rejuvs is actually healing for 250% (the initial 50% tick of a normal 6 tick (300%) rejuvenation has to be excluded when considering the secondary effects of haste) * 1.2 or in total, a 16.67% boost per rejuvenation (250 * 1.2 + 50) / 300. Thus, each rejuv here is actually healing for 350% spellpower whenever you are at 20% haste.

Thus, the 24 casts are now healing for 350% spellpower each. This in total is an 8400% heal over the 30 second duration. Thus, 20% haste actually resulted in 1.4x healing for the rejuvenation example (8400% vs 6000%).

Obviously other classes don’t have spammable HoTs as strong as druids, but healing rain and riptide HoT along with HST/HTT do in fact easily contribute 20-30% of our output in a raid and they all behave the same as above.

And in a high HPS scenario (i.e. the whole point of following a stat weight program like mine or yours), you are going to be chain casting / gcd capped. I think this is something I should have clarified. Obviously if you are not chain casting, then haste isn’t going to be useful and I completely agree with that, but I am not seeing the relevance of addressing an issue like this because it isn’t something you would gear around for mythic content.

When it truly was a ‘healing’ wipe in early mythic HFC (i.e. not from failure of mechanics) on fights like mythic council / mythic gorefiend, healers were frequently GCD capped at certain points of the fight where a “healer wipe” could occur (bugged marks week 1 in combo with the whirls on council as I am sure you are aware of, or feast phase gorefiend). Thus, I think simply not taking this value of haste into account isn’t exactly fair.

Going back to riptide (or other CD spells like HST/HTT), you are right that it is limited by a 6 second CD, but the comparison still makes sense AGAIN assuming there is some sort of GCD capping of spells. Again, you are completely correct if we are simply stopping in between casts that haste doesn’t deserve to double dip, but again, I would point out that isn’t exactly what people tend to gear around when they are using sims to tell what stat to stack. If you are truly wiping to some failure of a healer (and not from mechanics), chances are high that you are GCD capped and chain casting for a long period of time. You can't use riptide again for 6 seconds, but you are filling in the void with other spells which also have their GCD reduced from haste. And then, the HoT component that is spawned from a riptide shouldn’t simply take into account the extra ticks from haste, but also the fact that the GCD was lowered and you are immediately casting something else as well right after the riptide. Similarly with HST/HTT.

That was my logic anyhow on why I gave haste a bit of a weight. I did something for crit as well (because of QA), but I think people tend to overvalue it both from an HPM consideration (resurgence is pretty laughable in the grand picture for stacking something like 20% crit) but also not taking into account the plethora of passives (tidal waves, floodwaters, empowered droplets) we have that actually de-value stacking crit even further. Haste, on the other hand, is not damaged at all from QA / tidal waves / wavecrash if you are again assuming a high HPS scenario where you are chain casting.

[Koor]

I did assume that we're chain casting (GCD capped). The reason is that riptide has 250% direct heal and 300% hot. When I assumed that 1% haste increased throughput by 1%, this applied both to the hot and the direct heal part of riptide, not only the hot.

The hot part, the 300%, is increased by 1% by having 1% more ticks in the same 18 seconds period.

However by assuming that our entire throughput is increased by 1%, I also assumed that the 250% direct heal part got 1% bonus as well. It isn't actually stronger of course, but this takes into account the reduction of the GCD by 1%, estimating the gain as 1% gain to the direct heal part of riptide.

Had we stopped casting, the weight of haste would have needed to be reduced further below the 1% linear assumption, since it would then only affect the hot part and not the direct part.

Maybe we can argue that in that time we saved by lowering the GCD, we could have casted another higher hps spell (say healing surge) instead, and therefore my assumption of 1% gain on the 250% spellpower direct part of riptide was too low. But then it starts to get more complex to model without looking at the entire cast sequence.

I agree that resurgence is weak. Both resurgence and QA make crit more desirable, but then again crit has the usual problem of being an unreliable stat due to RNG, which is also not being modeled when we look at averages.

[dadev]

I think what Gardiff tries to say is that the HPET (heal per execute time) of said spells double dips from haste, and indeed it does. For that metric (and those specific spells) haste scales not linearly but quadratically. Whether the spell has a cooldown or not is completely irrelevant to its hpet.
Obviously, if all we cared for was hpet then CH + High Tide wins the competition and has no cooldown, but spamming that will oom you quickly and it's not working that well for spread out groups so choices have to be made.

EDIT:
I just did some quick napkin math, and unless I had some error or something, the double dipping shows in real situations. For instance, if you have 25% haste, then during lust your RT has higher hpet than CH (with high tide) before crit, mastery and vers multipliers. crit and vers are straightforward, but counting in mastery might prove a little tricky and can pull the weight either way. However, the mere fact that a spell which costs almost third of the mana of another spell while doing similar levels of healing for the time it took you to cast it indicates that haste is indeed quite valuable stat for some spells.

[ryklin]

every stat gets worse as you get more of it, including haste, and nothing scales linearly unless you're incapable of basic multiplication and division, the issue with stat weights is the more you get of any stat, it gets weaker while the other 3 get better

and haste's big weakness despite giving more healing than other stats like crit per point, even at low levels is that it doesn't increase hpm for direct heals like crit/mastery/vers can, but I mean it's definitely still good (especially if you can get most of it out of queen's ascendant relics for when you need it for heavy chain heal spam mana dumping without gimping other stats too much)

crit, we get a ton of from our artifact, but is definitely still strong, and provides resurgence, so it ends up being a bit weaker when your main concern is just surviving heavy damamge and getting out the most healing, but is great when mana matters

mastery, can depend on the fight, but it's pretty fantastic as a go to for progression, and may not actually be that good on some fights, but you'd have to look at your logs and decide for yourself and is generally good enough that you can just treat it as strong

and vers just has god awful rating to % value, but when you start getting a lot of other stats, it starts to catch up pretty quickly as a flat increase to all healing while for instance crit will only improve the 65% of your heals that aren't critting

and flat weight values to compare them by will only just get more and more inaccurate the more and more gear you get

[Gardiff]

I think what Gardiff tries to say is that the HPET (heal per execute time) of said spells double dips from haste, and indeed it does. For that metric (and those specific spells) haste scales not linearly but quadratically. Whether the spell has a cooldown or not is completely irrelevant to its hpet.
Obviously, if all we cared for was hpet then CH + High Tide wins the competition and has no cooldown, but spamming that will oom you quickly and it's not working that well for spread out groups so choices have to be made.

EDIT:
I just did some quick napkin math, and unless I had some error or something, the double dipping shows in real situations. For instance, if you have 25% haste, then during lust your RT has higher hpet than CH (with high tide) before crit, mastery and vers multipliers. crit and vers are straightforward, but counting in mastery might prove a little tricky and can pull the weight either way. However, the mere fact that a spell which costs almost third of the mana of another spell while doing similar levels of healing for the time it took you to cast it indicates that haste is indeed quite valuable stat for some spells.

Right this is what I am trying to say. And from raid testing on beta extensively, I can assure you that haste doesn't magically make you OoM. Bad play makes you OoM. Healing with the wrong spells at the wrong times makes you OoM.

Stacking 25% crit does and gaining maybe an additional 300000 mana in an encounter where your passive regen alone makes up 3 million mana, isn't going to make a dent in pretty much anything.

I dunno, I was a proponent of crit early on but when I looked into it and how it interacted with QA and other things on a more in depth level, it really doesn't deserve the love or weight that people give it. And yet I still see some shaman valuing it above mastery (not using sims mind you, but just what they believe) and I am just perplexed how much misinformation has spread and what it does. It is GOOD if you are truly massively CH spamming and this does happen on a few fights, but no where near the amount that people think. And on those fights that it doesn't crit becomes nearly worse than versatility because of how much passive crit we have on some of our spells.

[Koor]

I think what Gardiff tries to say is that the HPET (heal per execute time) of said spells double dips from haste, and indeed it does. For that metric (and those specific spells) haste scales not linearly but quadratically. Whether the spell has a cooldown or not is completely irrelevant to its hpet.

It's slightly less than quadratically, but basically it's correct if we look at HPET: Baseline riptide has 250% sp direct heal, and 300% hot, therefore its HPET with a GCD of 1.5 seconds is (250+300)/1.5 = 366.66. With 25% haste, we'll have HPET = (300*1.25+250)/(1.5/1.25) = 625/1.2 = 520.83. Therefore the scaling in HPET is 520.83/366.66 = 1.42, significantly more than 1.25 (but slightly less than 1.25*1.25 = 1.56).

But is HPET the correct metric to look at? Ideally we want to use HPS to derive stat weights, not HPET. A druid can convert the HPET benefit into HPS since rejuv can be spammed, but we are locked with the 6 seconds riptide cooldown. It's not relevant to the HPET of riptide, but it's very much relevant to the HPS we can gain out of more haste. Yes, 25% haste gained us 0.3 seconds. Out of 1.5 seconds it looks like much, but in contrast to a druid, we can only benefit from those 0.3 seconds gain once every 6 seconds, not every 1.5 seconds like a druid, so the HPS benefit from haste is four times less than the HPET numbers would suggest.

- - - Updated - - -

every stat gets worse as you get more of it, including haste, and nothing scales linearly unless you're incapable of basic multiplication and division, the issue with stat weights is the more you get of any stat, it gets weaker while the other 3 get better

That's true of course. The spreadsheet and the equations I linked take this into account. The weights depend on our current stats and aren't static, and don't scale linearly (except if we start with 0 of that stat). The stat weight will automatically decrease the more we increase a specific stat.

[Gardiff]

It's slightly less than quadratically, but basically it's correct if we look at HPET: Baseline riptide has 250% sp direct heal, and 300% hot, therefore its HPET with a GCD of 1.5 seconds is (250+300)/1.5 = 366.66. With 25% haste, we'll have HPET = (300*1.25+250)/(1.5/1.25) = 625/1.2 = 520.83. Therefore the scaling in HPET is 520.83/366.66 = 1.42, significantly more than 1.25 (but slightly less than 1.25*1.25 = 1.56).

But is HPET the correct metric to look at? Ideally we want to use HPS to derive stat weights, not HPET. A druid can convert the HPET benefit into HPS since rejuv can be spammed, but we are locked with the 6 seconds riptide cooldown. It's not relevant to the HPET of riptide, but it's very much relevant to the HPS we can gain out of more haste. Yes, 25% haste gained us 0.3 seconds. Out of 1.5 seconds it looks like much, but in contrast to a druid, we can only benefit from those 0.3 seconds gain once every 6 seconds, not every 1.5 seconds like a druid, so the HPS benefit from haste is four times less than the HPET numbers would suggest.

The point he and I are trying to make is that you are casting other spells which normally benefit from haste as well (which also normally benefit from crit/haste). It is clear we can't spam something like rejuv since all those things are on CDs. But what we are saying is that neglecting the subsequent value of haste for everything in between the riptide doesn't really make sense. Every cast at 25% haste for something like HS will also be 0.3 seconds faster since the GCD is lowered by that much. It is just that riptide / hr / hst / htt also benefit secondarily which is what I was estimating as 20-30% increased weight for haste since in raid testing I did, that is what these components comprised.

It doesn't have to be just riptide that we are casting over and over, it can literally be ANY spell since all the direct heals still benefit directly from haste as they do from crit / mastery. It is just that when we do use riptide with haste, it secondarily benefits.

[Koor]

The point he and I are trying to make is that you are casting other spells which normally benefit from haste as well (which also normally benefit from crit/haste). It is clear we can't spam something like rejuv since all those things are on CDs. But what we are saying is that neglecting the subsequent value of haste for everything in between the riptide doesn't really make sense. Every cast at 25% haste for something like HS will also be 0.3 seconds faster since the GCD is lowered by that much. It is just that riptide / hr / hst / htt also benefit secondarily which is what I was estimating as 20-30% increased weight for haste since in raid testing I did, that is what these components comprised.

It doesn't have to be just riptide that we are casting over and over, it can literally be ANY spell since all the direct heals still benefit directly from haste as they do from crit / mastery. It is just that when we do use riptide with haste, it secondarily benefits.

I agree with the basics of this, just not with the numbers. With 25% haste, my equations did attribute 25% throughput increase to the direct heal component of riptide to account for those 0.3 seconds, I didn't ignore it, I'm just arguing it's already included without adding extra stat weight. I can also agree that it could be worth more if we use those 0.3 seconds to cast healing surge.

But it's still 0.3 seconds gain every 6 seconds. And for HST/HTT it's even less - it's 0.3 seconds gain every 30 seconds for HST, and every 3 mins for HTT, basically it's negligible. It won't be correct to say that since riptide/HST/HTT amounts to 20%-30% of our healing breakdown, we increase that stat weight of haste by that 20%-30%. For HST/HTT the gain is basically only the increased tick rate.

(EDIT: and of course the 0.3 seconds benefit to every healing surge in between riptide isn't ignored when calculating the weights -- that's included by default for all spells by the (1+H) component in the healing equation that multiplies the spellpower coefficient).

[Gardiff]

Is there a reason you are not including HR? It is the biggest contributor and also the biggest reason why haste is valuable. In all the beta tests I did, it routinely made up 10-15% of my output alone.

[Koor]

Is there a reason you are not including HR? It is the biggest contributor and also the biggest reason why haste is valuable. In all the beta tests I did, it routinely made up 10-15% of my output alone.

Sorry, just missed HR in the copy/paste, but I think HR should get the same treatment of riptide, just with 10 seconds instead of 6 -- we get increased ticks (the normal haste effect), and the secondary effect is the 0.3 seconds gain (with 25% haste) every 10 seconds.

[Gardiff]

So this is how I am seeing it and what I am trying to convey.

If you disregard completely the bonus effects of haste w/ riptide/hst/htt/HR, then 25% haste (as we have been using in our example) will linearly increase output by 25% because you will have increased casts in a set time period.

I.E., in a 15 second intensive window, let's say you are purely casting CH. You get 6 off with no haste, or 7.5 off with 25% haste so its a 25% bump in output. This is true for every spell we cast whether it has a CD like HST/HTT/riptide/HR and it doesn't change as long as you are chain casting.

Now a separate issue - Look at the secondary effects of haste on those above spells. The fact that you get more HTT ticks, more HST ticks, more riptide ticks, more HR ticks are all completely separates issue that has to be looked at independently of the fact that you got 25% more casts. That is why I am saying its not just a 0.3 seconds gain every 10 seconds (or 6 seconds or 30 seconds or 3 minutes). It is 0.3 seconds gained for every spell (that happens to be 1.5 seconds, obviously larger if the base cast is longer). Then on top of this, we count secondarily all the ticks which is usually somewhere in the ballpark of augmenting each of those above spells by ~70-80% (since you never count the initial tick) of whatever haste value you are at. And since in my personal logs, those 4 combined to be about 20-30% depending on the fight, I was giving a weight for haste based on this value at about 1.2x its normal value. It might be closer to 1.15 depending on playstyle, but it still is stronger than people give it credit for.

[Koor]

In the model I used (1+H) factor for haste, for all spells, regardless if they are hots, direct spells, or combination of those:

L = (1+V)(1+C)(1+H)[1+M(1-P)]IE

(The complete equations are at http://www.mmo-champion.com/threads/...y-stat-weights).

For a direct spell, let's say healing surge, 25% haste would multiply the healing amount in the model by 1.25. In practice it shortens the cast time of healing surge by 0.3, which allows us to cast 25% more healing surges. So this effect is taken into account without additional stat weights.

We don't have a pure hot, but if we had, again 25% haste would attribute 25% more throughput for it in the model. The mechanism would actually be 25% more ticks, and this is again taken into account.

The secondary gains -- reductions in the GCD when casting riptide/hr/hst/htt, occur infrequently -- they are 0.3 gain every 6/10/30/180 seconds. In the model above I take them into account since I multiply all the healing by (1+H), including the direct healing components of the spells that contain both direct heals and hots.

So for riptide, for example, the model adds 25% throughput not only to the 300% hot, but also assumes that the direct part is now 1.25*250 = 312.5% spellpower because of the additional 0.3 seconds we gained. We may argue that it's too low if we use those 0.3 seconds to cast other spells such as healing surges. That I can agree, though it'll make the model complex since it'll then assume that our spell breakdown changes.

But beyond that, I don't see where additional stat weights are required. The (1+H) factor already accounts for both cases -- more casts for direct spells, and more ticks for hots. We can't have more casts of riptide/hr/hst/htt themselves because of the cooldown.