The Restoration Itemization Thread for 3.3 - Page 2

♦ Carebare · 01/18/10, 5:11 PM

I don't think I follow your question honestly. My point is that if you must gem haste for any reason doing it with a combo of [Runed Cardinal Ruby] + [Quick King's Amber] is a loss compared to [Reckless Ametrine]x2. I would not personally advocate dropping 3/3 CF until your gear reaches a point (without gemming haste in any form) that you would gain less than half of the haste value of 3/3 CF (put more simply you have around 800-810 haste from ungemmed gear and need only 50-60 haste socketed to meet the cap). At that point it is at least worth considering gemming the difference (via Reckless) and transitioning your points back to either 11/0/60 or 14/0/57 depending on your healing style. As you can see currently on my armory, I have 3 Reckless to meet the cap with 3/3 CF.

• Hamlet · 01/18/10, 7:30 PM

Also, you don't have to pull all the points of CF at once. When you hit 775 haste on gear only, or maybe with good yellow socket bonuses only, you can drop 1 point anyway. 815 haste, 2 points. Then you can decide whether to gem the difference to free up two more points.

The most haste you can get while maintaining Trauma and 4T10 is 786 before any enchants, gems, or cloth.

goodolarchie · 01/19/10, 4:31 AM

If you're sticking with 4T9, keep in mind that crit > spirit in terms of throughput, obviously spirit wins in regen. There are cases where +spirit/haste gear is better than +haste/crit, but I've prepared a wishlist that emphasizes crit as a tertiary stat. Bare in mind this set is not that realistic since it uses a fair amount of DPS-ripe cloth and so forth. But hey, that's what wishlists are for! It is optimized for the haste cap, Trauma (buff please), using no CF (presumably Natural Perfection instead).

Chots™ Wishlist:
Profiler - Wowhead

Spell
3370 Bonus Damage
3479 Bonus Healing
30.83% Critical
35.98% Haste

modicumofrespite · 01/19/10, 11:24 AM

It seems the greatest challenge presenting itemization guidance is comparing SP (throughput) with Mana Rejen. The tank threads approach a similar topic ingeniously by measuring time till death. By assuming a canonical set of gear and typical encounter, tanks compare the value of stamina, mitigation and even threat through the "time till death" mechanism. They then present a single, and very simple chart directly comparing the value of every stat to an "equivalent stamina." In many ways their challenge is greater than ours as healers.

The question is: Is there a way to provide a single stat value (equivalent SP comes to mind) to compare all stats, including regen? Can we create a quantitative and direct "apples to apples" stat, even if only 100% valid in a single representative case? Could we let the reader extrapolate from that one case? Could this be accomplished by looking at "time to oom?" Our goal is simplified because we can assume constant casting and tree optimal "time till oom" is three minutes. More and one is "over-regen'ing" and SP may be traded for the wasted regen. In this way, all stats might be compared and the fuzzy and imprecise statement: "if your having regen problems...," may finally be eliminated.

btw: Kudos for starting a thread with a valuable list of stat comparisons. Many of the new threads seem to have removed stats in favor of a list of thread rules. Thanks for putting in the effort.

• Hamlet · 01/19/10, 11:35 AM

Is that the biggest question right now? Mana is such a nonissue for us, and has been since mid-TBC.

But here's how I've generally thought about it. First, imagine a setup which is geared for maximum throughput only. Then, for each possible concession you could make for mana, list the HPS loss and then MP5 gain (this is easy using TreeCalcs). Examples: Replace Ember meta with Insightful meta, replace IDS with Solace, replace Idol of Flaring Growth with Awakening, etc. Then you can rank these by ratio of MP5 gained to HPS lost and apply them in that order until you stop running out of mana.

Last time I checked this for myself, for example, the order or priority went:
1) Replace Ember with Insightful (this will probably always come first, Ember isn't very good).
2) Replace Sif's Remembrance with Spark of Hope.
3) Replace Flaring Growth with Awakening.

Idol of Awakening is worth so much MP5 that you'll never need to go beyond that. I haven't looked into this since then, as mana has become largely irrelevant.

Arentios · 01/19/10, 11:38 AM

Rawr can provide a measure of time to out of mana with a given rotation. However, the number of fights where you'll be using a predictable rotation is going to be rather small, if it's greater than zero. We change healing styles over the course of most fights, and many produce events which prevent us from casting entirely (raidwide stuns/silences, which reduce mana out to 0 for the duration) or force movement (which prevents the use of non-instant spells for the duration of the event). Thus, there is likely no 'single representative case' which has much practical use.

You can model 5x Rejuv + WG fairly easily, but at the same time most druids can keep such a rotation up for more than the length of a modern encounter without doing anything special for regen. On top of that, a large portion of healing is on the spot and is not going to be found in any rotation or model. You don't plan on Nourishing or Swiftmending the DPS who stands in the fire, gets hit by malleable goo, or whatever, but you're going to do so. Mana is just not a limiting factor right now in any encounter we have seen.

As many people have noted in the past, we simply don't have many choices in terms of itemization. Outside of trinkets and gems we aren't really choosing regen versus throughput, at least not in very large quantities. Even in the non-trinket cases where such a distinction exists, it's a matter of using non-spirit gear to get crit+haste instead of crit or haste, which in turn lowers spellpower by removing the spirit.

Kirbie44 · 01/19/10, 2:29 PM

Originally Posted by Arawethion

Is that the biggest [...]

Idol of Awakening is worth so much MP5 that you'll never need to go beyond that. I haven't looked into this since then, as mana has become largely irrelevant.

One thing to note again, and I do even do this sometimes, is you can equip [Idol of Awakening] in combat. Using something like this for lower damage (less healing needed) parts of an encounter. IC Hardmode was a good example. You could get away with Awakening for the first 2 phases, but when the "healing burn" came up, you could swap for a more powerful idol. While keeping Ember and Throughput trinket it, this is a way to conserve mana on parts of an encounter that don't require intensive healing.

Demagogue · 01/19/10, 3:43 PM

Originally Posted by Carebare

I don't think I follow your question honestly. My point is that if you must gem haste for any reason doing it with a combo of [Runed Cardinal Ruby] + [Quick King's Amber] is a loss compared to [Reckless Ametrine]x2. I would not personally advocate dropping 3/3 CF until your gear reaches a point (without gemming haste in any form) that you would gain less than half of the haste value of 3/3 CF (put more simply you have around 800-810 haste from ungemmed gear and need only 50-60 haste socketed to meet the cap). At that point it is at least worth considering gemming the difference (via Reckless) and transitioning your points back to either 11/0/60 or 14/0/57 depending on your healing style. As you can see currently on my armory, I have 3 Reckless to meet the cap with 3/3 CF.

I think I get your explanation now Carebare. I went and messed around with a WoWhead profile today and essentially took my current gear and made a CF build to compare to my current build using your own gear as an example on how to gem/enchant. In the end it looks like switching to CF would net me an additional 165 SP. (Give or take 2-3 SP as I would be losing 12 Spirit) There are some minor changes to crit/regen/int/spi as well but nothing drastic.

Non_CF
CF

In my case it looks like I have to make a choice between CF + 165SP vs Emp Touch, LS, and 2/3NP

In a more general note, I would actually suggest that people who are torn between the two 'paths' (Going CF vs gemming a ton of Haste) to do something similar to this. Once you have you can take your role into consideration (Raid Buffer/5x1 spam, vs more of a direct healer) and pick the one more suited to your needs.

Rijndael · 01/19/10, 3:53 PM

Originally Posted by modicumofrespite

OK, there are three people who effectively say "it can't be done." Nobody can come up with a single (spellpower) equivalent stat, so that we can evaluate gear/gems/enchants on paper. They say it's impossible or valueless and offer other methodologies to try and get-around the issue (e.g. "just keep changing gear until everything works").

Well, I think the contributers here are a lot smarter than these three give credit and that some bright person can give us a proper analysis. Are healers dumber than tanks? I think not. Show us what u got, Tree of calculation. Relate all stats to an "equivalent spellpower," so we aren't stuck "changing gear until everything works."

Can you come up with a single number (call it "awesomeness") to evaluate cars which seamlessly combines top speed, acceleration, and mileage? How important do you suppose is your car mileage if all you are doing with your car is going to a store, which is a 5 minute drive, and you get free gas between store trips?

Incidentally time-to-live, avoidance and threat are not combined into a single number when evaluating tank gear (because avoidance is about the average case, and time-to-live about the worst case). And at any rate, avoidance is generally not very useful on hard content as long as tank healers can spam and not run out of mana.

Norfair · 01/19/10, 4:43 PM

Originally Posted by Rijndael

Can you come up with a single number (call it "awesomeness") to evaluate cars which seamlessly combines top speed, acceleration, and mileage? How important do you suppose is your car mileage if all you are doing with your car is going to a store, which is a 5 minute drive, and you get free gas between store trips?

Incidentally time-to-live, avoidance and threat are not combined into a single number when evaluating tank gear (because avoidance is about the average case, and time-to-live about the worst case). And at any rate, avoidance is generally not very useful as long as tank healers can spam and not run out of mana.

I agree. Way back in vanilla I already tried to somehow make a comparison between mana and spellpower, to see where the break-even point is (X mp5 = Y spellpower). The best I could come up with, is figuring out how much healing you can do with a single point of mana and then using mp5 on an average length fight to see how much mana it will give me (and therefore knowing how much extra healing I could do with it). Then compare this number to the extra amount of healing the extra spellpower would give. Of course, this only applies if you run oom every single time, which we don't. Also you need to have a pretty straightforward cycle, as each spell gives different HPM.

Another way to value mana versus spellpower is using itemlevel (e.g. on a gem, 20 spirit is equal to 23 spellpower), but "we" do not value mana/spellpower the same way Blizzard does.

Anyway, as pointed out earlier, we don't really have to choose between throughput or mana regeneration at the moment, so therefore making an equivalent-to-spellpower-stat for resto druids is not necessary. The shaman department already tried to do this, but as you can see from the opening post this is not an easy task: [Resto] Healing Equivalency Points.

I suggest letting it go and accept that as a healer there never will be a mathematically-proof way of gearing in a similar way dps classes can.

• Hamlet · 01/19/10, 5:26 PM

Okay, the stuff about mana is pretty tangential to anything important and the things people are saying above are right. I'll just try to add a few interesting things before we move on.

1) Stat Weights. Is a "good" set of stat weights one that always picks the better item of any two? I posit the following:

Spellpower: 1
Haste: 1
Spirit: 0.2
Int: 0
Crit: 0
MP5: 0
Hit: -0.1

Well this ever lead you astray if you toss it into the Wowhead item comparator and use it to pick gear? I don't think it will. That being the case, trying to refine a "better" set of stat weights is not a high priority, as it won't matter for the rest of this expansion.

----

2) Equivalence between spellpower and MP5. My response above implicitly contained the answer to the question modicumofrespite is asking. In the list of tradeoffs you can make between spellpower and regen is "give up X spellpower for Y MP5." If that's more efficient that one of the things you're already doing than, than you'll do it. If it's not, then you won't. This allows you to define an equivalence between spellpower and MP5.

For example, dropping the Ember meta for the Insightful meta trades about 25 spellpower for about 75 MP5. If, like me, you're not even choosing to make this trade, then you can conclude that the MP5:spellpower equivalence ratio is less than 1/3. If you are doing it, then you can conclude the opposite. Dropping Flaring Growth for Awakening trades around 234 spellpower for around 267 MP5 (while using a 5x1). Most people aren't doing this, so they're implicitly valuing MP5 as less than 0.88 spellpower.

The theory is: make trades of spellpower for regen in order of increasing efficiency. When you no longer need to do so because you have enough mana, the efficiency of the last tradeoff you made is your effective spellpower:MP5 equivalence ratio.

-----

Turn around here if you don't like math.

3) Deep theory of mana consumption. Way back in my first WoW life as a Mage, I developed some theory regarding the optimization of mana efficiency in the presence of mulitple spellcasting options:
Mathematical Magery
[Math] The Two-Cycle Theorem of Spell Selection

The game has changed, but the mathematical ideas haven't. Arcane Mage theorycraft is still heavily based on this concept today. I'm going to keep this short because it's so far afield, but if anyone is really interested in the question of how to perform calculations regarding mana constraints, I'd advise you to start by understanding the second linked post.

kaellia · 01/19/10, 5:46 PM

Originally Posted by Norfair

Anyway, as pointed out earlier, we don't really have to choose between throughput or mana regeneration at the moment, so therefore making an equivalent-to-spellpower-stat for resto druids is not necessary. The shaman department already tried to do this, but as you can see from the opening post this is not an easy task: [Resto] Healing Equivalency Points.

I suggest letting it go and accept that as a healer there never will be a mathematically-proof way of gearing in a similar way dps classes can.

The current version of the resto shaman HEP (healing equivalency points) program analyzes real combat logs and creates stat weights that take in to consideration glyphs, spec, personal healing style (through spells chosen and when) and combat mechanics. Stassart has done an incredible job making this tool viable for shamans. It would take some work, but it would be possible to create a similar tool for resto druids as well.

It doesn't give you a "mathematically-proof" way of gearing, but it does help to understand the personal value of each stat, set bonus, glyph and talent point.

Edit: here's the link to the first post: [Resto] shaman_hep reports

Fallenangel · 01/19/10, 6:54 PM

Regarding mana usage and 5x1 rotation, not sure how everybody is floating around in endless mana land.
Perfect 5x1 costs 2395mp5. Going by my own armory, I'm standing at 638 mp5 unbuffed with 2 stacked solaces. Adding the following (figures taken from TreeCalc, rounded): Meta (85) + Kings (79) + Wisdom (109) + Divine Spirit (34) + Int (20) + MotW (39) + Replenishment (263) + Innervate (222), or a grand total of 1489mp5. This is a deficit of 906mp5, which puts OOM time at about 170 seconds with runic mana potion.
While this scenario might be a bit unrealistic, tweaking it with a small delay won't change the result by a lot. This is what druid regen and mana expenditure really look like.

• Hamlet · 01/19/10, 7:22 PM

Yeah, back on topic, we shouldn't exaggerate that too much. After all, most people still use Insightful for a reason. I don't think we're really in infinite-mana territory. The main point stressed in the guide is that you solve your mana problems with simple things that confer large chunks of mana (Potion, glyph, trinket, Idol, etc.) rather than scrapping for a little extra Spirit and Int on you gear. As an ultimate backstop, anyone worried about mana has the giant sledgehammer of [Idol of Awakening], so you don't have to analyze much past that (and yes, the Idol swap is surprisingly efficient).

For one, you're forgetting your Solaces (e: nope, just didn't see them before the long list). Other things:
--Mana Tide is better than it looks since it's frontloaded. Same is true for Glyph of Innervate.
--Revitalize isn't included in the sheet in any way.
--There are a lot of times where you probably don't spam from start to finish.
--Innervating yourself is okay. If you have to do it every fight, you should probably bump your regen a bit, but having the option available helps you run things pretty tightly the rest of the time.

Lightflower · 01/19/10, 10:28 PM

Endless mana land is also a function of just how rigidly you stick to 5x1. I know I often forego the WG if healing looks fine or if there is not an optimal group to cast it on (since melee are pretty much taken care of by Glyphed HL and Chain heal). I have also been known to give up a GCD to use a Clearcasting proc on LB for the minor mana regain; something I picked up on Vezax.

Perfect 5x1 is the response to massive raid damage. I would say that what most people, me included, describe as 5x1 is an approximation thereof which is tailored to the specific fight and which has a bias on rolling as many RJ as possible.