First post. For the sake of this discussion, please leave pvp out of this. This is for comparing pve bullshit only, and likely only relevant to raids.


So, might as well introduce myself. Ex wow player, played a raiding paladin on the Cenarius server. I began raiding right around the time Ragnaros was killed, and stopped not long after AQ came out. Guild was easily within top 3 on server, at times #1 guild, though that may depend on who you ask. Quit due to real-life priorities and guild somewhat collapsing.

Anyway. I looked up the stats for the pally tier3 coming out of E3 (http://www.malicesb.net/wapnew/tier-3.php) and decided to run some numbers on it, just to see how ridiculous it might get. When i quit wow, i was using full T2 with some ZG equipment, and the silithus lord neck.


Before we get into the meat of this post, here's some numbers/gear.

For these numbers, i'm assuming a talent build similar to this and several pieces of other 40m raid epics. They are listed below.

Wormscale Blocker http://www.thottbot.com/?i=53397
Ring of the Devoured http://www.thottbot.com/?i=53097
Cloak of Clarity http://www.thottbot.com/?i=53384
Angelista's Charm http://www.thottbot.com/?i=52757
Shard of the Scale http://www.thottbot.com/?i=36048
Rejuvenating Gem http://www.thottbot.com/?i=44657
Scepter of the False Prophet http://www.thottbot.com/?i=53400

Rest of the gear is full Redemption. Enchants are as follows.

+22 intellect on weapon.
+30 healing on gloves.
+4 mana per 5 on bracers.
+4 stats on chest.
+7 stamina on shield.
+7 stamina on boots.
+8 intellect on head.
+8 intellect on legs.

Added into that was all non-party specific buffs. PWF, AI, mark, iBoW, and Kings.

Total stats with above gearing and enchants worked out to be as follows.

Strength 139.9
Agility 89.3
Stamina 506.4
Intellect 523
Spirit 111.3
+heal 790
SpellCrit 27.73% (~29.5int = 1% SCrit)
+m/5 158.6
Hitpoints 6266
Mana 9076



So, with the Illumination talent, that means Fol (Flash of Light) and HL (Holy Light) only cost, on average, 72.27% of their listed mana cost. Spells must be above L20 to recieve full +healing, to my knowledge. Blessing of Light adds 115hp to FoL, and 400 to HL as long as they both recieve full value from level based spell reduction.

FoL1: 35 mana, 1.5s cast. ~75 hp
HL4: 190 mana, 2.5s cast. ~373 hp

These two spells are the lowest spells a L60 pally can use and maintain +Heal. Here are their mana cost per 5 seconds, before adjustments.

FoL1: 35/3*10= 116.7 mana per 5 seconds spent.
HL4: 190*2= 380 mana per 5 seconds spent.

Up above, you can see said pally would have 158.6 m/5 gain. You can already see the problem with FoL1. Now lets add in our crit modifier.

FoL1: 116.7*.7227= 84.315 m/5 spent chaincasting FoL1.
HL4: 380*.7227= 274.626 m/5 spent chaincasting HL4.

And now we add in our average mana gain over time, 158.6 m/5.

FoL1: -84.315+158.6= 74.285 mana gained per 5 sec.
HL4: -274.626+158.6= 116.026 mana spent per 5 sec.

So, if a dwarf paladin with the above gear, talents, and buffs, starts chaincasting HL4, it will take said paladin 6.5 minutes to empty out their 9k mana pool. Conversely, if said paladin is perfectly chaincasting FoL1, they will gain 891.42 mana per minute, and will take a little over 10 minutes to fill their entire 9k mana pool back up from zero. With no consumable use. (As a side note, when i quit, with full T2 and bwl/zg gear, FoL was effectively drainless.i had approximately 15% chance to crit, and much less m/5 though.)

Before you say these cheap heals are entirely worthless, here's the math on the +heal and BoL. also, 27% of the time they will heal for 1.5x the value below, but i'm not going to model that since it's too random to depend on.

FoL1: ~75+115+(1.5/3.5*790)= ~526 hp per FoL1.
HL4: ~373+400+(2.5/3.5*790)= ~1337(!) hp per HL4.

So, a note on the effectiveness of said heal numbers. Individually, chaincasting FoL will not save your tank from spike damage. With 4-5 paladins doing it, that's 2000-2500 hp per 1.5s. It might save the tank from sudden spike damage, but it's fairly random. I personally think on high spike damage fights, FoL1 isn't worth very much, even with several paladins doing it. Too much hp needs to be replaced quickly for FoL to do it. Now, on a fight with steady damage on the tank, FoL1 is milk and honey. Conversely, HL4, or a higher rank of HL, is wonderful on spike damage fights. With it's substantially higher impact value, your tank has a much better chance of surviving if your heal lands at the right time.


That's not really the best part, though. The value of Illumination scales, not only with the amount of SCrit you have, but also with the cost of the heal spell you're casting. And it doesn't scale linearly.

FoL1: 116.7*.01= 1.167 m/5 saved per 1% SCrit
HL4: 380*.01= 3.8 m/5 saved per 1% SCrit
HL9: 1320*.01= 13.2 m/5 saved per 1% SCrit

So, as an exercise, here's some numbers to play with. Take HL4, give ourselves 50% crit, and assume we're running nightfin soup, mageblood potions, and brilliant mana oil, with otherwise identical gear to what was being used above.

380*.5-190.6= total cost of HL4
190-190.6= .6 mana gained per 5 seconds, with a 50% crit rate, and 190.6 m/5. At that point, chaincasting HL4 with no lag becomes drainless, or atleast until your consumable buffs run out. HL9 looks like this with the same stats.

660-190.6= 469.4 m/5 lost. Granted, this is a significantly higher crit rate than what we're close to right now. But the m/5 is likely to be attainable in Naxx, and will probably be a much higher value, actually.

For the hell of it, here's how HL9 looks with 75% crit rate.

1320*.25= 330 mana spent per 5s on average. when pallies get up to 75% crit rate, i don't think 330 m/5 is at all unreasonable to expect.




So, what's so bad about this? Horde don't have it. If things proceed the way they are for a few expansions, paladins will be able to chain-cast max rank heals endlessly, and alliance healing on endurance fights becomes a complete joke. Horde certainly don't have that option. Druids and Priests get static spell cost reductions in talents, and get scaling spirit regen, but no other healer gets scaling spell cost reduction.


The problem is, in my opinion, multi-part.

The root of it is in the way the skill system in WoW is designed. When WoW first came out, you may remember MC wasn't effectively itemized. Drops were a hodgepodge of T1 and T2, and spells, effectively, didn't scale at all with items. Physical damage characters progressed, with new weapon drops and higher attack power, but casters only got a tiny percent higher cance to crit. The skill/spell system in WoW is designed for a level-based progression system, i.e. reach a new level and gain a new spell. Jump up to today. Now people have reached max level, and have almost solely item-based progression, and the skill/spell system is beginning to show signs of problems, though it's not really broken yet. As stat bonuses from items becomes closer and closer to n, the way several things scale begins to break. Pally spells become free. Give rogues 95% hit rate dual wielding, and 70% base crit rate. Every backstab becomes a crit, Eviscerate becomes completely laughable, and the item-generation formula ceases to function correctly. Give mages +5000 spelldamage, and not only does fireblast cease to function as spike damage in a casting sequence, but spellcrit becomes hideously awesome as a means of increasing damage, and also, whatever spell being cast does not significantly matter, as long as it gets full level based bonus to spell damage, so Evocation and total intellect become entirely worthless. For warriors, given100% crit rate, Impale and Flurry become the two best DPS talents out there, while Deep Wounds ceases to function with weapons over 3.0 speed due to constant refreshes before it gets a chance to tick.

So why am I talking about paladins? Well, for one, I know their mechanics. Number two, with their specific scaling mechanics and already low spell cost, they will break, as a class, sooner than others (meaning that choices made by players stop actually affecting the outcome.) To start out with, Paladin intellect is worth 1.21x times as much as Shaman intellect, due to the +10%int talent and kings. Hastening things along, is a paladin's low spellcrit modifier, at ~29.5int per SCrit. I don't know what shaman get, but i know mages are 59.5int per SCrit, and i assume shaman are higher than paladins, with blizzard assuming pallies will have lower intellect. Which is true for leveling up, but not true after item progression has been let loose. Lastly, Illumination is, although i hate to say it, is a broken scaling mechanic. If you took away illumination, and only illumination, with stats=n, paladin mana consumption once again comes back into the picture. Pallies still would have an advantage over Shaman, with the way int scales, but as a class, they are not entirely broken.

The only real way to fix this, that i can think of, is basically to rebuild the WoW character system with item progression in mind. Lock down things like crit rates so that they don't scale out of control, design the skill system so that the relevant stats affect the the spell being cast. I.e., both damage and cost of frostbolt become modifiers of how much Int the mage casting it has. Healing strategy based around low-rank heals changes fundamentally.

However, rebuilding much of WoW's character engine isn't going to be easy or quick. As far as band-aid fixes go for paladins, nerfing the shit out of Illumination, or drastically scaling back the +SCrit in the game would be one place to start, although Blizzard would piss off even more the already irate paladin player base. For other casters however, scaling up the mana-cost of spells, to reflect the ever-increasing +damage stat, is a bad solution. Mainly because new skill books are designed to be available to everyone, in 20m instances, and not everyone has ridiculous +damage. Can you imagine the shit-storm raised by non-raiders from putting new spell ranks in 40m content only? Or, sticking new ranks within 20m content, but making it only worthwhile to use from an efficiency standpoint if said caster already has +3000 dmg, that's only available from bleeding edge 40m raids?


So anyway, yeah. Maybe someone else can rap about how +healing scales with healing wave's speed increase, and how possibly overpowered the new +18% to HP recieved talent that the tank might get, for shamans. Think that's all i got for right now though.

Oy, that's a type of scaling I hadn't even thought of before. But yes, scaling has worried me for a long time (I think my only recent rant on the topic is here). Everyone can see that the system wasn't built from the ground up with the long-term in mind.

Briefly since it's late, some of it is already visible. Cone of Cold gets eclipsed more and more as +dmg gets higher. Also, when I chaincast Rank 5 Frostbolt in a raid setting, my mana goes up.

The most broken thing imo with current mechanics is how +heal doesnt scale with spell ranks (above lvl 20). If you look at the spells without any +heal you see how each rank follows a nice curve and gets more and more manaefficient, once you start getting even a moderate amount of +heal the opposite is true the highest rank spell is the least manaefficient.

Can't wait to see my new ranks of level 70 spells and realize how much of a downgrade they are and that I never will use them.

You know, I have been trying to explain this to people for some time. Purely from a ilvl perspective, caster damage actually scales faster than melee damage. Obviously there are various mana and aggro issues that prevent mages and warlocks from being top on the dps charts, and different issues in PVP. But my point is that when people complain that melee damage is scaling faster than magic damage, they are wrong because the situation is actually the reverse.

iirc the cap of +healing on a given spell is that you can at most double the amount healed of said spell.
I dont know how that impacts on your math. The major point of your post (mana effectivness of pally heals scales with gear) still remains correct though.

I read something from a blizzard offical that spellcrit from int is the same (59,5 per % crit) for all classes at level 60.
I think the post is deleted though. He first posted that mages need those 59.5 int per crit and later posted another reply that its the same value for all classes.

Frankly, there are worse scaling problems than this one. Blizzard is going to have to change all sorts of things when the expansion hits, or shortly thereafter, let alone after multiple expansions are on the books.

I made a huge post about scaling on the FoH boards back in the fall of 2005. It's largely about mixing additive and multiplicative factors, or multiplicative factors with different magnitudes. I dug up the old post, and here are a couple of excerpts:
I then discussed specific cases. One big one, which hopefully will be remedied in the rogue revamp in 1.12, is the fact that high enough rogue AP/weapon DPS will actually make finishers obsolete entirely for dagger rogues. There's a theoretical point at which Backstab will be a superior use of energy to any S&D even, and at that point, the rogue class breaks. This whole class that was designed around building up combo points will now just be about hitting your "setup" move over and over.

Probably 50% of a tank's aggro generation comes from weapon damage; the other 50% is from fixed-threat bonus values associated with their abilities (unlike druid and shaman tanking, which uses a straight multiplier on threat). So if the raid's DPS increases by 20% across the board, everyone else is generating 20% more aggro but the tank is generating only 10% more.

Stuff like new skill books from AQ20 are a bandaid to this, but the expansion is when the real overhaul needs to come.

I agree that item-scaling and performance-scaling is a problem, and I hope they seek to correct it (no small task) in the expansion.


Regarding healer-scaling, I think the real problem is the a lack of need in raids for sustained healer throughput. Lets look at +healing. This stat increases throughput -AND- efficiency, in the same way increased spell-ranks always gave a lot more throughput, and a little more efficiency. When +healing first became ubiquitous on healing gear in the tier1 revision and DM release (I don't remember which came first), everyone loved it because previously, healers were all about Spirit. Healing was primarily about managing your regen and using the 5-sec rule well. With the simultaneous introduction of Mp5 gear, managing regen became notably less important.

But lets look where are today. Now, nobody cares about the sustained throughput (not burst throughput), we care about the efficiency. Because the unforeseen problem of +healing is that if you take two different ranks of a spell, the throughput increases linearly at the same rate for both, but the efficiency does not. The efficiency scales linearly in both cases, but the lower ranks of the spell scale at a higher rate. Take two sample spells, 3.0 sec cast. Rank 1 is 600 health for 200 mana (3.0 hpm). Rank 2 is 1750 for 500 mana (3.5 hpm). Rank 1 efficiency scales at 0.005 hpm/+heal and Rank 2 scales at 0.003 hpm/+heal. Rank 1 scales about twice as fast.

So what’s a potential solution? Make sustained throughput more important in raids. There are very very few fights in the game where this is important (Vael and Huhuran come to mind), and in the former, the fight basically breaks most of rules of the game with mana/energy/rage. Currently, healing is about efficiency and dealing with spikes. We need more fights that are about pumping massive hitpoints into a target over a period of time longer than 2 seconds after he gets MSed / ISed, to make the "free heals" unusable due to their low output. Consider this raid boss:

He has an ability that he uses every 60 seconds.
This ability buffs his current target with +50,000 of max hp, and then boss becomes enraged and starts swinging for 15,000 DPS, with a 1.0 attack speed. This ability lasts for 12 seconds. So, you'd have to deal with about 4-5 of these during the course of the 5 minute fight.

Now granted this is a unique encounter like Vael where certain limitations of a character are ignored, but it’s an example I came up with quickly that would test a raid's throughput. A much simpler version would just be to increase a boss's DPS against the tank but decrease the spike damage and make it more linear. Tanks have a ton of HP these days, but the raid designers are just putting in boss abilities that can instantly hit for such a huge portion of a tanks life that its imperitive to keep a tank topped off all the time or risk instant death. Lets have some bosses where you can comfortably leave a tank at 50-75% of his life, so you can be sure you aren't wasting your time casting max-rank HTs, gheals, HWs, and the like. This would also remove the requirement of Titans etc and lessen the BoK imbalance. As always, more HP would give you a better buffer and make healing easier when you learned, but eventually you could wean away from them as you got more comfortable instead of when your tank naturally gets 1200 more hp. I'm sure the creative minds at blizzard can come up with good raid bosses to take advantage of this stuff; the whole point is that we can lleviate the broken-ness of low-rank "free heals" by stressing the importance of throughput in contemporary raiding.

Wow, that is an issue with scaling of Paladin's heals. It is nice a pally can chain cast, but it seems in the raids I have been in a lot of the heals are overheals with pallys (around 30%). I wonder what the average Shaman overheal % is.

However, like Gurg said, their are other issues that Blizzard will have to address to keep the game balanced, especially Warrior threat (which affects Horde more).

Eh, better to have a small amount of overhealing then to have your healers scared of it and cancel their heals for fear of another healer landing one before they cast. :)

Edit: I'm not saying healers should just spam heals no matter what. Better safe than sorry, though.

Well, if you think tanks have a lot HP now, wait until next patch. 7k unbuffed on a tank is already a reality; with the new 10% hp talent and the plethora of stamina on Dreadnaught, well over 8k unbuffed should be a real possibility in the coming months.

Hmm.

1 +spelldmg gives me 0.814/2.5 = 0.326 DPS. (assuming my miss rate is rougly similar to my crit rate).
5/3 AtkPwr (the corresponding item budget value, I believe) gives 1.6/14 = 0.114 DPS. For, let's say, a Rogue, this amounts to an overall increase of 0.114*1.75*1.3=0.26 white DPS. Also, since the Rogue's crit rate is much higher than his miss rate, this number will be even bigger.

But, the damge I gain to my Frostbolt is all the damage I get, because that's all I do. In fact, even I burned mana to work in, say, Fireblast, I don't get any additional benefit, because the +dmg is scaled to down specifically to prevent this.

But for the Rogues, the AP contributes to his specials in a way that adds, without any loss of efficiency, on top of the white DPS increase. I really don't know much about Rogue theorycraft, but for a coarse approximation, if you're Sinister Striking every 4 seconds, those 1.6 AP are giving you 1.6*2.4/14*1/4 = 0.069 DPS. This is already enough to push your gain over ours, and there are significant factors I'm not yet considering (Crits, Sword Spec, Windfury, and assorted other things that further multiply your AP gain).

Fixed. The conclusion is still correct.

It was imprecise for me to say "melee" as honestly, I probably know even less about rogues than you do. But for warriors you get much less AP multiplication because the far fewer number of specials and crits. Even for rogues, I think if you apply even a modest average armor reduction of say, 40%, and then apply a realistic resistance rate to casters, you would find different results. And while rogues have a higher normal crit rate, spell crit is cheaper than melee crit on items.

And let's not pretend that casters don't have +damage multipliers as well. Master Demonoligist, Arcane Power, Power Infusion, Improved Shadow Bolt, Netherwind Focus, etc. - these all have a similar multiplicative effect to Sword Spec, Windfury, and all that jazz.

Although, honestly, as I said above, the whole point is moot as casters are hamstrung by other factors in every situation. I am merely trying to realistically examine whether melee damage actually does scale faster on items.

The only reason to not overheal is to conserve mana, and since paladin's literally don't have to worry about mana, any amount of overhealing is fine. I know a couple of paladins who actually have emergency monitor healing macros and have some crazy automation mod or something that basically means when they get put on patch healing detail, they turn on their mods and alt tab. It's not that they are bad players, it's just that they literally don't need to do anything else, and have gotten bored with it.

To reference, one of the paladins uses mid 700's of +healing

Perhaps one solution is to use the expansion and the new suite of level 61-70 skills to address the weakest points. Add an "epic" form of Eviscerate that scales properly. Add a passive ability for higher spell ranks that reduces threat so you can simultaneously ease down the aggro limits while giving people a reason to use them. Brainstorm some new mechanisms for healing and encounter design that downplay spamming a button. What if you gave healers special heals that also carry buffs which need to be selectively applied in fights? One type of heal could, for example, give the character a +50% chance to resist curses for 6 seconds, while another could be put on a long cooldown and give a thunder-clap-esque effect to anyone who attacks the target. These would have to be selectively applied to protect players in addition to just topping off their health bars.

In other words, instead of just re-scaling the numbers so that you can have yet another breaking point set beyond the player's gasp, perhaps Blizzard could just tweak the gameplay as the numbers get bigger and bigger so that it's not just a numeric reshuffle, but also produces new ways to play? If healing is getting trivialized as numbers get bigger, give healers something to actually do.

Or is that too gimmicky?

Aye, thooloo, one possible solution for blizzard would be to introduce new, scaling ranks, and leave the current game the way it already is. I think this would fix some things, like rogue Eviscerate, but some of the mechanics for characters with mana bars are so totally fucked i don't really see how tacking on a new, altered spell level would fix anything.

Your thought on healing, though, I think is somewhat of a different subject. Overall healing strategy in WoW is interesting, and you're kind of touching on it. Currently, the majority of raid healing works reactively. That is, the tank loses hp, and then the healer repairs it. There are very few proactive abilities for healers, namely PW:S and regen spells, but both are nearly irrelevant for a tank in a raid setting. What you're talking about adding would be proactive damage prevention abilities, and in my opinion has the possibility, if implemented correctly, to change the healing game from a reactive twitchfest, (or a mindless spamfest if you have enough m/5) to a more proactive style of gameplay encouraging mental activity instead of repressing it.

Which would be really cool if WoW would do, actually.


As one other thing, here's an idea i had on how blizzard might be able to salvage the illumination talent would be to simply give it a 10s cooldown on returning mana or some such. Still useful, but the maximum mana saved goes down significantly.

You mean when they release the expansion? It seems like as good a time as any to rewrite much of the game.

You mean the time when they have a million new zones full of bugs and unbalanced encounters/rewards, many of which are probably not even fully completed? That is when they'll find the time to rebalance fundamental class concepts? Experience shows that Blizzard (and other MMO publishers) are far better at balancing shit when their content is mostly established and working, due to the basic principles of resource allocation. Class balance has a lower priority than simply making stuff work.

My point is that you are complaining that, "in 6 months, all their work has to be redone."

Well, no surprise there. The entire game will have to be redone in six months; priests are not special in that regard.

I think he foresees something more cyclical, where the "redoing everything" phase begins after the expansion launches, and all the class-design theorycraft has to be redone in light of the actual content. I don't think we necessarily want another round of massive class overhauls because the new 9-tier talent trees broke everything, do we?

I'm not sure I agree with his premise, though. Blizzard knows a lot more about practical implications of class balance than they did 18 months ago. Or at least, they'd better. They should be in a position to think through the issues now, with the benefit of what they've learned about scaling, raid mechanics, people one-shotting each other in PvP, and so forth. If not, we're in for a messy ride.

The problem is, some of the needed changes are drastic.

They need to change how hit, skill, and defense work in the combat equations. That may screw up itemization, talents, and class balance, but it needs to be done.

Right now, with 310 skill, glancing blows effectively do not exist. With 440 defense, crits do not exist. With various amounts of +hit depending on who you are, misses don't exist. That's a problem. Unless they redo PvE combat heavily, at level 70, people with 310 skill now will have 360 skill and still never lose damage to glances. People will have 490 defense and still never get crit by 73's. And so forth. And that's with current, pre-expansion itemization. It leaves no room to grow in any of those areas.

Well, they don't necessarily need to change the equations. The ease with which one can max out hit, weapon skill, and defense can easily be counterbalanced by game mechanics - Unbalancing Strike is fairly elementary example of this, but it suggests the encounter designers are aware of the problem.

To speculate for a second, one could easily imagine mobs with somehwat low base damage but much higher crit rates, or mobs with higher than standard defense - in fact, I think C'thun is either level 64 or has a very high defense skill.

The tools for adjusting those problems already exist, it is simply a matter of the designers taking advantage of them.

I'm honestly a bit surprised we HAVEN'T seen a boss with higher than base crit rate, or other abilities designed to mess with the current itemization paradigms: say, a boss with Imp Overpower or a similar ability.

That said, since healer theorycraft is a lot fuzzier than DPS theorycraft, I think they'll have to revisit the healing classes at some point to deal with these issues. One example of how healer theorycraft is different: what possible metric do you use to measure the effectiveness of a healer? Because it's sure as hell not sustained HPS, especially when the main danger a current raid boss presents is spike damage, and the other possible metrics are either difficult to nail down due to the presence of 14 other healers or difficult to quantify. This could be solved to an extent with a fight like Sirloin is suggesting: best example I can think of is Rajaxx, where the NPC tank has a lot more HP than a warrior but takes a good bit more damage.