Offense vs Defense

[Bolty]

People hear me quote the whole "offense is outpacing defense" thing and wonder what the hell I'm talking about. It's something you will have noticed in WoW if you played a "defensive" (aka healer) class over the past 2 years.

A Recent Post on the WoW Priest forums is one of those gold nuggets in a sea of trash, as the WoW boards tend to be. It expounds in more detail what is meant by "offense is outpacing defense" in World of Warcraft. Since such posts will eventually disappear off the forums, I'd like to capture Xusilak's article here. He writes in relation to spell power only, but there are equivalents with ranged and melee classes as well and the way their gear scales them ever upwards in damage while healer gear is far more linear in growth. This, in effect, makes characters like my healing Priest less and less useful as time goes on, and nobody likes being marginalized.

A phenomenon I frequently notice on these forums (and in-game) is that many people simply do not understand the concept of proportional scaling. They see +damage on one item, +healing on another, and note that the +healing is almost twice as high, then think, "Wow, healing scales really well." This reasoning is because they only understand rates of power. They believe that because Ability A gains X DPS per +damage, and Ability B gains X HPS per +healing, everything is fine, because after all, healers are getting better returns per item level point, right?

Honestly, Blizzard perpetuates this line of reasoning with the way they have built their spell progression system: their system is completely based off addition of rates. For the most part, adding X spell power will increase all spells by Y power per second (PPS). It does not, in any way, take into account the fact that Spell A might have a lower base PPS than Spell B. What happens if you then proceed to take two hypothetical spells, Spell A of 10 PPS and Spell B of 100 PPS, then using Blizzard's system of rate addition, add 10 PPS to both spells? Spell A becomes 20 PPS, and Spell B becomes 110 PPS. This is all fine, because both gained the same rate, right? Wrong.

Let's discuss proportions for a moment. Take two classes, Healer and Nuker. Healer is designed to be able to heal about three times as much damage per second as Nuker can deal. This is so that Healer can have a significant effect as a support class in group PvP, enabling group strategies, and also giving Healer a bit of a bonus for playing a support class instead of a damage class. However, this is not about class design philosophies; it's about proportions. So, take Healer and Nuker. At base level, Healer is healing three times as much per second as Nuker is dealing damage. This ratio is decided to be the ideal point for class balance. Still, note what has happened here: I've established a ratio of power between two classes' abilities. This is also known as proportion. Nuker's power is 0.333x Healer's power. This is important, because if this value ever changes, then class balance has changed.

Now, currently, Nuker has absolutely no way to defeat Healer in one-on-one combat. So, let's give him some precisely tuned abillities that, if used correctly, will allow Nuker to defeat Healer directly about 50% of the time. Note that these abilities are tuned for this 0.333x proportion of power. If this proportion increases, then these abilities will be too powerful. Consider that for a moment.

Okay, now that we've set up our two classes and they can frolic in joy while PvPing in a balanced setting, let's add some gear into the equation. We need gear that improves their power output, right? So let's use Blizzard's system of rate addition, since it's already there. Let's say that, right now, Healer has 300 PPS and Nuker has 100 PPS. Let's say that our new gear stat, Spell Power, increases PPS by 1 per point. Okay, we've got some gear. Now let's outfit Healer and Nuker in some nice high end gear of 200 Spell Power. Healer now has 500 PPS and Nuker has 300 PPS. Looks good. Except, wait a minute, Nuker's proportion of power to Healer is now 0.6x! Nuker will now win every single encounter with Healer, because his abilities that were tuned to confront Healer when his power was 0.333x are now grossly overpowered! Not only this, but our entire design philosophy for the two classes is broken. Nuker can now practically ignore the presence of Healer in group PvP. Using the basic Calculus concept of a limit, we find that as Spell Power approaches infinity, the ratio between Nuker and Healer approaches 1:1. If you have trouble understanding this, then just keep imagining the ratio between their power at higher and higher levels of gear; at 500, at 1000, and so on. As it keeps increasing, Nuker becomes closer and closer to being able to equal Healer's healing rate. Clearly, if the proportion ever became 0.9x or more, there would be serious balance errors.

Consider the previous paragraph carefully for a moment. It is crucial to understanding overall game balance. If relative power changes, then game balance has changed.

Now, let's go back to World of Warcraft. Nuker and Healer directly parallel Mage and Priest. The proportion of Healing:Damage is approximately 3:1 at the base level. This proportion changes as gear increases. For the same reason as before, this is a balance failure. Blizzard has applied many attempted fixes to their Spell Power system to try to correct the more egregious problems resulting from differing base rates among spells, but as demonstrated, this system is fundamentally flawed when dealing with abilities that have varying base rates. The only way to fix it is to give every ability its own, unique coefficient based off its power in relation to some baseline.

By far, the easiest way to construct a Spell Power progression system is to make it multiplicative. Multiplicative bonuses aren't even foreign to World of Warcraft; Critical Strike bonuses are multiplicative. A simple linear rating value, much like critical strike, that multiplied your power output would be perfect for preserving varying rates of power. This is known as proportional scaling. When people talk about scaling, what they really mean is proportional scaling. Scaling is the concept of taking an object and then multiplying it up to a different size. When someone is talking about different scales, you don't hear about additions or rates; you hear about things like "1:5200th scale." Scaling, in reality, always refers to proportions.

With that in mind, the next time Healer tells you "my heals scale terribly" and you laugh at him because you saw an item that had almost twice as much +healing as your +damage items, remember this post. He is talking about proportional scaling.

On page three of the thread, Niimue chimes in with a more "tldr" (too long, didn't read) version:

Very well thought out post, although somewhat lengthy.

A more consice way to state the OP's point is that the spells are designed initially at around a 3:1 heal:dmage ratio. Additional gear scales around 3:2 heal:damage, so that damage goes up much faster than healing.

Here is an example:
Priest GHeal 5 (no gear, untalented) heals for around 2080
Mage Fireball 12 (no gear, untalented) hits for around 680
This around a 3:1 ratio of healing to damage

Priest 8/8 Tier 3 (no enchants) adds +483 healing
Mage 8/8 Tier 3 (no enchants) adds +313 healing
This is around a 3:2 ratio of healing to damge

With just the Tier 3 gear, the mage now hits for a non-crit average of around 1000, and the priest heals for a non-crit average of 2500. What used to be a 3:1 heal:damage ratio is now a 2.5:1 ratio. It continues to get worse as gear and enchants are added until it levels off to the 1.5:1 ratio of the added gear, and not the original spells.

With equivalent gear, damage dealers outpace healers.

Xusilak's post isn't really a balance post, though. He's not so much complaining about balance between offense and defense; rather, he's pointing out how a scaling system makes more sense, is cleaner, and benefits all caster classes. Two more quotes explain this:

Yeah, Stamina certainly came to mind many times while writing that post, but that's a seperate issue entirely, and supposedly one that is being addressed in The Burning Crusade.

I've considered scaling systems like what you propose before, but realistically, just switching to a pure multiplication system is so clean and simple that I really don't understand why Blizzard went with this addition-based system at all. Back in the initial WoW testing phases, I noticed the +damage mechanic and immediately went, "Wow. That would be really broken for small, fast spells." Cue the numerous band-aids, such as cast time-based coefficients and level penalties. As time goes on, the system just gets more and more convoluted. I'd really like to hear a justification from Blizzard for why the choice was made for an addition-based system, when a multiplication-based one is so much cleaner. With the newly devised rating system they have, it wouldn't even have the problem of scaling with level; Spell Power rating could be less of a multiplier as you gain levels.

To put it simply, at base, a Mage deals damage at a certain percentage of the rate at which a Priest can heal. This percentage must stay constant at all gear levels, or class balance has changed. Right now, the percentage is increasing in the Mage's favor as both the Mage and the Priest gear up.

That isn't even really the point of my post, though. The purpose was to try to inform people about the exact nature of proportions vs additions of rates, and the mathematical consequences thereof. That's why it's so long. I'm simply tired of hearing things like "Heal scaling is fine, L2P noob priest, theres to much +healing already." Complete with incomprehensible English and other such gibberings.

I couldn't really agree more with Xusilak, but I'm wondering if any math wizards can find flaws with this system. There are some spells that Blizzard clearly doesn't want to scale 100%, such as Mind Flay, but I'm sure that can be handled even with the more simplified Spell Power system. Also, I'm not too big a fan of the concept of a single stat that boosts a Priest's damage AND healing at a similar rate that that it would a pure DPS class. This would make Priests overpowered. Cleoboltra's ability to heal at gargantuan rates cannot be coupled with the ability to nuke or you have some serious imba. Even still, there could be a healing spell power rating and a damage spell power rating, keeping the 3:1 ratio.

-Bolty

[lemekim]

The big issue here probably is scaling. Blizzard has been a long time opponent of multiplicative bonuses. One reason, for example, would be that if your gear increases your damage by 100% at 60, it will also increase your damage by 100% at 70. Author takes Critical Strike as an example of multiplicative damage bonus, yet Blizzard already implemented a system to reduce effectiveness of low level gear at higher levels.

Now, if we do decide to change +damage system from +spell damage to %damage, adjustable by level, such as the current rating systems, I don't see much difference between these two systems (well, there are some). And honestly, I'd rather prefer static +damage and at least do the same damage at 70 as I do at 60, rather than actually losing damage every time you level up if you don't buy a new rank of the spell, because %damage gear scaled back on me.

Now, a much simpler solution would be to make the cost of +healing lower, to the tune of 3:1 ratio with spell damage. Also, Stamina cost would have to be decreased even further - because greater healing power would result in need for higher DPS bosses, resulting in bigger hits taken by tanks, resulting in too many situations where meager scaling of health pool would not be enough to prevent 1-2 shots.

[Drasca]

I couldn't really agree more with Xusilak, but I'm wondering if any math wizards can find flaws with this system. There are some spells that Blizzard clearly doesn't want to scale 100%, such as Mind Flay, but I'm sure that can be handled even with the more simplified Spell Power system.

Sure, if all else remained equal that'd work. Except, it doesn't, does it? You've got convoluted health/mana efficiency ratios, instants, cooldowns, crazy resistence checks.

While I do agree offense outpaces defense as equipment scales, the system in whole is messy and complicated enough that one answer just won't do the trick. That might work for fireball and greater heal. PoM Pyro x5? Nuh uh. Cleanse, purge, kill?

You'd need some revolutionary (for wow) defensive abilities to close the spike gap. . . and given the current boss philosophy of 'do more damage requiring more equipment', that's just probably not going to happen.

[Concillian]

Now, if we do decide to change +damage system from +spell damage to %damage, adjustable by level, such as the current rating systems, I don't see much difference between these two systems (well, there are some). And honestly, I'd rather prefer static +damage and at least do the same damage at 70 as I do at 60, rather than actually losing damage every time you level up if you don't buy a new rank of the spell, because %damage gear scaled back on me.

"Losing damage" as you level up is a little bit of a misconception with the 'rating' based system. You still do the same amount of damage to the same opponent as when you were 60. It decreases only when your opponent's level increases, which makes a lot of sense, really.

This was true for many stats before the rating system was implemented. Armor, for example. If you show at something like 45% DR at level 60, you'll mitigate very significantly more than that if you start running Wailing Caverns or The Deadmines.

The rating system is a necessary fix to bring the +x% modifiers back into line with the normal stat bonuses like +agility, armor, etc... that already scaled in exactly the same way as the 'ratings'. I think the rating system is very clean, it just needs a page in the spellbook or character sheet that mentions what rating = what % at various opponent levels.

[lemekim]

"Losing damage" as you level up is a little bit of a misconception with the 'rating' based system. You still do the same amount of damage to the same opponent as when you were 60. It decreases only when your opponent's level increases, which makes a lot of sense, really.

This was true for many stats before the rating system was implemented. Armor, for example. If you show at something like 45% DR at level 60, you'll mitigate very significantly more than that if you start running Wailing Caverns or The Deadmines.

The rating system is a necessary fix to bring the +x% modifiers back into line with the normal stat bonuses like +agility, armor, etc... that already scaled in exactly the same way as the 'ratings'. I think the rating system is very clean, it just needs a page in the spellbook or character sheet that mentions what rating = what % at various opponent levels.

Yes, but using the same gear, with straight +damage I will do, say, 500 DPS against level 60 mobs at 60 and 500 DPS against level 70 mobs at 70, but with scaling ratings, against level 60 mobs I might do 500 DPS at 60, but it might be only 400 DPS against level 70 mobs at 70, using the same gear and spells.

I understand your point, but when I am leveling, I only care about fighting mobs at or around my level. And against those mobs, my DPS is actually decreasing, unless I upgrade my gear and spells.

Ratings are needed for balance for a specific reason - these "ratings" are applied to percentage spells, where you have a certain ceiling to how for you can go - your crit rate can only get as high as 100%, your armor can only mitigate up to 75% of damage, etc. There is no such ceiling for spell damage, hence there is no need for rating. And I prefer it that way.

[Concillian]

I understand your point, but when I am leveling, I only care about fighting mobs at or around my level. And against those mobs, my DPS is actually decreasing, unless I upgrade my gear and spells.

This is exactly how is should be. It is also exactly how it is for melee classes, as mob armor values and agi per % crit increases and such.

If it were on a percentage basis that changed with level, then the percentage would get smaller, but you base would increase with new spells, and you start evening out... again exactly like happens with melee. The system makes sense and fits in with existing mechanics on multiple levels.

[lemekim]

This is exactly how is should be. It is also exactly how it is for melee classes, as mob armor values and agi per % crit increases and such.

If it were on a percentage basis that changed with level, then the percentage would get smaller, but you base would increase with new spells, and you start evening out... again exactly like happens with melee. The system makes sense and fits in with existing mechanics on multiple levels.

If you want look at melee comparison... Let's say a player has top-notch gear at 60 - 95 DPS 2-hander, DPS gear, etc. He probably won't replace any of it before he hits 70, simply because there is no better gear. And yet, his crit rate will go from 30% to 20%, and his hitrate will take a hit as well, so as he levels up, his actual DPS against unarmored mob of his level will be higher at 60 than it is at 70.

Now, crit rate has a hard ceiling, so it is quite understandable why it should change in this case. However, I think in situations that we CAN avoid it, it's silly to make people feel as if their power is decreasing as they level. Mobs already take longer to kill because they have more life and more armor, there is no need to compound it further, especially when all ratings are multiplicative of each other. I guess some people might like it that way, but I am not masochistic enough for that.

As for spell casters, you should probably be aware that right now, damage increase in spells you get by going from 60 to 70 is not nearly enough to cover the 35% drop in effectiveness of rating across the same level gap. I just don't see the need to make things overly complex and change mechanics so drastically when a much simpler solution is available (Reduce cost of +healing).

[Artega]

"Losing damage" as you level up is a little bit of a misconception with the 'rating' based system. You still do the same amount of damage to the same opponent as when you were 60. It decreases only when your opponent's level increases, which makes a lot of sense, really.

This was true for many stats before the rating system was implemented. Armor, for example. If you show at something like 45% DR at level 60, you'll mitigate very significantly more than that if you start running Wailing Caverns or The Deadmines.

The rating system is a necessary fix to bring the +x% modifiers back into line with the normal stat bonuses like +agility, armor, etc... that already scaled in exactly the same way as the 'ratings'. I think the rating system is very clean, it just needs a page in the spellbook or character sheet that mentions what rating = what % at various opponent levels.

From posts I've read in the Blizzard forums, it's possible to simply extrapolate the numbers of the rating system to get values for other levels. I believe they used this method to determine that 25 Resilience is -1%/-2% at Level 60. You could likely do the same for other ratings versus higher or lower levels (+14 Crit is equal to 1.30% at Level 48 and 1.76% at Level 37.)

EDIT: It's also been a very long time since I've done any kind real math, so I could be wrong. I'm just relaying what I've seen in posts at other places:)

[Xame]

\

To put it simply, at base, a Mage deals damage at a certain percentage of the rate at which a Priest can heal. This percentage must stay constant at all gear levels, or class balance has changed. Right now, the percentage is increasing in the Mage's favor as both the Mage and the Priest gear up.

That isn't even really the point of my post, though. The purpose was to try to inform people about the exact nature of proportions vs additions of rates, and the mathematical consequences thereof. That's why it's so long. I'm simply tired of hearing things like "Heal scaling is fine, L2P noob priest, theres to much +healing already." Complete with incomprehensible English and other such gibberings.

I couldn't disagree more with Xusilak. This is looking at two unrelated classes and trying to make some rational discussion about how they should scale.

Let us first look at PvP. In PvP a mage trys to kill a healing target in <15-20 secs. That is because he must use cool downs to get that level of burst. If a mage fails to get a fast kill the 3x healing gear means that any healer will be back at full health. In any fight >15secs the healer's mana pool will outlast the mages and without the cooldowns the mage cann't kill a healer. Without the cool downs, priest could and have just wanded mages to death. Most people who don't play mages equate fast kill with overpowered. But mages with their very low health pools die equally fast to other classes. If we extend the scaling to level X, mages either get crazy nukes or priest's healing must become equal or less then mages nukes (ie. if a mage nukes and a priest heals, the mage should make some forward progress on the priest's starting health pool and shields) to produce a coin flip to see who wins.

In PvE, the comparison is also different. A mage needs to convert their mana pool into damage equal to the mobs health or in parties/raids last the fight (note: damage done by the mob is NOT a function of mage design since we don't get hit). Priests need to heal through the damage incoming compared to damage put down by the mob (raid or solo). Thus in mage design the health of the mob verses mage dmg+mana pool is important but in a priest it is a function of incoming damage and healing through it.

Thus mage and priest are an apples/oranges comparison. I would suspect that priest/warrior are balanced together and would make for a more interesting scaling discussion. Both priest/warrior have been scaling too wildly pre-2.0 patch and Blizzard needed to clamp down on both classes to build interesting instances. Warriors with endless rage and priests with never ending mana pools (because of down ranking and crazy +healing gear) makes for some pretty stupid fights (see the never ending group of Ony 5-10 man'ed videos).

If I had to beg Blizzard for one thing, it would be a collection of defensive skills/spells. The mage "spell steal" and the new warlock dot "unstable affliction" are nice starters (but also very offensive). But there should be more for melee and different spell schools. Currently, we only have our health bar and some classes have shields. Thus all we can talk about is dmg + stuns vs health bar + escapes as an offense vs defense discussion. Sadly I dream for more.

[Icebird]

I think the warrior/priest situation is a good example where the relatively gearing of the characters makes all the difference.

I have a warrior mostly equipped in the low-level pvp gear. Weapon-wise he switches between an Iceblade Hacker/Warblade of the Hakkari off-hand, and The Unstoppable Force. I generally prefer dual-wielding against casters for more chances to interrupt their casting.

I've had one-on-one battles which I had no chance of winning, because I simply couldn't do enough damage. A priest can shield themselves every 15 seconds. So my effective damage time was 15 seconds minus time taken to break the shield. Your Tier 3 warrior probably breaks through it seconds. Against a well geared priest I'd be lucky to have 5 seconds of "damage" time. Then they re-shield and can heal themselves back to up full with nothing more complicated than a Renew. Meanwhile even Shadow Word: Pain will kill me eventually.

Chris

[Brista]

As Xame indicates stamina and survivabiity skew the fights

Let's suppose Nuker Class beats Healer class if he gets the healer down a bit so healer has to heal then uses Counterspell and bursts him down the rest of the way before Counterspell expires. There is a window of opportunity between the healer's heal being prevented and the counterspell expiring and a fresh heal being cast

That's the only way he can win, if he can't do sufficient burst in that window of opportunity he loses

Early in the game level 60 Mages simply lost to level 60 Priests. The only way you could not win that fight as a Priest is to let your mana get carelessly low then get counterspelled. Keeping yourself over 75% life and wanding the Mage = win

3:2 is probably fairer than 3:1, at least I think so. 3:1 with pretty slow fights was too much

Please also bear in mind that some of Xusilak's calculations assume the healer is using damage and healing gear. That's probably a fair assumption for the level 60 bg farming but it won't be a reasonable assumption for the Arena matches

I do think that this last month of pvping frenzy has indicated a pvp healing problem and that damage to survivability is too high with level 60 gear and the talents meant for the level 70 end-game

But it's balanced for level 70 and in good level 70 gear players won't die fast and pvp healing will be crucial. Moreover the nature of the Arena fights is much more hardcore (in the Diablo sense). If you die you're out. If someone on your team dies you have probably lost unless you can immediately get a kill in return.

Wait till level 70 Bolty, most 5 man pvp teams will require a minimum of two good healers and healing will win fights.

[Bolty]

Wait till level 70 Bolty, most 5 man pvp teams will require a minimum of two good healers and healing will win fights.

Oh, this I completely agree with - Cleoboltra will be most powerful relative to the other classes in the game as she first hits level 70.

Problem is, as itemization increases, the old issue will crop up again. People will simply die, herself included, so quickly that her healing power becomes irrelevant. I guess it's hard to get across how much it changed over the last 2 years in group PvP environments. I used to be able to keep people alive and force them to have to take me out first to get anywhere. At the end, before the expansion came out, I could be totally ignored by DPSers intent on killing a target because I'd only extend the target's lifetime by a few seconds, tops.

Or to put it more directly in terms of my own character: I had a mixture of blues and purples which would double the damage output of my Holy Fire or Smite spells. I also had a mixture of Tier 2 and Tier 3 epics for pure healing that would increase the power of my max Greater Heal by 50%. That seems a bit off to me.

*shrug*

-Bolty

[Brista]

Oh, this I completely agree with - Cleoboltra will be most powerful relative to the other classes in the game as she first hits level 70.

Problem is, as itemization increases, the old issue will crop up again. People will simply die, herself included, so quickly that her healing power becomes irrelevant. I guess it's hard to get across how much it changed over the last 2 years in group PvP environments. I used to be able to keep people alive and force them to have to take me out first to get anywhere. At the end, before the expansion came out, I could be totally ignored by DPSers intent on killing a target because I'd only extend the target's lifetime by a few seconds, tops.

Or to put it more directly in terms of my own character: I had a mixture of blues and purples which would double the damage output of my Holy Fire or Smite spells. I also had a mixture of Tier 2 and Tier 3 epics for pure healing that would increase the power of my max Greater Heal by 50%. That seems a bit off to me.

*shrug*

-Bolty

Bear in mind though that while itemisation increases the ability of dpsers to kill fast in general it also can increase survivability.

I honestly don't know how the level 70 end game we currently are heading towards is balanced. I suspect it will be possible to keep people up pretty effectively in arenas as long as they take advantage of all the +stamina and +resilience rating gear. If dps fails to win by burst then healing still beats it

I have no doubt though that as new instances with shiny new toys are introduced that this problem will become more acute

I do share your concern. WoW Arena will be much less of a team game if the best tactic is for 5 people to stack as much damage gear as possible and focus whatever comes closest. And when the fights are too short pvp healing is pointless. And that makes it a worse game, less about team play and more about nuking

[Warlock]

If an assist train can kill someone in less than the casting time of a heal spell how fast healing scales isn't going to matter much.

[Artega]

If an assist train can kill someone in less than the casting time of a heal spell how fast healing scales isn't going to matter much.

Organized PvP is a bit like follow the freight train, isn't it?

Choo choo.

[Malakar]

I think the current spell damage/healing coefficient system's main problem is that it is based on cast time, and not on each individual spell.

The good is that it multiplies an added number, so it has the possibility to scale well without affecting higher level base spells in an unintended way. The bad is that it multiplies by generic numbers based on cast time.

Edit: My sig is not directed at this thread.:P

[Drasca]

I think the current spell damage/healing coefficient system's main problem is that it is based on cast time, and not on each individual spell.

Actually, it is per spell, the cast time is merely a rule of thumb. You won't find the cast time % formulae to work well for most spells outside the basic fireball or shadowbolt.