Muphrid, correct me if I am wrong in this:

You would prefer that a system be built where a relatively equal geared and skilled player of class A and class B would do approximately the same amount of relative worth of damage whether they be in T1 or T7?

This system being based around the idea that there should be a general idea of a chomping order of how things should look (Arcane mages should be just above BM hunters, but a tinge below Destro locks, or whatever), and that before encounter specific situations or weirder conditions become involved this general order should be preserved regardless of gear level assuming that the individual makes an assumed set of choices about their gear (assumed to be close to if not the optimal choice).

From there individual gear levels and content levels can vary who will do better based on specific encounter design, slight variations in the item budget ("We want FFB mages to come into their own in T9 content, lets include some extra crit on the gear beyond normal"), or set bonuses ("It's been a while since DPS Warriors had some time in the spot light, lets increase X talent (Obvious that I don't know warriors at all, eh?) via the T8 set bonus).

---

Q u o t e:
Muphrid, correct me if I am wrong in this:

You would prefer that a system be built where a relatively equal geared and skilled player of class A and class B would do approximately the same amount of relative worth of damage whether they be in T1 or T7?

This system being based around the idea that there should be a general idea of a chomping order of how things should look (Arcane mages should be just above BM hunters, but a tinge below Destro locks, or whatever), and that before encounter specific situations or weirder conditions become involved this general order should be preserved regardless of gear level assuming that the individual makes an assumed set of choices about their gear (assumed to be close to if not the optimal choice).

From there individual gear levels and content levels can vary who will do better based on specific encounter design, slight variations in the item budget ("We want FFB mages to come into their own in T9 content, lets include some extra crit on the gear beyond normal"), or set bonuses ("It's been a while since DPS Warriors had some time in the spot light, lets increase X talent (Obvious that I don't know warriors at all, eh?) via the T8 set bonus).

---

Yeah, I'd say that's about right. I'm still wary of the notion that a certain spec should have its time in the spotlight at a particular gear level. To me, it begs the question of why now and not earlier or later, but that can be justified in the face of encounters that are otherwise less friendly to that spec or class than others were in earlier content. Since we can't quantify how one level of content compares to another in terms of favoring or disfavoring certain specs, it's kind of hard to figure how much that should occur.

---

Q u o t e:
Right, and the problem with spell power is that your base damage values mean less and less as time goes on. You maintain a difference (Damage 1 - Damage 2) that stays the same (if coefficients are tuned properly)

That, is, a problem. The fact that the lines curve indicates a shift in worth that results in the wild undulations of classes that we have observed in the past as gear increases.

---

I think this indicates that spell coefficients were not (and in many cases intentionally not) tuned correctly more than a problem with coefficients and spellpower to begin with. It's fairly easy to model mathematically and prepare for. Why they haven't done it in the past, I don't know.

---

Q u o t e:
The buff and nerf cycle, which doesn't need to exist.

---

The buff and nerf cycle takes more things into account. Among which are suitability of class mechanics to given fights. I severely doubt they tune class abilities based solely on theoretical maximum damage. If I had to guess, I would say that this is one of the reasons behind the uneven scaling we've seen before.

---

Q u o t e:
Linearity is the problem, proportionality is the solution.

---

I honestly don't think it is. It's just as predictable and easy to model as proportional increase in spell-damage. It has the added advantage that they can play with base damage and intentionally make spells differ in value at different predicted levels of gear. The problem wasn't that this existed, it's that they didn't give every class/spec at least one spell that scaled similarly.

---

Q u o t e:
And actually procs aren't that big of a problem to balance into a proportional scaled system, since all they are are a percentage chance that another proportionally balanced event happens. This can easily be taken care of in the multiplicative terms of the damage equation to balance them in initially, and then maintain that balance as time goes on.

---

Difficulty is a relative term. Non-linear scaling is simply more difficult to shape than linear scaling. Hotstreak in particular throws a whole lot of non-linearity into things. Crit value becomes more and more valuable as your crit value increases. Offset slightly by the diminishing return of base critical strikes.

I agree that compared to additive spellpower, proportional spellpower would be easier to balance. But it'd also eliminate the ability to have multiple spells that one class has access to differ in value given gear level. It think that is a mechanic that makes playing the class more interesting.

---

Q u o t e:


Yeah, I'd say that's about right. I'm still wary of the notion that a certain spec should have its time in the spotlight at a particular gear level. To me, it begs the question of why now and not earlier or later, but that can be justified in the face of encounters that are otherwise less friendly to that spec or class than others were in earlier content. Since we can't quantify how one level of content compares to another in terms of favoring or disfavoring certain specs, it's kind of hard to figure how much that should occur.

---

Ghostcrawler said (somewhere, I lost the link) that they wanted balance to shift as time goes on.

While I think the current system provides far too much shift in ways that is ugly and damaging, I also would tend to agree that if a Rogue always does 2% more damage than a Mage always does 2% more damage than a Hunter always does... It could be a bit boring.

The only way I can think of to allow for relative shifts without destroying balance is flooding or withholding a given stat at a particular gear level, or providing a specific trinket / set bonus that provides a temporary shift in value.

---

Q u o t e:


I think this indicates that spell coefficients were not (and in many cases intentionally not) tuned correctly more than a problem with coefficients and spellpower to begin with. It's fairly easy to model mathematically and prepare for. Why they haven't done it in the past, I don't know.

---

Even when the coefficients are tuned correctly the proportions shift.

350 damage per cast spell, 3.5 second cast, does 100 DPS.
175 damage per cast spell, 3.5 second cast, does 50 DPS.

Add 100 spellpower.

450 damage per cast at 3.5 second cast, 128.57 DPS
275 damage per cast at 3.5 second cast, 78.57 DPS

100 / 50 = 2
128.57 / 78.57 = 1.63

Still a 50 damage spread, but now a 63% advantage vice a 100% advantage. That 50 damage spread means less at 100 spellpower than it did at 0 spellpower. That is why the lines curve.

---

Q u o t e:
The buff and nerf cycle takes more things into account. Among which are suitability of class mechanics to given fights. I severely doubt they tune class abilities based solely on theoretical maximum damage. If I had to guess, I would say that this is one of the reasons behind the uneven scaling we've seen before.

---

This isn't necessarily about theoretical maximum, since any system can be balanced for any level of play. However, the current system is nigh impossible to balance at *any* level of play, and maintain that balance as you increase gear.

---

Q u o t e:
I honestly don't think it is. It's just as predictable and easy to model as proportional increase in spell-damage. It has the added advantage that they can play with base damage and intentionally make spells differ in value at different predicted levels of gear. The problem wasn't that this existed, it's that they didn't give every class/spec at least one spell that scaled similarly.

---

The problem is that the lines curve in a defined pattern, once they go up, getting them back down again requires buff/nerf. If there were thresholds, where a given spell went up in value from 1000-1250 spell power better than another did, but then the other took over, then that could be acceptable. But, again, is needlessly complicated.

---

Q u o t e:
Difficulty is a relative term. Non-linear scaling is simply more difficult to shape than linear scaling. Hotstreak in particular throws a whole lot of non-linearity into things. Crit value becomes more and more valuable as your crit value increases. Offset slightly by the diminishing return of base critical strikes.

---

Yes and no. The limiting factor of Crit with Hotstreak is that as you add more and more crit, damage and haste become relatively more valuable. It is true (especially with FFB builds) that crit and haste are in close competition for which is more valuable point for point, but crit vs damage is still not a competition, even with hot streak. The effect is mitigated by the worth of other stats.

---

Q u o t e:
I agree that compared to additive spellpower, proportional spellpower would be easier to balance. But it'd also eliminate the ability to have multiple spells that one class has access to differ in value given gear level. It think that is a mechanic that makes playing the class more interesting.

---

It doesn't eliminate it, it forces you to actively make those decisions using set bonuses, trinkets, or stat (non)availability, instead of letting the lines curve.

---

Q u o t e:
Every tanking class is different and every dps class is different. There is no way that scaling will be linear for each class.

---

Strictly speaking, linear scaling is bad. The ratio of linear to linear is ambiguous: it could be constant, increasing, or decreasing, all depending on the base values.

---

Q u o t e:
The buff and nerf cycle takes more things into account. Among which are suitability of class mechanics to given fights. I severely doubt they tune class abilities based solely on theoretical maximum damage. If I had to guess, I would say that this is one of the reasons behind the uneven scaling we've seen before.

---

I see no reason to make this a part of scaling. I feel this should be a part of base values.

---

Q u o t e:
This is a very interesting thread.

Our general philosophy is that we don't want relationships to be too predictable. Much of the fun of improving your character is seeing the effects that different kinds of gear have on different rotations. To use an extreme case, if every piece of armor just said "improves stats by 1%" getting loot would be a lot less fun.

Similarly, we don't want classes to be too similar. It's interesting and fun when you switch classes from say a warrior to a rogue and see how different stats affect you differently.

Scaling is very important to us in the sense that we don't want certain classes or specs to completely fall down at certain gear levels. This has happened in the past. But what those levels are is very important. If mages dramatically overtake warlocks when everyone has 20,000 spellpower, then I just couldn't care less except possibly as an item of trivia. If and when WoW ever reaches numbers of that level, the class mechanics will have changed so much in those years that its just not worth worrying about at this time.

---

Thanks for dropping by, Muphrid definitely generates great discussions.


My big concern with the way that Spell Power (and I'd assume Attack Power aswell, although I've never done physical DPS math) works is that linearity aspect.

I'd tend to agree with your (Blizzards overall) stance that things should vary over time, and definitely agree that things shouldn't be super boring (This gives 2% happy fun bonus, wee), but I'm concerned that we face the same situation that came up in BC, where we watched tables start out un-balanced in 80 Dungeon and T4 content, buff nerf cycle happens, T5 content happens and the nerfed are annoyed they got nerfed, while things stayed in a general balance, T6 starts happening and it gets out of balance again, another buff nerf cycle, same result, end T6 happens and things are all out of whack.

Not that I doubt your (Blizzards overall) prowess and ability to try to make this work. What I doubt is that it *will* work. It either results in buff nerf, or gross imbalance. All of this, I feel, is a product of Spell Power, and the idea of linear vs proportional gain.


I maintain, all other variables ignored, the highest valued actor in a given encounter should be decided by encounter specific mechanics, and active selection from above on who should be doing well in a given gear level. Not by whose M/R ratio is the best.

---

Q u o t e:
This is a very interesting thread.

Our general philosophy is that we don't want relationships to be too predictable. Much of the fun of improving your character is seeing the effects that different kinds of gear have on different rotations. To use an extreme case, if every piece of armor just said "improves stats by 1%" getting loot would be a lot less fun.

Similarly, we don't want classes to be too similar. It's interesting and fun when you switch classes from say a warrior to a rogue and see how different stats affect you differently.

Scaling is very important to us in the sense that we don't want certain classes or specs to completely fall down at certain gear levels. This has happened in the past. But what those levels are is very important. If mages dramatically overtake warlocks when everyone has 20,000 spellpower, then I just couldn't care less except possibly as an item of trivia. If and when WoW ever reaches numbers of that level, the class mechanics will have changed so much in those years that its just not worth worrying about at this time.

---

All well and good, but in the past the scaling has been such that magic based classes scale worse than physical at the level cap.

---

Q u o t e:
Even when the coefficients are tuned correctly the proportions shift.

350 damage per cast spell, 3.5 second cast, does 100 DPS.
175 damage per cast spell, 3.5 second cast, does 50 DPS.

Add 100 spellpower.

450 damage per cast at 3.5 second cast, 128.57 DPS
275 damage per cast at 3.5 second cast, 78.57 DPS

100 / 50 = 2
128.57 / 78.57 = 1.63

Still a 50 damage spread, but now a 63% advantage vice a 100% advantage. That 50 damage spread means less at 100 spellpower than it did at 0 spellpower. That is why the lines curve.

---

And in each case, it was a 50 dps increase. Linear increase is not multiplicative increase. But that's not a problem. It's a "curve" because you're measuring percentage improvement. Absolute dps is linear with respect to spellpower.

It works the same for all of the other stats. A 1% increase in crit adds to dps linearly but iterative increases in crit percentage increase total damage by less and less.

Again, I must wonder why you think this presents a problem.

---

Q u o t e:
This isn't necessarily about theoretical maximum, since any system can be balanced for any level of play. However, the current system is nigh impossible to balance at *any* level of play, and maintain that balance as you increase gear

---

I'm not sure why you think this. It's more complicated than straight percentage improvements based on base damage, but hardly a mathematical unified field theory.

---

Q u o t e:
It doesn't eliminate it, it forces you to actively make those decisions using set bonuses, trinkets, or stat (non)availability, instead of letting the lines curve.

---

But that eliminates the challenge of it. For instance, take FFB and FB. If the two spells scaled differently such that FFB specs become superior at T8 while FB spec was superior at T7, then all the TC'ers would have a job. They'd (like me) waste time at work not working and instead doing math on which spec is better.

A trinket that says "Adds an additional 5% for Fireball" that belonged to T7 gear and a trinket that said "Adds an additional 5% for Frostfire Bolt" that belonged to T8 would make it blatantly obvious.

my take on this,

i do not want every class to scale equaly as it would take away from the diversity of class however, i think we should all sclae simularly:

ex: mages scale well with crit/dmg, rogues scale well with agi, warriors scale with str and crit rate

and i think everyone scales equally with haste as it's likea direct +X% dmg boost.

i'm am no wiz when it comes to theorycrafting but say:

from Tier A to Tier B all classes should go up ~X% each to keep DPS or HPS as closed to balanced as possible.

Mages get theri with crit rate and spell power and haste

rogues gte hit rating / agi / ap

wars get hit / crit rating and strength

these are just examples but still, with various way to gear you char depending on which stats are most beneficial based on your current stats ( some cases dmg > crit and vice versa)

so long as all class see a relativly equal increase in performance as we go up in tier levels scaling should not be a problem.

---

Q u o t e:
my take on this,

i do not want every class to scale equaly as it would take away from the diversity of class however, i think we should all sclae simularly:

ex: mages scale well with crit/dmg, rogues scale well with agi, warriors scale with str and crit rate

and i think everyone scales equally with haste as it's likea direct +X% dmg boost.

i'm am no wiz when it comes to theorycrafting but say:

from Tier A to Tier B all classes should go up ~X% each to keep DPS or HPS as closed to balanced as possible.

Mages get theri with crit rate and spell power and haste

rogues gte hit rating / agi / ap

wars get hit / crit rating and strength

these are just examples but still, with various way to gear you char depending on which stats are most beneficial based on your current stats ( some cases dmg > crit and vice versa)

so long as all class see a relativly equal increase in performance as we go up in tier levels scaling should not be a problem.

---

Shaman, due to the majority of Elemental Damage coming from Lava Burst and Lava Burst being on a cooldown do not scale well from Haste. 8 second cooldown is 8 seconds. The spell hits hard- so much so that you cannot delay casting it if you want to max your DPS, so you often have dead space in your rotation where you dare not cast a Lightning Bolt in place of Lava Burst lest you see your DPS DECREASE.

It's also worth keeping in mind complexity, and working with an existing design.

WOW is a pretty complex system, and though I imagine the developers make some effort to code things in a general way and reuse code as much as possible it seems like there's also a design principle that if they have to sacrifice simplicity and modularity for enhancing the 'fun factor' of the game they do.

From a systems engineering point of view (and I'm talking generally not purely comp-sci) just modelling WOW in a comprehensive way is a bit of a daunting task. Especially since it's a system where a lot of users are trying to find ways to make parts of it do things the designers may not have intended it to do.

To avoid surprises you'd have to run models for everything possible in WOW, and I cringe to think of the kind of computing resources that would take.

So you have imperfect ability to predict what changes will do, combined with legacy code that probably is not optimized around making life as easy as possible for the developers.

For a new game a fundamental scaling mechanic or system of balancing would make a lot of sense if you want PvP to be a significant component.

For an existing game that started out with the goal of appealing and distinct PvE flavors (or at least that's what I infer looking at WOW from a user's standpoint) I'd say that the developers do a pretty good job of taking care of scaling in PvE, and an adequate job of balancing in PvP.

As a consumer I can live with needing continual maintenance on WOW. It's not like having a car where you wind up getting charged extra to fix things if something goes wrong.

---

Q u o t e:
Wow

This thread was dead for almost as long as i was.

Not too bad a point but the difference class mechanics wont allow for even scaling of abilities.

Every tanking class is different and every dps class is different. There is no way that scaling will be linear for each class.

---

Yup.

A class with a 2.5 second cast spell as their main nuke will scale differently than a class with a 3.5 second cast spell as their main nuke. So, to make everyone scale the same, we would all need 3 second cast spells, with the same talents and the same everything. Until we have four classes: tank, melee damage, caster damage, and healer (with no talent trees)... we're all going to scale differently.

---

Q u o t e:


And in each case, it was a 50 dps increase. Linear increase is not multiplicative increase. But that's not a problem. It's a "curve" because you're measuring percentage improvement. Absolute dps is linear with respect to spellpower.

It works the same for all of the other stats. A 1% increase in crit adds to dps linearly but iterative increases in crit percentage increase total damage by less and less.

Again, I must wonder why you think this presents a problem.

---

It presents a problem because the 50 DPS value doesn't stay 50 DPS. That example assumes coefficients that behave.

The reason I preach about proportional gains vice linear gains is because the curve means it has reduced worth as time goes on. Lets consider this example:

http://zaldinar.bounceme.net/tcom/index.php?id=109

What it shows is the relative worth of the two spells as you scale up spell power. What we get from this is just beyond 2500 spell power, fireblast isn't worth casting anymore. Its worth diminished to the point where just straight fireball spam is better.

What does this mean? On the small scale, it means that mixing fireblasts in is no longer a DPS increase, which is bad, utility spells should always be worth something, and 2500 +damage is not necessarily outside the realm of possible gear for this expansion.

On a large scale, if fireblast represents Warlock, and Fireball represents Elemental Shaman, then once you get to that 2500 +damage range the Warlock gets unseated, and the Elemental Shaman has the potential to gain so much worth that the Warlock is no longer worth bringing. IE, mages in sunwell.


The raw value lines themselves may be linear, true, but the fact that they have different slopes is what causes strife.

Of note, if we scale up crit in those lines I linked:

http://zaldinar.bounceme.net/tcom/index.php?id=110

The lines are parallel. Each spell gains the same proportional benefit from the crit.

---

Q u o t e:
I'm not sure why you think this. It's more complicated than straight percentage improvements based on base damage, but hardly a mathematical unified field theory.

---

GC, please do not take this as an attack on Blizzards policies and actions in the past.

Have they gotten it right yet?

Not without buff/nerf, which should not be an active part of a balance strategy, it should be a reaction to bugs or player creativity that was not anticipated by the devs.

Instead of spending time intensely worrying about class balance, they could be keeping an eye on it, but spending time designing encounters and effects that make the game more interesting on an encounter for encounter basis.

---

Q u o t e:

But that eliminates the challenge of it. For instance, take FFB and FB. If the two spells scaled differently such that FFB specs become superior at T8 while FB spec was superior at T7, then all the TC'ers would have a job. They'd (like me) waste time at work not working and instead doing math on which spec is better.

A trinket that says "Adds an additional 5% for Fireball" that belonged to T7 gear and a trinket that said "Adds an additional 5% for Frostfire Bolt" that belonged to T8 would make it blatantly obvious.

---

But what happens if you wear the T7 trinket in T8 content with T8 gear? What happens if you mix trash purple drops with set gear? what happens if you force your crit rate up with gems?

Theres still plenty of theory left to do. Plus, all that theory is the boring fairly easy stuff, the real complex theory is things like double shatter combos, haste advantage windows under mobility conditions, precasting mechanics. Those are the real fun things.

---

Q u o t e:
This is a very interesting thread.

Our general philosophy is that we don't want relationships to be too predictable. Much of the fun of improving your character is seeing the effects that different kinds of gear have on different rotations. To use an extreme case, if every piece of armor just said "improves stats by 1%" getting loot would be a lot less fun.

Similarly, we don't want classes to be too similar. It's interesting and fun when you switch classes from say a warrior to a rogue and see how different stats affect you differently.

---

Totally. That we still want gear to do different things for different classes is a diversity I appreciate. This is why I've backed off the strict notion that all stats must necessarily do the same thing for all abilities (even though I feel there are many cases where stats that don't, as of right now, should: spell power and attack power being the most extreme case, where the large and complicated set of spell coefficients can have devastating consequences for rotations--see, for example, how Ice Lance gets obsoleted for use in the phantom 3rd charge of Fingers of Frost at or around 1500 spell power, if I recall).

With that restriction on stats all having the same effects relaxed, however, comes an increased responsibility to assure that stat point allocation is fair--I honestly don't believe the stat-point budget formula really allows for this. I've solved for the gradient of this function several times in just a simple two-variable case, and the result is extraordinarily complicated. This, to me, tells me that at best the fairness of stat point allocation is checked by hand (and "by hand" I mean with human eyes, using computers and calculators, no doubt) for several relevant gear setups at a subset of item levels.

Which is...not unreasonable. But there are significant issues with classes that use a smaller number of stats: again, I reference the warlock concern about affliction and haste. As I'm sure you're aware, GC, an item with 3 DPS stats has a higher maximum DPS value than an item with 2 DPS stats, even if they both spend the same number of item points on those DPS stats and are of the same item level. By severely curtailing the scaling with haste for affliction, it requires much, much more scaling with the other two stats (power and crit) that I feel cannot be guaranteed.

---

Q u o t e:
Scaling is very important to us in the sense that we don't want certain classes or specs to completely fall down at certain gear levels. This has happened in the past. But what those levels are is very important. If mages dramatically overtake warlocks when everyone has 20,000 spellpower, then I just couldn't care less except possibly as an item of trivia. If and when WoW ever reaches numbers of that level, the class mechanics will have changed so much in those years that its just not worth worrying about at this time.

---

I appreciate that the extreme gear level cases are not immediately relevant to balance now, but this was also part of my point: that while, in this example, mages only outstrip warlocks when everyone has 20k spell power, they are constantly gaining on warlocks even at reasonable levels. As we all know, viability is not just DPS: it's survivability and utility as well. Even not "completely outstripping" another class in terms of DPS can mean significant consequences for overall viability, for who gets a spot in a raid or who gets chosen for an arena team.

---

Q u o t e:


Yup.

A class with a 2.5 second cast spell as their main nuke will scale differently than a class with a 3.5 second cast spell as their main nuke. So, to make everyone scale the same, we would all need 3 second cast spells, with the same talents and the same everything. Until we have four classes: tank, melee damage, caster damage, and healer (with no talent trees)... we're all going to scale differently.

---

This is a result of poor definitions of "scaling the same."

Here's what I mean when I say scaling the same: proportionally. Where a 5% DPS gap is always a 5% DPS gap. This means that coefficients would be at a fixed ratio to base damage--instead of casting time determining coefficients (and let's be honest here: a fair number of spells don't have their coefficients determined by casting time; even heals have the 1.88 scale-up factor to correct for the +healing change, but this also makes one wonder why should heals have a 1.88 scale factor and not higher or lower. This factor was necessary to keep the healing done the same, but it also undermines the whole theory behind coefficients based on casting time. We can see that, for everyone having the "same" spell power, it doesn't hold water.), base damage or healing would. And this would be a much fairer system, with rotations and specs that are not summarily rejected at higher or lower gear levels due to forces and desires we cannot understand.

By no means am I trying to eliminate the complexity of the game. In my mind, every class, spec, and ability has a niche, a uniqueness about it that is what makes it special. It is the poorer scaling regimes in this game that destroy those niches, erase that uniqueness. I'm only advocating that these obsoleted abilities retain their niche, at least until Blizzard consciously chooses to have them supplanted by a new ability, class, or spec.

---

Q u o t e:
It presents a problem because the 50 DPS value doesn't stay 50 DPS. That example assumes coefficients that behave.

---

The 50 dps value will vary depending on the spell coefficient...

But for a given spell, any given spell, a set amount of spellpower will yield a set amount of dps throughout all spellpower values.

This means different spells gain more or less potency in relation to other spells. As I've said before, I think this adds to the game, not detracts from it.

If you absolutely wanted two spells to scale the same, it's a simple matter of adjusting the coefficient. The dps vs spellpower graph is linear.

---

Q u o t e:
The reason I preach about proportional gains vice linear gains is because the curve means it has reduced worth as time goes on. Lets consider this example:

http://zaldinar.bounceme.net/tcom/index.php?id=109

What it shows is the relative worth of the two spells as you scale up spell power. What we get from this is just beyond 2500 spell power, fireblast isn't worth casting anymore. Its worth diminished to the point where just straight fireball spam is better.

What does this mean? On the small scale, it means that mixing fireblasts in is no longer a DPS increase, which is bad, utility spells should always be worth something, and 2500 +damage is not necessarily outside the realm of possible gear for this expansion.

---

This is where our disagreement is. I don't think this is a bad thing. The spell rotation changes (granted to a more boring one, but that's a matter of specific implementation) depending on gear level. I don't know about you but I like that idea. It makes the class more complex. None of this makes spell scaling "out of control". It presents the opportunity to add dynamics to gameplay.

---

Q u o t e:
On a large scale, if fireblast represents Warlock, and Fireball represents Elemental Shaman, then once you get to that 2500 +damage range the Warlock gets unseated, and the Elemental Shaman has the potential to gain so much worth that the Warlock is no longer worth bringing. IE, mages in sunwell.

---

Unless of course, everyone gets a fireball, or a spell that scaled like it. It's not impossible to make two spells with different base damage and different coefficients scale closely within, say, the 1 to 5000 spellpower range. I think it's a fair trade-off for allowing more interesting mechanics.

Your objections seem only to be with specific implementation and specific coefficients. None of the problems you presented can't be resolved by simply adjusting numbers.

---

Q u o t e:
The raw value lines themselves may be linear, true, but the fact that they have different slopes is what causes strife.

---

Slope is controlled by the coefficient. If the slopes don't match, adjust the coefficient...

---

Q u o t e:
Of note, if we scale up crit in those lines I linked:

http://zaldinar.bounceme.net/tcom/index.php?id=110

The lines are parallel. Each spell gains the same proportional benefit from the crit.

---

Unless of course, you compare spells with different crit coefficients:

http://zaldinar.bounceme.net/tcom/index.php?id=111

---

Q u o t e:
Have they gotten it right yet?

Not without buff/nerf, which should not be an active part of a balance strategy, it should be a reaction to bugs or player creativity that was not anticipated by the devs.

Instead of spending time intensely worrying about class balance, they could be keeping an eye on it, but spending time designing encounters and effects that make the game more interesting on an encounter for encounter basis.

---

I am not very much in favor of dumbing down class mechanics to make designing easier. I agree that their track record isn't very good. But that doesn't mean we should give up on the principle. They simply need to do their math more carefully.

---

Q u o t e:
But what happens if you wear the T7 trinket in T8 content with T8 gear?

---

If this was viable, there'd be no point to gear progression. If a trinket that dropped from T7 content was just as good, provided you use fireball instead of frostfire bolt, in T8, then why bother?

---

Q u o t e:
What happens if you mix trash purple drops with set gear?

---

If T8 shoulders increased spellpower by 10% and T7 shoulders increased it by 9%. The T7 bonus was a 5% increase in fireball whereas T8 was a 5% increase in FFB, the obvious choice is T8.

Without scaling differences in the spells, % increase is % increase. There is no variability.