[First time posting about theorycraft issues here, please be gentle.]

Previous discussions have talked about the issues of the itemization formula Blizzard has chosen: it gets more expensive to increase the amount of a given stat on an item as the amount of that stat increases. This means that to provide meaningful upgrades, item levels have to climb rapidly--"mudflation". There was a fair bit of discussion about this when TBC first came out, because Karazhan and SSC loot were not huge upgrades in the way MC/BWL//Naxx loot were, due to item levels and poor stat allocation. Obviously, 2.1 featured massive stat reallocation that resolved many of the problems--Karazhan loot, for instance, provides nice upgrades over level 70 blues--and people are not complaining anymore, having happily returned to killing dragons, demons, etc. However, the itemization formulae are still in place and the core issues have not changed. If I have followed discussion on the board correctly, this is part of the issue behind shadow priest complaints--there are a very small number of stats that are relevant to a shadow priest and these stats are running into problems with the itemization formulae.

I think Blizzard realizes this and is attempting to provide more stats to raise, thus enabling meaningful upgrades without drastically inflating item levels over time. Examples of this phenomenon are resilience in PvP, the changes to haste rating for channeled spells in 2.1, and the "ignore armor" bonus on the warrior tier 6 DPS set. (Speaking as a paladin healer, I also think that this is what Blizzard is trying to do with spellcrit and mp5--balance the effectiveness of both to provide for more meaningful upgrades in paladin gear, but I'm not convinced they have found a good balance of usefulness for spellcrit and mp5.) Basically, with the current itemization formula, if you have a given item level increase and five stats to raise, you can raise all of them by a small amount and provide a better upgrade than if you only had three stats and could raise those by a small amount plus epsilon.

What other bonuses do you think Blizzard could provide that would be useful and would not be imbalanced PvE- versus PvP-wise? (PvP bonuses or PvE bonuses, either one.) The "ignore armor" bonus has been cited as an elegant way of providing PvE utility without unduly affecting PvP--are there any other ideas out there?

For starters they could make spell penetration useful.

I think they could add a "Critical Strike Damage Bonus rating", which assuming it scaled decently would be a pretty nice stat for high crit rate classes. It would also help be a gear counter to resilience, giving it some PvP practicality.

Past that, all I can come up with maybe gear bonuses that reduce mana/energy/rage costs, but I'm not too sure those are well suited to anything other than set bonuses. And mana cost reduction is really just mp5 in flashier clothes.

Actually, they don't. The formula looks roughly like:

cost = ( stat1^1.5 + stat2^1.5 + stat3^1.5 + ... )^(2/3)

So if you multiply each stat by a constant, then the cost is multiplied by the same constant. The nonlinear scaling only comes into play when you're shifting the point allocation between stats on the item.

You're only raising one stat. Try raising both from 10 to 20 and you'll find that the cost doubles exactly. More generally, if you increase all stats by X%, then the cost will also increase by X%.

Because you're changing the relationship between the two stats.

You get the most benefit when all stats are realitively equal points wise:

IE:

+10 str
+10 agi
+10 int

The futher they deviate form eachother is when it gets expensive.

Take another example where the points scale the same way.

10 str, 10 agi = 14.14 (Not sure where 15.87 came from, sqrt(200) is ~14.14)
Make it 15 str 15 agi = 21.21 which is 7.07 points higher
Make it 20 str 20 agi = 28.28 which is 7.07 points higher

As long as the ratio of the stats remain constant, they scale linearly.

I thought that was the point, raising just one stat in relation to the others, example being how spell damage and only spell damage scales well for Shadow Priests.

As long as they implemented this as a rating, that might be pretty cool... Would you gain more from 44 spell damage, or an extra 3% of your spell damage added to your casts, etc etc. Really depends on what spells you cast and what coefficients they use, and such. It would make for some complicated gear decisions as well.

This kind of stat is available in the game Titan Quest (and probably many other games I've never played): both flat increases to damage as well as percentage increases. It works for that game because your character grows in power by orders of magnitude, and it's balanced for single player. But it will quickly spin out of control in WoW. Bob Raider's fireball already hits 35% harder than Joe Casual's, but with another tier's worth of damage multipliers, Bob could more than double Joe's damage. We started to see that near the end of WoW 1.0, and it wasn't good for the game. Plus, the developers have a history of losing control of gear scaling even without percentage multipliers, and the corrections are rarely pleasant.

I've noticed people tend to use 'mudflation' to represent a lot of different things. What I have always thought it meant was that the power of everyone in the game gradually increases over time (mostly due to new content offering gear that completely outclasses previous gear) which then has the side effect of trivializing old content.

What do other people mean when they say it?

It's a problem of relative power gains.

Say you start out with an item with +10sta +10str +10agi. If you go up 5 ilvls (hypothetically) you get a new item with +15sta +15 str +15agi. You've increased in power by 50%. Go up another 5 ilvls and you get +20sta +20str +20agi. You've increased in power 100% from the first item, but people aren't looking at that first item, they're looking at that last item, from which it's only a 25% upgrade.

Extrapolate this out, and once you get to the point where you have +50 or +100 of a stat to start with, you have to increase it by more than 5 ilvls (or 10 or 20) for it to have a meaningful effect. This is then compounded the next time you increase stats, over and over until any upgrade at all is virtually meaningless.

It's kind of a catch 22. If you have a linear scale, it gets to a point where upgrades are meaningless, since you reach a point of such small increases relative to existing stats, that it's unnoticeable as an upgrade. If you have a % increasing scale, it eventually grows so large that the numbers you are using are meaningless.

Personally, I'd rather have a carefully planned out % increase type of progression path, combatted in part by introducing new and varied stats in addition to existing stats. They've done this to some extant, but they could go much farther as far as meaningful new stats:

-Spell penetration could be made useful by adding resists to pve mobs. Spell penetration would be needed to do max dmg.
-A stat could be introduced to combat the frequency and severity of glancing blows. They just need to figure out a way to seperate it from the weapon skill stat, since that is tied to leveling up, which they don't want anyone to have an unfair advantage in.
-They could add a pve function for dodging/missing/avoiding physical ranged attacks, so hunters have to collect something to make them be able to counteract that effect.
-Besides more armor, they could have mobs have a greater than normal amount of avoidance, so tanks/melee have to collect a stat that reduces dodge/parry. Again, divorce this from weapon skill, so you do not unbalance things when it's time to level up again.
-Add new talents that reward collecting a certain type of a new stat. Similar to how mages have a talent that gives them more benefit to the +dmg they have on certain spells, or several classes (hunters, warriors) have talents to increase the benefit of stamina.
-Add mobs that can reflect spells, and then give casters a stat to counteract this effect.

These are just off the top of my head, i'm sure others can come up with more imaginative ideas .

There's many directions they can go with stats, the important thing is to give people a reason to want those stats, and to make them broadly available and applicable to the chosen focus of the player, be it pve or pvp.