Avoidance Diminishing Returns?

[Jericho]

Alright, first off, I have not been lucky enough to get a beta key. All I have the ability to do is to go on the PTR and AOE grind for fun at the moment.

However, after this push I noticed while looking at my stats that there was "diminishing returns" listed under all the avoidance tooltips. Forgive my ignorance since I have not read anything about this. What does this mean?

[Hypatia]

They're changing the rules so that your actual avoidance is reduced from what your gear "should" give you. This is a *good* thing. Traditionally, avoidance has been something that has gotten better and better as you get more of it, which causes problems later in the game. It means the devs can't put avoidance on gear quite as happily as they can put other things on. It means they have to pull tricks like Sunwell Radiance to make up for people who just plain can't be hit any more. :X

Same thing as armor, really: You'll notice that you get "less and less" damage reduction for each bit of armor you get. However, the way it's tuned, a given amount of armor always provides the same effective health benefit. Presumably, avoidance will be tuned to the same end: Put a certain amount of dodge rating on, and it will always benefit you the same amount no matter how much other avoidance you have on.

Basically, every bit of avoidance you get will still be good--it just won't get better the more of it you have. You'll notice that your avoidance numbers are lower than you would predict from your gear. I'm sure we'll have addons and formulas (like the EH calculator addon) that tells you the actual avoidance benefit you get from a piece of equipment by the time the game goes live.

[Bayho]

Does anyone have any numbers on how avoidance will plateau? I also wonder if this means that as dodge becomes higher, to a point, will you then focus on getting parry higher as it becomes more effective skill point per point. Will a balance of avoidance be more effective than stacking dodge over parry? Will 25% Dodge and Parry be easier to get than 30% Dodge and 20% Parry?

[phaze]

Does anyone have any numbers on how avoidance will plateau?

Preliminary numbers show roughly 5% drop in effectiveness per 100 rating.

For example: if 100 dodge rating converts to 5.29% dodge pre-LK, it will convert to ~5.04% dodge in LK, followed by the next 100 rating only giving ~4.78%, and so on.

I'm guessing the Blizzard UI doesn't have the actual formula available, like it does for Armor's mitigation?

Will a balance of avoidance be more effective than stacking dodge over parry?

Yes. That makes Defense Rating a bit more attractive as well, since it promotes a well-rounded stat distribution.

[Klimpen]

@phaze

A clarification. That's not a 5% reduction in avoidance per 100 rating, it's a 5% reduction in effectiveness per 100 rating.

Just asking a'cos it's a bit ambiguous.

[orcstar]

Avoidance without diminishing returns:

It's important here to talk about remaining hits and extra avoidance.

Three scenarios, you have 0%, 50% or 90% avoidance. We're gona add an extra 5% to each of these.
In the first situation, you get at 5% avoidance, avoiding 1 in 20 hits, so you get about 5% less damage.
In the second situation you will also avoid an extra 1 in 20 hits, but because you already avoid 10 in 20 hits the actual reduction in incoming damage will be 10%.
In the third situation you're already avoiding 18 in 20 hits, and adding another 1 in 20 to youravoidance will reduce the incoming damage by 50%.
5% added when you already have high avoidance does a lot more then 5% added when you have low avoidance.
Now let's take the same example and add 10%. You will notice that in scenario 3 you won't get hit at all...........the extra avoidance icreases your lifetime to unlimited and you will die of boredom before reaching 0 hp.

Hence accelerating returns. This is what's going to change.

[phaze]

@phaze

A clarification. That's not a 5% reduction in avoidance per 100 rating, it's a 5% reduction in effectiveness per 100 rating.

Just asking a'cos it's a bit ambiguous.

Yeah, that's what I meant. =)

[Satrina]

I'm guessing the Blizzard UI doesn't have the actual formula available, like it does for Armor's mitigation?

It does not, no.

[Radhja]

This is just crazy enough to work...

...so it promotes balance among avoidance types, but isn't it ultimately a nerf?

[Orenus]

Yes. That makes Defense Rating a bit more attractive as well, since it promotes a well-rounded stat distribution.

It sounds like this will hit bears harder than us, since they don't have any other stats to spread around.

[phaze]

so it promotes balance among avoidance types, but isn't it ultimately a nerf?

Kinda. It prevents the crazy scaling by limiting your avoidance stats, so in that respect it's a negative. However, it also prevents the future need of another Sunwell Radiance, and that's a large plus both for players and for devs.

End result: it's a good change to the game mechanics, to prevent problems.

It sounds like this will hit bears harder than us, since they don't have any other stats to spread around.

The devs still have the option to modify the DR curve for Feral Druids, so that it affects them to an equal degree.

[Hypatia]

It's also possible that the diminishing returns is done to avoidance as a whole, and the proportions of various types of avoidance are kept flat. If that's done, then spreading out among different avoidance types won't give any advantage, but neither will classes with access to less types of avoidance be at a disadvantage.

I should go test stuff. But my pipe is small and patching takes soooo long.

[Kazeyonoma]

That's what I read when i first saw "diminishing returns"
not that the individual avoidance of each is nerfed as it gets higher and higher, but instead that the formula no longer went
1-(M+D+P)

and is instead (1-M)(1-D)(1-P).

Some quick math using my normal stats:
1-(11%+20%+18%) = 51% chance to be hit.

using new diminishing formula:
(.89)(.80)(.82)= 59% chance to be hit.

now lets add 10% say... dodge.

1-(11%+30%+18%) =41% chance to be hit
vs.
(.89)(.7)(.82)= 51% chance to be hit

So you see that the old way, allowed a straight 10% avoidance added, which eventually leads to unhittable tanks.

The new way makes it so that as you get higher and higher in avoidance levels, you eventually cap out at a high number but never 100%. We see here that 10% avoidance only decreased your chance to be hit by 8%.

Could this be what they meant by diminishing returns on avoidance?

[2ndNin]

Its not actually diminishing its linear returns, that is

1 dodge rating = x% reduction in incoming damage.

Taking a silly example, if you have 0% avoidance and stack to 50% with a single item it will reduce your damage taken (as an average over many many hits) to 50%.

If you then add another "god dodge item", in traditional WoW you would add 50% avoidance, taking you to 100% (increasing returns). Instead you would get 50% dr from it, taking your 50% dodge to 75%, adding another takes you another 50% dr to 87.5%... and so on.

Armour is linear, the same +armour = the same DR at all levels, its just that it appears to diminish as you now need 10,000 to get 1% rather than 1,000 because each 1% you reduce damage taken by is much more valuable. Diminishing returns would scale off, so that stacking 2 "god dodge items" would instead result in < 75% damage reduction that a linear increase would. If its diminishing returns then we have a problem, we will reach a point where it is no longer effective to stack the diminishing stat and instead stack linear ones (health etc), with linear we get consistant benefit but likely favouring an EH approach still after a certain point, and with increasing returns we favour high avoidance tanks but can still afford to stack increasing other stats.

The issue with making things linear is the ilevel formulas are relatively linear, a tier is roughly 13 ilevels, however the content difference can be much higher, the increasing returns made the TBC content difficulty ramp in the end game easier, each tier was essentially much better than the last for Tanks as we gained avoidance, LK will need to be carefully balanced if its going to truly be a linear advance as pots etc will have a larger effect relative to the tier gap.

[Del]

My one worry, and maybe Hypatia or someone else could shed some light on this, is that the ton of avoidance on tanking gear will be flatly inferior to stam because of it's unpredictable nature.


Avoidance is viable in BC because of the super linear sailing, but not so much so that it is a clear cut choice in gearing. Hence the numerous EH vs Avoidance debates. Where this really hurts us is in gem sockets. I'd hate to see a point where stamina is flat out better than avoidance in every gem slot, regardless of socket bonus.

[phaze]

That's what I read when i first saw "diminishing returns"
not that the individual avoidance of each is nerfed as it gets higher and higher, but instead that the formula no longer went
1-(M+D+P)

and is instead (1-M)(1-D)(1-P).

Could this be what they meant by diminishing returns on avoidance?

That would require reworking the combat table mechanics. I don't think this is the case.

From the EJ testing I saw, each rating just has an independent conversion formula (with diminishing returns now built in), they still combine additively on the combat table.

Its not actually diminishing its linear returns, that is

1 dodge rating = x% reduction in incoming damage.

For LK, it's definitely DR: your avoidance gain per rating point is decreasing as your rating goes up.

the ton of avoidance on tanking gear will be flatly inferior to stam because of it's unpredictable nature.

The unpredictable nature of avoidance is unchanged. It's still useful for survivability (through EDF), just as before.

As you move deeper into BC raid content, Avoidance becomes better and better because you're climbing farther up the linear function. For LK raid content, the value of Avoidance should remain much more steady.

I think the goal is for avoidance to mirror the behavior of armor:
Avoidance stats are subject to diminishing returns, but total avoided damage is not.

[Kazeyonoma]

I didn't think so either phaze, and haven't kept up with EJ testing,

Basically the more dodge rating you have... the more dodge rating it'll take for you to get another 1%.

[Hypatia]

What thread is the EJ testing being discussed in? (It's not a benefactor's bar thread, is it?) I'd like to take a look, but have somehow missed finding it. :X

[phaze]

What thread is the EJ testing being discussed in? (It's not a benefactor's bar thread, is it?) I'd like to take a look, but have somehow missed finding it. :X

Combat Ratings at level 80

Looks like Whitetooth's updated the first post with more info: each class has a different formula, which means Druids aren't being punished as severely for piling everything into only Dodge.

EDIT: I'd copy over the formulas, but it doesn't look like the EJ notations transfer over.

[Satrina]

Yeah, the EJ site supports TeX directly via mathTex. Very nice feature, that.

Looks like they also instituted hardcaps for dodge and parry as well.