Mitigation Comparisons – 4 tanks

[ziggyunderslashone]

There’s been a lot of talk about mitigation recently. Who is in range of who, who has too much and too little. Frankly, mitigation is somewhat subjective. It’s based on gear levels, buffs and also the perception of the healers involved.

This is all a little fuzzy for me and my compulsions, so I set about doing the maffs. Many thanks to the testers here and at EJ for the numbers that make up this sheet.

Edit: Current version of graphs on page 12


The Spreadsheet

http://www.savefile.com/files/1827360

updated 08/10/08

Notes:

This spreadsheet was built in somewhat short order, feel free to check for accuracy. I’ll update analysis as I go. It's a little rough and makes some assumptions, but it will show any large or glaring errors in class scaling.

Currently I do not have base health or stats for dks and bears at 80. Until I do, I’ve left both off. This will disadvantage warriors somewhat.

I understand diminishing returns will apply to miss provided by defence, but in the absence of the math, this isn’t present, this will advantage bears a little. Many thanks to Whitetooth for [url=http://elitistjerks.com/f31/t29453-combat_ratings_level_80_a/]his work on diminishing avoidance returns.
[/url]
The numbers I’m using to represent mitigation are “Average seconds before healing is required before death.” This number rates health pool, armor (and associated –damage buffs, def stance et al), as well as avoidance and even attack speed debuffs if required (expertise’s effect on parry should you so wish)

I’ve also included a Raw Mitigation figure, a worst case zero avoidance string. Anyone who’s tanked knows it’s not just what you can survive, but what you will.

The Scaling factor multiplies the stats derived soley from gear, this allows the sheet to chart how the classes compare across the gear progression spectrum, assuming that spending remains consistent. (Spending is unlikley to remain consistent, but current spending is the only trend information available)

Edit: Updated with information provided by the good people of the forum. The spreadsheet now includes: Shield Block (with correct multipliers), Icebound Fortitude, Unbreakable armour, and ranged slot statistics.

Graphs

Avoidance:




Raw:



Conclusions

Paladins and warriors.

As I have more accurate information pertaining to warriors and paladins, namely the base stats, I present a more detailed analysis.





We are slightly behind warriors in terms of raw survival. About 4% to be accurate. The increased stamina scaling we have may one day make up this deficit, but I suspect it may be too little.

In long term damage reduction, the picture is even less rosy, we begin nearly 5% behind, and due to warriors rather impressive block scaling, this gap increases to anything up to 7%. Far from ideal.

And it’s a shame, with relatively few improvements in accuracy to the sheet, we saw these charts move from nigh on identical mitigation, to a sizable and increasing gap.

One thing that isn’t included in this analysis is Ardent Defender, but the same reason it’s close to impossible to model, it’s not terribly effective as a survival mechanic. It cannot be relied upon due to its leapfrog nature. I’ve always been keen on the talent, a gambler at heart, but even I wouldn’t attempt to analysis it against more straightforward damage reduction.


Death Knights

Interesting one. At a disadvantage on big hitters, prone to that bad luck string or parry gibbing. But with a significant advantage on fights where taking low damage or hits is a premium. Bloodboil and void reaver style bosses are most obvious, but anywhere where healers are stretched thin or otherwise engaged.

I’d be a touch worried that such an advantage/disadvantage system could see them become a niche tank in the eyes of many (and we know how that goes) but overall, I can see them balancing out well across content. I suspect that might be slightly too strong scaling in the deeper end of avoidance, the differences between this and other tanks becoming increasingly pronounced.

Interesting, and somewhat risky, but job well done, blizz. Probably equal but definitely different.

Bears

Oh dear. And it was going so well. Starting off massively strong and scaling poorly, I’m getting this awful sense of déjà vu.

I suppose they're supposed to be the opposite to the death knight, the non avoidance tank. But where as DK's fit that role at both ends of the scale, druids start with the best of all worlds, and then just scale badly in order to fulfill the raw survival role.

History will repeat if this remains unchanged.

[PsiVen]

I can't take a detailed look right now, but I get the impression that what you're modeling here is essentially our worst case effective health. You are quite correct that talk about mitigation is subjective -- this is actually the only area that we feel comfortable in. Unlike warriors and DKs, paladins get to look at no CDs, zero avoidance scenario and and still account for pretty much all of our mitigation abilities. So really we ought to be well ahead of both here.

I'm not sure how meaningful it is to model avoidance without accounting for average shield block uptime etc., in an arbitrary-time sampling you have a chance to dodge just as you have a chance to have shield block up.

Does your scaling factor account for diminishing returns on avoidance gear? If not I imagine that would pull the DK avoiding tests down to earth a bit.

[Bobness]

I'm not sure how meaningful it is to model avoidance without accounting for average shield block uptime etc., in an arbitrary-time sampling you have a chance to dodge just as you have a chance to have shield block up.

/agree

[ziggyunderslashone]

I can't take a detailed look right now, but I get the impression that what you're modeling here is essentially our worst case effective health.

Correct ya, I find the worst case is the position one should gear and assume around, the rest is bonus.

I'm not sure how meaningful it is to model avoidance without accounting for average shield block uptime etc., in an arbitrary-time sampling you have a chance to dodge just as you have a chance to have shield block up.

Very true, as it's over a time period. Seeing these results I'm compelled to go back and apply them (both SB and the dk cd's). Initially they interupted my idea of it being a sort of "average snapshot" by being essentially two states, one of which doesn't represent worst case, but I think you're right, that feels flawed if we're talking extended periods.

Does your scaling factor account for diminishing returns on avoidance gear? If not I imagine that would pull the DK avoiding tests down to earth a bit.

For both dodge and parry yes, whitewooth alludes to the fact that miss is effected too, but without the "cap" I haven't worked this in. It's an inaccuracy for sure, but is fairly consistent, effecting the relative place of druids slightly with their lower def from gear.

[2ndNin]

Run some tests with averaged shield blocking, emergency only shield blocking (assume a 25/50/75/100% ideal shield block [so the Warrior will save their shield block for a hit that a) hurts, b) would kill them c) needs blocked rather than using it as mitigation]), if you just look at raw numbers without taking a major mechanic into consideration it won't work.

[Faeth]

My 2 cents for now (if you didn't already intend to do so)...Instead of replacing any info with the onest including SB, etc, i'd ideally like to see them put next to eachother for clear comparsion...

Nice work btw

[ziggyunderslashone]

Same analysis, but reduced to paladins and warriors in order to include base stats (these were accurate as of the recent major health buff). 10/60 Shield block up time included in avoidance. Lower uptime will appear somewhere between both graphs. EDIT: 10/60 is incorrect, corrected in further posts.

Avoidance:



Raw



As expected, the health difference makes them nigh on identical in raw survival, with a marginal edge to warriors initially, and shield block pulls warriors slightly ahead in avoidance as gear moves along. Largely due to multiple scaling on block.

[Io.Draco]

Nice

[Arnock]

Same analysis, but reduced to paladins and warriors in order to include base stats (these were accurate as of the recent major health buff). 10/60 Shield block up time included in avoidance. Lower uptime will appear somewhere between both graphs.

Avoidance:



Raw

I'm really liking the looks of this

Pallies and warriors should generally be the same avoidance/migitation wise, due to the fact that we both use sword/board unlike DK's druids

on the original graphs, Druids seem to be the EH/Migitation tanks, while DK's seem to be the avoidance tanks and pallies/warriors are somewhere in the middle, The whole system looks kinda like they still each have their 'niche's' but more along the line of HOW they reduce incoming damage, not how much is reduced

[Holyfuri]

nice work. This makes me think an analysis like this is exactly where GC and the devs got their statement that we are "right up there" with warriors, with their shield block giving them an advantage in higher levels of gear.

[ziggyunderslashone]

Does anyone know where I can find base stats for all four classes by the by?

Seems a shame to be rounding off its inaccuracies (cooldowns, glyphs, ranged slot) and be missing something so basic.

[Andryana]

Same analysis, but reduced to paladins and warriors in order to include base stats (these were accurate as of the recent major health buff). 10/60 Shield block up time included in avoidance. Lower uptime will appear somewhere between both graphs.

What major health buff are you referring to? Have i missed something??

And thanks for this great information

[Worldie]

I tend to agree with the OP.

While it'll be probably true that we will take overally more damage than war / dk due to their CD, it's also true that unconstistent damage is what kills tanks.
I'm fine with that, maybe it's what GC and the devs mean. Even if i'd not dislike a "mini shield wall" after all.

[ziggyunderslashone]

What major health buff are you referring to? Have i missed something??

A few patches back they increased base health for all classes.

[Snake-Aes]

Isn't the shield block uptime 10/40?