View Full Version : Death Knight Damage taken

02-11-2011, 11:02 PM
So I been doing my own little tank theory crafting to myself and silent readers in my guild forums for a little while. I came up with a way to compare damage taken between different DK's on the same type of boss fight. The following is a copy paste from our forum. I have found the weight of my task a little daunting and wouldn't mind some input or help from any other Dk tanks.

So I been searching through world of logs and wow progress looking for guilds that run DK tanks. While I can't count every factor I can look at what attacks the tank focuses on DS HS RS, the amount of damage taken and the amount healed. I have only found 3 tanks to review so far including myself. From the 3 two of the tanks took more damage on the fight while prioritizing DS but haven't had a chance to see gearing style on them. Self healing from both said tanks were not enough to make up for the difference between damage taken but I need to reconfirm that absorbs were factored into that. Its only an observation at this point. I want to review and monitor 10 tanks before I say bitch your being bad kind of things. I will say that it was quite interesting to look at. The difference in damage taken was over 1million healing done difference was only 100k and damage done by DS(and thus correlation of how many DS's done) was more then double in my initial findings. While the difference in gearing style is a lot harder to compare with every variable that happens in a fight mechanic attacks can be compared a lot easier. And with such clear distinctions in the numbers this approach looks promising.

So I have continued my search and shit it sucks. Keep in mind Im looking for groups that Run DK mt's meaning if the fight is a single tank fight the Dk will be the one tanking. I'm starting it off like this so that I can cut out more variables. Once I designate my 10 tanks Ill expand from there. To add to my search I need to find tanks are all doing the same fights on normal coupled with doing them on normal as of 4.0.6. So that right there cut my list in half ><. I think I need to find a new searching method because this one sucks ass.

Anyways while looking deeper into the fight I ran across some interesting finds.
- World of logs wont show things like blood shield for certain tanks. Unless I'm missing something there's no way a Dk tank doesn't get a single blood shield if their doing DS during their fight.
-I need to find out if it is possible to see how much blood shield they gained during the fight and how much of the blood shield is actually absorbs any damage vs just falling off. I may have to look into direct participation from the test subjects.
-Another thing that may very well be a major factor during the fights is debuffs. So far out of every DK tank I have had a chance to look at I take the least amount of damage on the fight that I'm looking at. I understand that this is inconclusive and unreasonable because what if my ass is sucking on other fights. Like I said I'm gonna get to that eventually but atm I need test subjects. Anyways while looking at the up time of buffs and debuffs between my self and two others I noticed that my frost fever and scarlet fever up time far exceeded everyone else.

The difference in damage taken by tanks for the particular fight Im looking at varies from 2mil to 4mil+. These numbers will be rounded.
-(T1) 2.1mil
-(T2) 2.9mil
-(T3) 4.7mil
-(T4) 4.2mil
Rounded time for FF(frost fever) SF(scarlet fever) and BB(Blade Barrier) LF(Length of Fight)
-T1 (FF-312s) (SF-278s) (BB-245s) (LF-6.44m)
-T2 (FF-131s) (SF-131s) (BB-249s) (LF-6.12m)
-T3 (FF-091s) (SF-188s) (BB-276s) (LF-7.35m)
-T4 (FF-116s) (SF-000s) (BB-335s) (LF-8.04m)

I did go back to double check where the damage was coming from. Everyone took equal amounts of damage from spells in relation to the duration of their fights. So the difference in damage is mainly coming from melee attacks.

Once again I'm still working on subjects so this is all just preliminary findings.

02-12-2011, 12:01 PM
I would greatly appreciate any volunteers for this project. 10 or 25m is not a problem. I ask that you guys be as close as possible to 9/12bosses down so I can expand onto other fights asap. When I post results I can keep you anonymous if you want. I'm not trying to point out who's good and who's bad. I want to see the areas that reduce the most damage between tanks and then bring that into the game play. Example, if we find that tanks who use dancing rune weapon more often during the fights take 10% less damage then tanks who only use it once. I would of course make sure that I'm using DRW as often as possible.

02-15-2011, 04:00 PM
I think I remember reading there's a problem in that the combat log doesn't distinguish absorbs and their sources. Feel free to use me though, not that I'm entirely sure what you're doing exactly;

We've had a couple of dodgy raids since 4.0.6 but should get things rolling again properly this week starting on Thursday;

http://www.worldoflogs.com/guilds/46671/ my character is Miniwrathy now >,>

02-16-2011, 03:36 PM
Appreciate it wrathy. Atm I'm looking at the fights that are solo tanked and compare it to other dk tanks on the same fight. After that I review world of logs and check to see if the amount healed compensates for the amount of damage taken. Further more I check to see if there is a difference in the up time on debuffs between the tanks taking higher or lower damage then others. Once I get my set number of tanks I will review their armory and in some cases contact them to figure out what stats they are pushing. Mastery vs avoidence kind of thing. I also want to look at how much DS their using during the fight and check to see if its helping to mitigate damage and to what extent. Sry for the jumbled post, doing this from my phone during lunch break.

02-16-2011, 04:50 PM
After that I review world of logs and check to see if the amount healed compensates for the amount of damage taken.

I'm a little confused by this.... are you talking death knight vs. death knight, or death knight vs. Other tanks? Are you monitoring buffs/gear/other variables that will change this such as IBF/VB/BS?

02-16-2011, 05:27 PM
I don't want to say its DK's vs DK's because I'm not trying to tell people "look your doing it wrong, you need to do it like this." I want to find what it is that allows some DK's to take considerably less damage during fights then other tanks. I will be factoring in your named buffsv and so much more, DRW, trinkets, army, AMS ect. It would be awesome for DK's everywhere if it got to the point where we could tell you what CD's or rotation you could use on a specific fight to capatalize on damage redution. After looking at my inital findings I became aware of so much more CD's I could have been using.

02-17-2011, 01:31 AM
Ok so an example of what I mean.

T1 http://worldoflogs.com/reports/rt-ddw98l3r9yqlyxva/sum/damageTaken/?s=6891&e=7297
T2 http://worldoflogs.com/reports/mifnu7u546wzepjr/sum/damageTaken/?s=4104&e=4534
T3 http://worldoflogs.com/reports/rt-wll4h7167r88tgab/sum/damageTaken/?s=3395&e=3814

These are not the exact same tanks and fights given on previous post's.

Using WOL again we can take a look at what cd's the tanks are using during the fight, like this one.
T1 http://worldoflogs.com/reports/rt-ddw98l3r9yqlyxva/details/6/?s=6891&e=7297
T2 http://worldoflogs.com/reports/mifnu7u546wzepjr/details/7/?s=4104&e=4534
T3 http://worldoflogs.com/reports/rt-wll4h7167r88tgab/details/8/?s=3395&e=3814

If you look closely the only real difference in damage taken it melee damage. In time we can look into what type of stats the tank is pushing, but I'm looking to get more tanks before we get to that because I feel following around 3 tanks would be very inconclusive.

I dont want anyone to feel like they need to be ashamed of what their logs might show but if you want me to keep quite about who is who I will. I'm T1 in this example if anyone was wondering, I apologise for not introducing my char earlier yet asked for you guys to post yours ><

02-17-2011, 11:00 AM
You also need to consider that random randomness is random.

The same tank on the same fight on 3 different occasions could take less damage or more damage on any fight (without using cooldowns/trinkets/etc.) based on the fact that avoidance is random and will change the numbers. So either you need to average the same tank on the same fight over an accurate sample size (determining outliers etc.) to get a look at damage, or simply look at what the cooldowns do, how much reduction, and appropriate use of such. This still won't help when looking at overal damage however.

ie. You only used bone shield xx amount of times over the course of a fight when it could have been used xx amount of times.

These types of analysis do not require WoL logs however. Also, I've never seen DK's take considerably less damage than any other tank. If they did, I'd be looking at the rest of the raid comp... external cooldowns being present or not would have a huge impact, same if you had a disc priest vs. a non disc priest.

02-17-2011, 02:56 PM
2mil vs 4mil is a lot more then randomness. Possible up times for debuffs BS and such should already be know to the person tanking. I was to point out what a tank taking 2mil less damage on a fight is doing differently from the other tanks. If it is raid composition and healer types that are contributing to this reduction then so be it. The goal here is to be able to tell people what debuffs, buffs or cd's cause a noticable redution in damage. Like wise with WOL we can see what attacks a tank is focusing on. If he's using 2x more DS then other tanks but doesn't take any less damage then the average we can say that for this particular fight its not effective to spam DS. None of this is something that we can figure out from looking at one group of fights, it will be a large and long process.

02-17-2011, 03:03 PM
I think there are a few too many variables here. Internal cooldowns, external cooldowns, avoidance randomness, length of fight, cleansed debuffs on the tank, debuff uptime on boss, absorb shields, gear levels, etc. You maybe could do this sort of analysis through WoL, but you'd have to keep so many controls on the data that you'd be hard pressed to find a statistically significant number of participants who fit the control criteria in order to make any real determinations. Much simpler to build the experiment yourself with all the controls in place (like in a spreadsheet or Rawr) than to try to find the circumstances you're looking for in WoL.

02-17-2011, 04:06 PM
So much to consider even as far as fundamental playstyle differences are going to make it very tough for you, I think. If you can extract useful info from it, then so much the better!

A fair bit of it can come from playstyle. For example I like to hold my cooldowns in reserve for unfortunate RNG or awkward moments when a healer has to move, even the little ones. The only ones I really spam are VB and RT. I like to keep my damage reductions on hold just in cast which is obviously going to mean I'll take more damage than if I used boneshield every time on cooldown, for example.

Tonight's log http://www.worldoflogs.com/reports/w16f9g5coc4mv57x/

02-17-2011, 04:15 PM
I think I'm going to have to agree with Reev that you're taking the wrong approach. Prove with math. Support with logs.

You're trying to find something that you don't even know what you're looking for. Try and find something in a haystack and then prove its what you wanted to find is a lot harder than proving you are looking for a needle and then finding it.

02-24-2011, 06:35 AM
I'm not sure using world of logs is the best thing for a DK tank. As far as me and my guild other tank can tell, it counts absorbs as damage taken.

If I look at omnotron, where both me and my druid co-tank are active I get large differences between world of logs and wow meter online.

For example, on our last raid in BWD, world of logs has him taking 3.2 million damage during the fight. Wow meter online has him at 2.4 million. That's a huge difference, and will be more noticeable on a death knight who is actively blood shielding as much as possible.

I know you talked about this, but it really does make world of logs data worthless for this purpose.

02-24-2011, 08:59 AM
Actually if you go into detail WoL breaks out everything, dodges, misses, parrys, and absorbs. It's all there in detail, you just have to know where to look.

02-24-2011, 11:41 AM
I've always found WMO's break down of data more easy and easier to find than WoL's. Sometimes you'll click links in circles at WoL to find the info you want, but as MellvarTank said, if you know how to actually navigate the beast, the info is there... My problem is why do they make it so difficult.

03-05-2011, 05:13 AM
Hi there,

I'm guild MT(and a blood spec DK), and i'm quite happy to voluenteer our WoL data for your research :)

We're currently 9/12 normals, and we cycle between myself and our offtank for single tank fights, so every other week you'll see a prot warrior instead of me.

My characters name is Dyzphazia. http://www.worldoflogs.com/guilds/76806/

I will periodically check this thread, but if you need any clarification on anything, don't hesitate to PM or email me :)