By Gravity
18 commentsI have verified some of the threat comparisons using the latest version of Kahorie’s sim. It’s had a few tweaks so I wanted to see if that refinement had led to much change in the comparative threat of DK tank specs.
I have also itemised the threat per talent of various blood alternatives.
Blood vs Unholy
I tested a mainstream Blood tank, with a survival bias in its spec, against an Unholy high-threat tank spec.
Blood was 6.4% more TPS. The gap is less than it used to be in earlier sims, when I found Blood doing nearly 10% more threat. However, in terms of threat, my comparison above is best-case for Unholy, and worst-case for Blood. So at a minimum Blood is 6.4% more threat.
Blood vs Frost
I compared the above Blood spec with a two-handed Frost spec.
Frost did 6.7% more TPS. I could reduce that gap by repecing Blood into more threat. This is a surprising result (see comments) but I cannot see any faults in the sim or settings. Take it with a grain of salt. I’d expected Blood to deal more TPS.
I couldn’t test DW Frost since there’s a bug with RS in the sim (v1.185) at the moment.
Key point, thanks Dreadskull (comments): Kahorie does not model the threat from self-healing, which can be a good proportion of your threat, depending on how much actual healing you do. Overheals do not generate threat, only the actual heal element. The TPS of rune tap and deathstrike can be quite a bit, but they can’t be relied on for sustained threat.
Blood talent threat values
Using this as the talent baseline, I then added one point of each talent to test the TPS gain and ran the sim.
Results are ordered.
- [spell]Necrosis[/spell]: 1.1% TPS per talent point (to get Necrosis, you will need to spend a point in tier 2 Unholy talents) [note this 1.1 value was before RS hotfix to stop it being proc'd by Necrosis; I now think Subversion is going to be equal, and thus the better choice due to threat-reduction when required, like Festergut]
- [spell]Subversion[/spell]: +1.0%
- [spell]Sudden Doom[/spell]: 0.7%
- [spell]Scent of Blood[/spell]: 0.5%
- [spell]Ravenous Dead[/spell]: 0.4%
- [spell]Black Ice[/spell]: 0.3%
- [spell]Morbidity[/spell] when glyphed for RS and DS: 0.019% (ie. it really sucks, this is an AOE talent not a single-target boost)
- [spell]Morbidity[/spell] when glyphed for RS and Dark Death: 0.02% (need more decimal accuracy, it was only like 0.002% more on the sim, which is below the +/- accuracy I’m running, regardless, talent still sucks for ST threat)
Blood glyphs
- [itemico size=small]45804[/itemico]Glyph of Dark Death, instead of Death Strike: 1.3% loss (ie. DS is better)
- [itemico size=small]45805[/itemico]Glyph of Disease and DS, instead of Rune Strike and DS: 1.4% loss (ie. RS is better than Disease)
- the sim uses a priority, so in a GoD sim IT and PS are barely used, so you get more HS. HS does a lot more TPS per rune than DS (like ~60% more), so my priority also has it higher in the ranking.
Learn more
You can read my detailed posts on frost, blood and unholy in 3.3, the stat weightings for each tree, rotations/priority to use and more such as ICC gearing choices or avoidance value of expertise.
All sims here used i245 stats and ICC aura for a total of 30% avoidance, which excludes your to-miss from defence/nelf/frost (ie. in this case, your actual avoidance would be ~40%).
From: 
Tags: 
feels a bit better if you can think at wow like...
Gravity: I love your sim but take it from someone who plays endgame in both Frost DW and Blood and I can garanty you that blood is supperior to frost in TPS
My frost is around 6-7 k tps while blood is easy doing 12-13k tps with normal rotations.
Frost is nice for pugging and so due to the aoe but if you have nuking dps blood owns that by a mile
Twitter: gravitydk
Yah, you have to take sims with salt. The results surprised me too, and I can’t work out why. Blood by all reports from you and others should be the winner for ST threat.
I wonder why it’s so much off.
I put a clarification in the post, thanks for reminding me of my own thoughts.
I feel that forst can do much better if the proc were increased as Getting a rime proc isn’t that often to make the TPS stable and that is one of the grand issues. It is to random to be save. sometimes you are hitting like 9k and then you fall back due to the procs are stoppping. I’m glad that we have 2 tanking builds
I use my Frost to OT and 5 m but still when I’m doing 10m as mt or in TOTC hardmodes I still prefere Blood I like never go low at all
Twitter: gravitydk
I might try to get some data parses to back-test Kahorie against real raids, that’d be helpful to verify too. I posted on dk.info to ask.
Frost is definitely proc-dependant: you only the best from it if you watch your procs carefully.
You didn’t include Necrosis in the talent values? It won’t be weighted as high as it used to be without the avoidance we had in TOC, or the T8 bonus we had in Ulduar. But, it’s bound to be better than some of the ones listed, such as Black Ice.
Twitter: gravitydk
I just checked, Necrosis is 1.1% per point. Updated post. Thanks Blue.
How dependent is Blood on armor penetration? I can imagine blood leaping away from frost by leaps and bounds when you’ve got a rouge exposing armor or a warrior sundering the same mob.
Generally, tanking gear seems to have no armor penetration at all.
Vigan´s last blog ..The end of the line of casual upgrades
Hmm, I wouldn’t be so quick to discredit your findings merely because they aren’t in line with preconceived expectations regarding the results. Anecdotally, I find these results are in line with my experience raiding with my high threat gen warlock. I consistently have less aggro issues with a well geared DW frost tank compared to a similarly geared blood tank. I think there’s certainly still a place for blood tanks, given their impressive survivability, but I’ve found that lately DW frost seems to be the build to beat for multi target AND single target threat.
Lasciell´s last blog ..Back on track
Twitter: gravitydk
@Bluedragon, an oversight. I’ll get to it over next few days, thanks.
@Vigan, you’re right: ArP is excellent for a blood tank. See these stat rankings.
@Lasciell, interesting. I am hearing such variable evidence. I want to back-test Kahorie against some real raid logs to be sure. Feel free to contribute data.
I think that baseline Blood spec used for comparative analysis isn’t terribly good for threat generation. I’d be very interested in seeing a more traditional Blood spec without Blood Tap measured up against the other contenders.
http://www.wowhead.com/?talent#jcEMVh0IcbofssxhxZ0gh:iGdmV0
Twitter: gravitydk
That was deliberate, as stated. Blood’s threat has a high variability due to its flexible options, so I tested a lower threat build.
Does Kahorie’s threat sim take healing into account? If so, what % of overheals did it assume?
Perhaps the threat discrepency is because the sim isn’t accounting for rune tap and death strike heals?
Twitter: gravitydk
Oh of course, I should have pointed that out earlier, thank you; no, Kahorie’s cannot model the threat from healing because of the overheal issue you mention.
Twitter: awfullyquiet
I believe that Dread is correct for the at least part of the threat leak that the sim is reporting… If you’re going by the results from Kahorie’s Sim for 2010/01/01… the Surivival Spec… this is all napkin math ahead…
assume 50k health… 5k healing per DS as blood… divided by 2 for healing threat base modifier, divided by number of mobs (lets assume we’re only tanking sindragosa) multiplied by our frost threat modifier (2.0735)… and we have a low amount of overheal… just suppose we get 5k healing… 5k/*1.0735 = 5.1k extra threat per DS… combined with the 3364*2.0735 threat (from attack damge), we are looking at 10k threat per DS with 0% overheal… as this is unrealistic, the number should be significantly less depending on % of overheal and % of death strikes over HS, diseases refreshing (either through UF or GoD) creates a tremendous YMMV situation… but what it does is at least give an account for what may be causing the threat leak in the model, somewhere along the lines of, 150 TPS under general circumstances, or, as listed, about a 3% threat increase…
But again, it depends on the fight… festergut, bq, and sindragosa will cause much more healing threat than on other fights… over our past few logs (constituting of the first 4 bosses, just for the low end of the spectrum) i’ve gotten around 2k effective healing average… 2.1k threat over 7.7% of attacks is an additional total TPS of 161 tacked on from the healing aspect of DS.
RT on the other hand… That’s a whole different ballgame once you factor in usage, damage, and timing… in a perfect world, it’s a good threat option of healing × 0.5 + 55, split between all mobs (this, like many other #’s i can’t remember off the top of my head were pulled from tankspot… albeit it’s a bit old, many of the numbers IIRC are still correct). If you need to use it in a close battle of threat and survivability… the threat generated is NOT bad… but, as with most cooldowns, when do you model them, and how?
Twitter: gravitydk
Testing pretty comments.
So i’ve gone through your spec comparisons and breakdowns for DK tanking as well as going through some of the theorycrafting over at EJ. One of the things i’ve noticed is that both focus on “well, you take this talent because it’s better threat”. But does threat really matter? If you were speccing to tank for raiding purposes, how much does threat really matter? There are misdirects and tricks and classes with aggro dumps. Does it really matter?
Check out my reply in Gravity’s “Blood tank spec in 3.3″. Basically, I’ve found that single target threat is much less of an issue than survivability and AoE threat. It is my personal recommendation to design your tanking build with that in mind, and give priority to AoE threat generation and survivability over single target threat.
Lasciell´s last blog ..Back on track
[...] posted an excellent point-by-point threat comparison blood talents which rates Necrosis the highest, followed by Subversion and then sudden doom. Based [...]