A few of us in guild have been discussing exactly how much demoralizing shout the like works; this thread gave a little hint about how that it might be a % dps reduction, but no real parses to back it up. So we did some testing:

Methodology:

Beast lore is a hunter spell that allows you to see the perinent stats of any beast. These stats include ac, hp, and most importantly their unmitigated damage range.. Using this, we won't actually have to spend time parsing against mobs to figure this out, we can just beast lore everything and see.

There's basically four spells that all reduce -atk (demoralizing, screech [owl pet skill], scorpid sting, and vindication), and one that obviously increases it (cor). We tested by using various combinations of cor scorpid and demoralizing (ie no vindication). In some cases, screech was tested instead of demoralizing, and the change was interpolated for demoralizing (by multiplying by 204 / 100). I tested to make sure that this interpolation is valid vs. a plaguebat (did it with demoralizing, did it with screech, and found the interpolation matched).

Tests were done with untalented scorpid sting (68str/agi debuff) and demo shout with imp demo 5 (-208atk).

This is about half the data I have, but it shows the interesting parts.

%ScorpRed is the % dps reduction from Scorpid sting. %Demo for demo, and %Combo for both Demo and Combo combined.

From the data a few things could be gleaned:

There is a cap of the effect at 30%. I have more uninteresting data of in-between levels (24-26 and 52-60) that all show that the %reduction of all effects combined cannot go above 30%. This is interesting, but not terribly surprising by looking at how pets work, roughly 30% of their dps is from their "Attack" value, the rest a base based on the mobtype. This cap does not seem to pertain to thunderclap/tf, as I thunderclapped a mob then demo'd it, and the mob's dmg range went down by the appropriate amount.


Demoralizing shout really isn't % based, but scales in some other manner. The lower spectrum of the parses are showing that demo won't drop dps more than 30%, because presumably that's "taking away more atk than the mob actually has". However the change from epl trash mobs to zg trash mobs shows demoralizing still giving a huge dps reduction, showing that it must scale in some other manner.

It has been suggested that raidmobs have fairly normal stats, and are just transformed via a "damage multiplier", and thus dmg reduction from -atk also goes through this dmg multiplier. This data seems to support it to some degree. However this theory also means that emps really aren't special, just we started noticing it here becuase 20% of 10bazillion (ubs) is a sizeable amount. Every mob tested so far exhibits this 20% melee dmg loss.

The obvious extension to the data would be to test demo/screech on The Beast, magmadar, and chromaggus. The formerly mentioned post says that antecdotally it works on Chromaggus, but using beast lore should be able to give a definitive answer. Unfortunately I have no plans to go anywhere near MC or BWL for the rest of the uh, century, so hopefully someone else can test these, as they can be tested basically mid-raid (beast lore is non-aggro, though it's range is fairly low and dogaggrorange is fairly high).


Some other things that I've found that I'm too lazy to post the data for:

Demoralizing shout and screech stack. Up to the 30% cap that is. I guess this was obvious already (since the two icons stack), but since I've never seen someone suggest screech or really demo shout for any of the "I'm getting one-shotted" threads (other than emps), I guess it's not something widely known.

Imp demoralizing shout is pretty huge. The -atk effects are linear (I tested this when testing the interpolation method), which means the dps loss is sizeable on any mob in which the cap does not play a role. For example a mob with 18% reduction via demo would have a 25% reduction with imp demo, that's kinda a big difference. I guess we knew this too, but again, not terribly widely known.

Also since the atk formula is linear, you can see that using curse of wrecklessness will increase dmg by a sizeable amount, but I guess we sorta knew that too.

Maybe all of this was all widely known already, donno. But atm there seems to be really no excuse for not having demo shout up on every mob, which is why we did these tests, to see if we need to start yelling at the warriors again.

Chromaggus is a dragonkin, but some very interesting info here. The testing seems sound but it doesn't line up with what I have noticed, and many other guilds including EJ have noticed, about how much CoR hurts on a boss. Obviously things like mortal strike and unbalancing strike are red flags against using it, but is it actually doing more harm than good? I don't know the answer but from this testing I can't help but be skeptical about its use anymore.

Well the one thing you have to consider is that we're only talking about melee damage here. In many fights (mostly bwl fights) you're getting a huge chunk of the mt damage from shadow flame etc, which cor (and demoralizing shout etc) isn't gonna change. Presumbly there'd be a sizeable dps increase on broodlord (his MS is multplicative, like 300% of wpn dmg), but honestly he's been nerfed so many times, his MS is way manageable no matter what for guilds past nef.

I was more interested in the mobs that currently hit like trucks (ouro, enraged kri, somewhat the emps). Even presumably patchwerk wouldn't fall under this category, since hateful strike is fixed damage and not based on wpn/atk.

Well, if you use demoralizing and screech, which should be more than 30% but is capped, and then use curse will you still be at 30%?

Well in theory you wouldn't be a 30%, you'd be around 26% ish probably (you gain 100 from screech but lose 90 from cor, so it'd basically a wash).

I'll see if I can get a warlock account to test it. One thing to note however is that screech is only 4 second duration, it's quite spammable and you can have multiple hunters using it, but who knows how well you can actually keep the debuff up with x number of hunters.

One pet can keep Screech up if the pet can actually stay alive. Soloing with my hunter I don't think I've ever seen the debuff drop while my pet is attacking a mob.

Great information. Thank you.

Edit - D'oh, didn't see the last reply date. Ah well, it's great information anyway.

Gonna post a summary of the test i did, all listed hits are from mobs white damage only and recorded by Sabrians CombatMonitor. Blocked hits are recalculated into average hit from block value and partial block count but in general its faster to test while turning your back to the mob and taunting when you loose aggro to healer.

Mob: Hederine Slayer lvl60 elite (Winterspring)
Listed mitigation: 63.9%

Without demoshout:
Hit count: 237
Average hit after mitigation: 241
Average hit before mitigation: 668

With demoshout (-146 ATP):
Hit count: 238
Average hit after mitigation: 188
Average hit before mitigation: 521

Hence demoshout made the average hit 668 - 521 = 147 less, before mitigation.
That is, ATP reductions on this mob have a value of 147/146 = 1,01 damage/ATP on average before mitigation.

However, well known formula for how ATP affects players state that 14 atp = 1 dps, so in the case of the Hederine Slayer which seems to have a standard 2.0 attack speed it'd mean 0.14 (2/14) damage/ATP.


Conclusion: Even tho hit counts should be trippled to make a correct determination of the mobs damage/ATP ratio its _extremely_ unlikely this mob got a standard (14atp = 1dps) conversion since the result was off by a factor 7 (1.01 vs 0.14). Hence, its probably pretty accurate to state that this mob scale somewhere around 6-8 times better then it 'should' with ATP and not too farfetched to see raid bosses with a scale factor of 20 or more.

Edit: Worth noting is following. 147 less average hit works out to be a 147/668 = 22% reduction in dps. Had i used imp demo shout seems reasonable to assume i'd reported 22*1.4 = 30.8% dps reduction which is atleast in line with the original posters findings.

30% damage mitigation hacks...

I know on patchwerk it cuts me from 2500 to 1800ish... But hey I'm being heavily anecdotal. This is full IDS.

Yeah, the second Demo drops off Patch the MT starts yelling for reapplication because it's a massive change.

Really interesting numbers. Confirms what has become more and more apparent to us over the past half year, but it's nice to see numbers backing it up.

I remember first being amazed when two identically geared and specced tanks tanks Twin Emps, one using Demo on Vek'nilash whenever possible, and the other not, and the latter was taking easily 25% more damage total.

On a semi-related topic, has anyone tested the effect of insect swarm? From what we've seen with little real testing, this debuff does near nothing.

Oh the topic got bumped.

Wasn't going to post these cus I thought it was old news, but since the thread got bumped, here we go:

Beast lore was used to determine the min/max damage of the mobs. The data was taken when neither frenzy (magmadar) nor enrage (maexxna) was up.

Magmadar:
Normal: 1867 - 2476 damage.
W/ demo: 1512 - 2121 damage. (-16.3%)
w/ demo+cor: 1731 - 2340 damage. (-6.2%)
w/ cor, no demo: 2085 - 2694 damage. (+10.0%)

Maexxna:
W/o demo: 5491 - 7281 damage.
With demo: 4448 - 6238 damage. (-16.3%)
(We don't cor it).

I'm going to assume these are untalented demos, but I'm not 100% sure. The magmadar was an open raid, and Maexxna has an extremely annoying hitbox so basically all our warriors spam demo all the time (for both small shits and maex).

Honestly it's been on my list of things to do, but just haven't been bored enough to try it. Anecdotally I'm not seeing a difference in raids though, which is why I've been too lazy to actually test it.

Speaking of open questions however: I really am curious on wtf improved scorpid sting does. The 30% demo cap applies to scorpid sting as well, so why on earth are all raidmobs flagged immune to scorpid? My guess is it either has to be the agi affecting critrate (the affect on ac is minimal -- 1agi=1ac), or that the sta suffers the same multiplication as atk (pure conjecture). The latter, if true, would move imp scorpid from the "incredibly worthless" pile to the "incredibly useful" pile. Any ideas on this? As with all Naxx hunters, I'm currently broke and don't feel like burning 40g on respecs.

My guess: An oversight no one bothered to fix. The stamina reduction with 3/3 improved scorpid sting is 20.4 - and I doubt it scales with anything since it only relies on the amount of strength reduced by the scorpid sting. I'm not 100% sure if it doesn't scale though (I wonder with what it would scale).

You're thinking of scaling in the wrong terms.

The way the -atk from demo works (and thus the -str from scorpid works), is it's assumed that a mob has X str and Y atk, and then his dmg range is created by the multiplication of that effect by some number we'll call the "bossness factor", for how bossy the mob is, and thus how hard he really should hit.

What is to say that sta is not multiplied by this same bossness factor? If this were the case, then imp scorpid sting would reduce a mob's hp by a flat percentage, not by 204 hp. Again, this is pure conjecture, but as we've seen above: npc stats do not work the way we think they do.

I guess I may have to grind gold this week if there's no independently wealthy hunters around.

I think that the effect of combined attack power reductions is capped simply because you can't reduce attack power below zero, and most level 60 mobs seem to have a static 238 attack power (for example, when I checked this with a level 60 warlock, all pets but the imp had identical stats). Assuming that attack power is responsible for about 30% of a mob's damage, this would explain the results pretty well.

Strength (possibly agility?) reductions are likely to affect attack power also, but will again be limited by not being able to reduce attack power below zero. I have not been able to come up with a consistent formula to calculate attack power from strength/agility, though. Here are some examples, in case somebody else wants to take a crack at it:

It would be interesting to know if Curse of Recklessness is applied before or after Demoralizing Shout/Scorpid Sting/Screech. If before, its negative effects could (at least in theory) be neutralized completely.

Note that you can use UnitDamage(), UnitStat() and UnitAttackPower() on enslaved demons (even though they don't have a pet screen), so it should be possible to do some more extensive testing by enslaving some level 60 demons with and without a debuff on them to figure out more about how debuffs affect damage and attack power.