Thursday, January 15, 2009

[Druid] The threat value of improved mangle

In an earlier post I talked briefly - hinted, really - at whether improved mangle is really all that now that swipe has been so significantly boosted. What I found out was pretty surprising to me.

Yes, I'm using the teaser trailer. Bear with me. :)

Pre swipe getting boosted, the best threat rotation for a bear was mangle, lacerate x 2, with a faerie fire thrown in whenever it was off cooldown or when it was applicable (stupid misery bug...). This meant that in an 18 second window, you would mangle 4 times and lacerate 7 times and faerie fire once (hopefully), along with 7.2 mauls. This would result in the following, assuming AP of around 5k, a crit rating of 40%, and the standard talents you'd find in most tanking builds:

Damage(mangle): 1644.7
Damage(Lacerate): 91.08
Damage(Maul): 2139
Damage(lacerate DOT): 815.1

Threat (mangle): 3406.9
Threat (swipe): 1578.763 (post 50% boost)
Threat (lacerate): 1162.154
Threat (maul): 5309.797
Threat(FF): 1933.057
Threat (Lacerate DOT): 844.21

4 mangle/7 lacerate rotation: 66991 threat, or 3721 TPS.

If you take out the improved mangle, you instead get a 3 mangle/8 lacerate rotation. Basically every 18 seconds you go to 65746 threat, or 3597 TPS. That's a loss of 124 TPS. That sounds substantial until you look at it as a percentage: it is only 3.35% of your total threat.

But that's not the rotation any more. The rotation for best TPS is now swiping whenever the cooldown is up and throwing in a faerie fire whenever the cooldown is up.

This assumes using the same 5000 AP and 40% crit from above.

Using the same methodology above, we have the following
4 mangle/7 swipe/1FF rotation: 64842 threat per 18 seconds, or 3602 TPS
3 mangle/8swipe/1FF rotation: 63014 threat per 18 seconds, or 3500 TPS
Net loss is 101 TPS, or only 2.8% total threat.

2.8% threat really isn't that big a deal in the long run. It essentially means that improved mangle is worth about 1% threat per talent point, which puts it ahead of master shapeshifter but behind rend & tear. The above is also the best case possible; it assumes you're mangling every time you can, that you're not using berserk (which helps the non-improved mangle case more), and it discounts rage/healing gains for threat.

The first surprise that you might have is that these numbers show that lacerate, not swipe, is the best threat. But...didn't I just say the opposite a couple weeks ago? Yep. The main reason is that that calculation assumed a 30% armor reduction, which is a bit unrealistic. 40% is closer to the mark. With that and a low-ish AP value (5k is pretty low for raiding tanks), lacerate wins out on threat. If you instead have 6k AP, swipe wins out again. So I was still just means that if you're on the lower end of gear, spamming swipe for threat is not as good as lacerating.

Okay, you ask, what about the best damage case? Since a lot of people seem to care how much damage a tank does, we'll go through with that. Using the same 5k AP/40% crit, the cycle for optimal damage is something like:

mangle whenever it's up
lacerate to keep the stack up (once every 12 seconds)

So in a 36-second cycle (we need it to be 36 seconds to fit in the lacerates), we have the following for max dps:
8 mangles
3 lacerates
13 swipes
14.4 mauls
+36 seconds of lacerate DOTs (12 ticks)

The damage for that is:
60624 damage, or 1684 DPS

For the same cycle without imp mangle, we'll have 6 mangles, 3 lacerates, 15 swipes, 14.4 mauls, and 12 ticks of lacerate. That damage is:
58351 damage, or 1620 DPS

The loss of DPS: 63.1, or 3.74%.

That was really surprising to me. It does mean that improved mangle is a really big bang for your buck in terms of overall damage - it's better than the 1% damage/talent point that the devs have stated is something of their goal, for instance - but it's not so big that you can't live without it. And do note that as AP increases, both the threat and DPS loss from imp mangle also increases. At (for example) 10k AP you're looking at a 3.9% DPS decrease and a 2.9% TPS decrease. Which isn't earthshattering, but it is something to note.

To me, this solidifies what I've been leaning towards for a while: improved mangle is a nice-to-have but hardly an essential talent. And it's one I'm going to drop, in favor of picking up more cat-friendly talents.

Curious about my numbers? The spreadsheet I used is over at google. The numbers were from EJ's feral values here and here.


felkan said...

Replace iMangle with something else to boost cat? Isn't iMangle great for cat (assuming you don't have T6)? Or is Predatory Instincts better than iMangle for Cat?

Using lacerate in your rotation boosts Maul if you have RnT? I thought you didn't max out RnT in your build? Maybe I'm thinking of another blogger...

So, what build are you thinking about using if you skip iMangle?

Kalon said...

Hey, Felkan - thanks for coming by. :)

iMangle isn't that great for cats. 6 energy off of an attack that ideally you'll use once every 18 seconds or so (assuming no trauma) is not that ideal. Predatory Instincts is about double the value imp mangle is.

But shredding attacks is nearly triple the value, and only costs two points.

Lacerate does boost maul, but so does any other bleed - and it's almost always guaranteed that deep wounds will be up on a single target boss. If not, you'd have to throw in some lacerates instead of mangles, and the rotation would look more like the 36-second rotation I posted. It ends up not mattering that much, and the TPS difference is constant since you'll be trading a mangle for a swipe in either case.

I do max out RnT right now - but I'm very tempted to change. RnT on two targets with lacerates going (or deep wounds) is really great threat. I'm tempted to go to something like this build. However, I think I'll miss the damage and threat from mauls too much. It's either that or this one. PI is a great DPS boost for cats, but RnT might as well be essential for bear threat.

I'll play around with it and see some.

And I'll be using your mail for my next article.

Anonymous said...

Thanks for the analyis, thats very helpful. I've also been playing around with maul/lacerate/rnt/swipe, to see how effecting maul is with a lacerate on the mob. I've concluded that for trash, its worth putting a lacerate on each mob when possible, mauling on every attack, and swiping/mangling in between lacerates. My maul is around 58% of total damage at times (raid trash + bosses). Does this sound accurate?

Karthis said...

My gut feel for the last week or so has been that Imp. Mangle wasn't all that essential for a bear, but I was too lazy/busy to do the math - thanks for running the numbers and confirming my suspicion.

I nearly dropped it before my raid last night, but ran out of time to get to a trainer for a respec.

Megalis said...

Thanks for doing the math. It verifies what I had sort of assumed. That is that Improved Mangle just isn't worth the points.

I wouldn't worry about dropping two points from Rend and Tear. I was running a spec that way recently and it didn't feel any different threat-wise. And I was still out DPSing most of the newer 80s I ran heroics with. As the tank.

I've been playing with different specs recently to try and balance out my mitigation with my damage. After I got all of the abilities that I really wanted for tanking I had one point left and at the moment it's sitting in Improved Mangle.

However, after a very poor performance DPSing a heroic last night (as Cat) I decided to move it - as well as one point out of Rend and Tear - to Shredding Attacks.

Paying 60 energy for Shred makes it really hard to put up decent numbers. 42 is much more reasonable, not to mention the discount on Lacerate.

This is the spec.

Cuer said...

Improved Mangle is useful for all kinds of feral, but not a "must have" for any. I can't imagine not having it now, though, after getting used to it. Between soloing and PVP, and boosting bear threat, I find it's a handy talent all around. For fast fights (i.e. all the trash in raids) it can be nice to have Mangle ready again so quickly, for that extra TPS boost.

I've ended up using a build like the first one listed above; I take the full 5 points in Furor for PVP purposes, but otherwise identical. It works solidly for all roles: main tanking anything in raids, doing some DPS when needed, and the occasional PVP bit of fun.

Flyv said...

There are folks who say Improved Mangle is great for cats. In particular, some DPS models show Mangle-spam being the same or higher DPS than Shred-spam if you spec for it. I don't play this way myself, and Shred seems to be scaling better for me with my gear.

BTW, you left Maul out of your threat calculations. It's my #1 DPS and I'm guessing my #1 threat, too. Hopefully someone else in your group is applying bleeds, but if not you want to use a Lacerate every once in awhile to get the 20% damage bonus.

Kalon said...

Megalis - hey. My DPS rotation was basically manglespam with an occasional Ooc-fed shred. I got around 3k in my very poor DPS gear. :) That being said, shred appears to be much better than forecast before when you start getting good gear, so I'm going to try that again.

Cuer - after playing with it I'm almost positive I never want to play without KotJ again. It's just so nice for cats, especially solo.

Flyv - I'm pretty sure that I had in maul for the threat rotation. The amount of mauls doesn't change from rotation to rotation (it's always 7.2) so I didn't put it in the rotation's name directly. It's just a flat constant anyway; both assume that you'll have mauling going on, and both assume the same damage. As to me - in general I don't bother going for max threat (though there are at least two warriors putting on deep wounds) and so always have lacerate going, as that's the most DPS. The exception is Malygos25, where lacerate stacks fall off anyway and max threat is most important. There, it's swipe-city.

And the calculations that have manglespam as the highest DPS appear to be wrong. Rawr thought that way for a while, but miscalculated shred's damage. Manglespam with imp Mangle and 2pT6 is good for lower-end gear, but it ends up quickly going under compared to shred with shredding attacks. It's serviceable - as I said, I can get to 3k with manglespam and that's not embarrassing - but it's definitely not on the order of what good dps can do.

Haunau said...

See, I've gone the other way -- I took Imp. Mangle not so much for the tanking part (though that was nice) but for the cat side, so I can foresake Shredding Attacks and R&T and I MANGLE YOUR FACE MEOW.

Yes, it's not a great trade for Shredding Attacks straight up, but with the points I'm getting back by not taking R&T, I'm getting stuff like Infected Wounds and Primal Tenacity and Brutal Impact -- all those kind of iffy tank talents you can't squeeze into a 2-way build.

And my DPS isn't awful. I run heroics all the time, and did DPS in Naxx-10 last night with a good showing. I know it would be better with Shred, but I want those tank goodies.

At least for now, maybe when we get to the next gear set, Shred will pull far enough ahead to make me reconsider.

Megalis said...

For the record, this is my DPS rotation (assumes no helpful spells/abilities from anyone else):

Faerie Fire -> Rake (for 1 point) -> Savage Roar -> 1 Mangle (increases Rake and Shred damage) - Shred to 5 points -> Rip.

If I get several crits in there and I get back around before Savage Roar/Rip are down, I go for a Ferocious Bite.

Idol of the Ravenous Beast and Glyph of Mangle help out here.

Phil Jackson said...

Funny that the day I start questioning my MT spec Kalon posts something that makes me want to test other builds even more. Even though I MT most of the bosses that we are progressing through, there are times when my dps in cat form is better for the raid than the OTing Warrior or Pally. I'm trying to get a build that I feel comfortable with that let's me get some more cat talents without losing the extra tanking talents I've gotten.

This is my current build. I'd love to get KotJ but I don't want to lose points in FA until I feel we have the raids on a more Farm status (no fury/arms warrior so sad T.T).

Well this just lets me know that I'm not done with what I can learn about my class.

Shamad said...

I actually already dropped iMangle for KotJ some weeks ago, more so because of the coming mangle fix/nerf, but also since I had a feeling the benefit was something I could live without. Whenever we do get dualspecs I'll prolly put the points back in as I hope to never again be asked to dps on my druid, but for the time being, it just makes sense to have the bear go cat when you don't need the extra tank.

Marino said...

Hey Kal,

I am blind. Either that or I am missing the TPS when you do lacerate. I only see DPS stated and you refer to it as doing most damage, not most threat.

To put it simple I am thinking about the following:
Every 12 seconds you trade in one swipe for a lacerate.
Using your numbers that is -1578 and + 1162. You also get 4 lacerate dots. 4 x 844.
that is an increase of (1162 + 4 x 844 -1578) = 3260 which is an increase of 272 TPS.
That is not even marginal, that is huge! Then why is keeping up lacerate in a swipe cycle not in your TPS calculations?

doug said...

I made this change last week to boost my cat dps. I wanted to get OOC and shredding attacksx2 so I had to evaluate what talents I could sacrfice and iMangle fit. In support of this I switched idols to the new wotlk badge idol.

As stated in earlier comments, iMangle is great for trash, pvp and soloing. But for a raid build, if I need to go kitty instead of bear, it just doesn't cut it.

Since I'm seeing more kitty action in this release than before I needed to maximize my dps as much as possible without sacrificing crucial tanking talents.

I just wish I had one more point. I'm with Flyv on nurturing instinct and wish I had two points to spend there instead of one.


Kalon said...

Hey, Marino -

I went back and checked out my numbers...and they were right. But they assumed a much higher swapping value for lacerate, but really, I think that there were some wrong assumptions all along the post.

If you can reliably swap a lacerate for a swipe once every 12 seconds, then (in a 36-second interval) you are trading 3 swipes for 3 lacerates + 12 DOT ticks - which does end up being higher TPS, and is so at basically every amount of AP, crit and armor that you can think of. It's about 5% more threat. So you're right - that cycle is the best, assuming that you can reliably keep a 5-stack of lacerate going.

And this is true for all sorts of values in between. Even doing a rotation of simply 2xswipe, 1xlacerate,1xmangle is better TPS if you can keep up that lacerate DOT stack.

Basically, I think that a lot of my assumptions on the math were wrong earlier. You can't compare the TPS to the threat in a cycle directly; you need to figure out the cycle first, then calculate the TPS from it. Between that and the armor thing I think that covers the error.

The long and short of it is this: the best threat you can produce does involve keeping up a lacerate stack. Swipe alone is not good enough. However, the more swipes you add to the rotation the higher threat you will produce, assuming that your stack doesn't fall off.

And if you're in a situation where your stack could fall off, you should almost always do swiping instead.

I'll go back and amend my post. Thanks for catching me on that. This is precisely why I publish the math. :)