Tuesday, October 7, 2008

Rend and Tear

When I first saw Rend and Tear, I wasn't too impressed, at least from a tanking standpoint. You see, for some reason, in the back of my head, I've always thought of Mangle as producing more threat. Well, with all this talk of Rend and Tear being such an awesome talent for bear threat, I decided to look into it. Here's a fairly representative report of Druid threat, taken from a Brutallus fight:

Clearly, Maul produces the most threat, about 38% of the threat for the Druid's entire rotation. What does Rend and Tear do for this? Well, the damage goes, up and the innate threat stays the same. Since Lacerate is up pretty much all the time, it's like a flat 20% increase to Maul damage. In the case of the druid above, this increases Maul threat by about 16%. This is about a 6.2% increase in threat overall.

This seems like a decent use of talent points. You get about 1.2% threat per point. But, we're not through yet. You see, any self respecting tank will also take Imp. Mangle, which increases Mangle to occur 33% more often. This will increase the amount of damage done by mangle another 33% on average. Maul still goes up 16% with the addition of Rend and Tear. It's just that the impact of this is less, with Mangle doing more damage. In fact, in this case, Rend and Tear is only a 5.8% increase in threat.

We're still not through. Apparently, all the threat modifiers have changed. Let's rework this table to see what the change looks like all around.

With all these changes, looks like Rend and Tear is about a 5.7% increase in total threat. I can't prove this with numbers right now, but Mangle - Bear seems to be doing a lot more damage on the PTR than it used to. However, this is likely due to the Savage Fury change (though the damage seems higher than 20% more.) We'll leave the analysis here though.

I'm still undecided about moving those points. If a 5.7% threat increase is what you need to keep above the DPS as a bear, there's something wrong with your threat. Still, it does allow for more DPS to the boss. I suppose I will have to wait and see how my guild's DPS shapes up in the expansion.


Kalon said...

Ah - they changed Maul in WotLK; it no longer has a static threat boost and instead just does huge amounts of damage. In the beta it's actually closer to 60% of your total threat in rage unlimited situations, and it's good enough that it's probably better to miss a lacerate and a mangle and maul anyway.

Kalon said...

Oh, also - it's probably not a good choice to go from a brutallus fight. Because you're constantly switching forms or not tanking, you're doing a lot more mangle/swipe than you are mauling.

Myze said...

Can you point me to information that states that Maul no longer has a static threat component? I've been through the EJ post on WotLK and I've found no mention of it.

As for my choice of fight, I wanted to go with a fight where the Druid would likely have "infinite" rage and also where the Druid tanking was focused solely on tanking. As you can see from the original report, this particular Druid was focused on tanking Brutallus the whole fight.

While there's no doubt that my analysis is imperfect (as pretty much all analyses are,) it does well enough to demonstrate that Rend and Tear is not as big as some folks are claiming it to be.

Maul does hit for a lot more damage, however Mangle does as well, and this will effect the percentage of threat Maul contributes to your rotation, which will in turn affect the value of Rend and Tear. Rend and Tear may provide more than 5.7% threat, but until I see evidence of the change to Maul, I have to believe it doesn't come close to 10% either.

Kalon said...

While a feral is going to be on Brutallus the entire time, they're not going to be mauling the entire time. They'll either not have the rage for it, switching to cat more, or simply don't want to because of the danger of pulling off the other tank if they get an unlucky crit at the wrong time. Thus, they tend to do more swipes/lacerates than they would normally in an infinite rage situation. For example, in our kill last night maul was only 34% of my total threat. It goes up to 40-ish if you take out the cat-only things and estimate the cat white damage, but the point is that I'm simply not mauling all the time on Brutallus compared to other fights. Heck, look at the linked person's swing damage. If they're doing almost as much damage with swings as they are with mauls it's clearly not an infinite rage situation. They had 108 mauls and 120 swings. Make that 220ish mauls and you'll see a much larger percentage of maul threat.

The removal of static threat was a comment a while back from GC in the blue forums. I don't know whether there has been specific data on how much has been removed. I do know they've done that for lacerate. From testing on the PTR nightcrowler appears to believe that they've reduced the base threat from 322 to 285, not eliminated it. Also in that thread, Astrylian comments that it's better to miss a lacerate or two than a maul.

Kalon said...

I call shenanigans on that report. They're doing 20 million damage to brut in 5 minutes? The bear is doing 1500 DPS without ever switching to cat? Seems bogus.

A better example would be their KJ kill, where they had maul do 48% of their total threat and were tanking it the whole time. And unlike the Brut kill it doesn't look like they messed it up.

Myze said...

Congratulations, sir, you win.

Does that make Rend and Tear a must-have talent for bear Druids? Are we hopelessly incapable of holding threat without it?

I think not.

You'll see me sporting it come 80, but not because of the bear threat component.

Kalon said...

You very well may be right that the 2% per point is too high of an estimate. I also doubt it's required for tanking to hold threat; the big advantage is that it gives both a large amount of threat and a large amount of DPS, unlike a lot of the other talents.

But at the same time, if you're going to use numbers to prove your point, you should probably make a better argument using better data. Heck, run the same numbers through using the KJ kill data instead.