Jun 142010

Note: Pretty busy with class this week, so posting will be light unless something juicy comes out.

There’s some interesting new information out there in regards to weapons enchants. It’s been a while since I examined them (and when I did, I just followed what EJ said, anyway), so I decided to crunch the numbers myself and came to some interesting conclusions. Let’s take a look.

(Note: several of the weapon enchants below are PPM procs. If you’re not clear on what that means, the proc chance is the PPM * 100 / number of autoattacks in a minute before haste, which for kitties is 60 due to baseline 1.0 attack speed. This means Mongoose(1 PPM), for example, has a 1.66% proc chance. While this may seem low, it can proc off specials as well, so you generally have over 100 attacks hitting each minute.)

The Old Guard

  • Mongoose (proc, 15s duration, 1PPM proc chance, proc of +120 agi, +30 haste rating…NOT 2% haste as everybody thought)
  • Berserking (proc, 15s duration, 1PPM proc chance, proc of +400AP, -5% armor)
  • Executioner (proc, 15s duration, 1PPM proc chance, proc of +120 ArP)
  • Massacre (static +110 AP)

These four have been the general focus of discussion for most of WOTLK. The general consensus in 3.0  was Berserking as your best choice, with Mongoose being a slightly inferior option for players who both tanked and dps’ed, and Massacre slightly worse than either. (Personally, I thought that Mongoose was a better overall DPS enchant.) In 3.1, however, ArPen received a huge buff when they changed the ratings to drastically raise the amount of ArPen you gained per point.. Well-geared ferals suddenly saw huge DPS increases from gemming ArPen with an ArPen trinket, and Executioner became popular. Blizzard eventually reversed course, nerfing the ArPen increase by 50% and putting in a hardcap. (Previously, you could actually stack enough ArPen via gems + trinket procs to cause a target to have negative armor.) Executioner lost popularity, and ferals went back to the other three and continue using one of these today. Unfortunately for Mongoose, someone finally realized that we’ve been quoting the wrong numbers all along; the 2% haste was actually 30 Haste Rating, or slightly more than 1%. This changes the numbers significantly. (See below.)

The New Breed

  • Scourgebane (static +140 AP)
  • Black Magic (proc, 10s duration, 35s CD, 35% proc chance on Mangle/Shred, +250 haste rating)

With the advent of ICC, Scourgebane became popular as a minmax choice. Unfortunately, it’s useless anywhere else and for certain fights in ICC. Black Magic, though, is very interesting. While its intended as a caster enchant, it still procs off Mangles and Shreds, for some reason. I don’t think this is intended behavior, but it makes it very usable for ferals.

The List

I decided to run the numbers twice; once with my gear, once with BIS gear, to see how the numbers change. (Rawr, as usual.)


  1. Scourgebane +132 DPS
  2. Berserking +110 DPS
  3. Executioner +104 DPS
  4. Massacre +104 DPS
  5. Mongoose +103 DPS
  6. Black Magic +97 DPS


  1. Scourgebane +144 DPS
  2. Black Magic +123 DPS
  3. Berserking +116 DPS
  4. Massacre +112 DPS
  5. Mongoose +99 DPS
  6. Executioner: +0 DPS (ArPen capped)


Scourgebane is obviously your best bet, if you purely run ICC. If not, then Berserking is probably your best choice. Mongoose remains a decent option for hybrids, but its not a contender like it once was. (sigh, time to switch enchants.) If you really hate proc effects, for some reason, then Massacre is a good choice as well. The behavior of Black Magic is interesting, but ultimately irrelevant since it does not provide an upgrade over Scourgebane.

 Posted by at 8:30 am

  14 Responses to “Kitty Weapons Enchant Roundup”

  1. Alaron, did you ever post your BiS list? Rawr currently sims Mongoose higher than Berserking for me, but I never considered testing out Scourgebane. I’d be interested to see what Mr. Robot sims these enchants at as well.

  2. Shadrilin – RAWR tells me that Mongoose is the BIS DPS enchant as well. I am pushing the ArP hard cap, but not the Crit cap. Perhaps, RAWR modeling is using the incorrect Haste calculation. I would love to hear some clarity on the RAWR issue.

  3. Yes, Rawr is currently using the incorrect Haste calculation for Mongoose. Astrylian is tracking it and will fix it in the next version. The corrected Mongoose calculation is above.

    If you want to see for yourself, it’s not too difficult to edit Rawr’s EnchantCache.xml file to add other enchants.

    • Thank you for the clarification, Alaron! I’ll switch the enchant tonight.

      I am very happy I found your site. It looks great and the content is both informative and timely.


      Steamwheedle Cartel

      • That’ll do it. So the static AP gain from Scourgebane outperforms Berserking, huh? I’ll give that a try for our raid tonight. Unfortunately we’re only working on PP hardmode for our 10/12 kill so I won’t see any significant increases.

        I am also ArP hardcapped with elixir/food, so I’m looking for more crit/haste now. I haven’t looked before, but is Black Magic even included in the Cat module in Rawr?

  4. @Shad,

    BM isn’t included in Rawr.Cat because it’s not even supposed to work with melee talents, as far as I’m aware. Try it if you’d like, but expect it to be hotfixed. Anyway, cut and paste this in to EnchantCache.xml:

    Edit: duh of course XML isn’t showing up right…look in the forums.

    • Alright, thanks Alaron – I was completely unaware of the discovery that we’ve been wrong about Mongoose. I stay away from EJ threads because they’re generally a huge mess of people throwing around ideas and I stick to well-informed bloggers and wikis. Keep up the awesome blogging!

  5. You’re welcome. :) EJ isn’t bad, but the trend there is for the first person to post a reasonable argument to be acclaimed as “right,” and any disagreement seen as “wrong.”

  6. I’m not a big fan of Rawr, preferring simulationcraft which provides more believable dot uptime numbers.
    With my stuff (ICC25 + a couple 277 pieces) Mongoose is still better, even after nerfing the haste from 2% to 1.19%.
    BTW all enchants are equivalent. a 30 dps difference is completely overwhelmed by the +-500 DPS fluctuation due to RNG.

    At times I feel that people let optimization go out of hand and start swearing by numbers, losing contact with reality. 30 dps over 12k dps is a 0.25% improvement. A 0.25% improvement in RAID DPS (not personal) means that a fight which lasts 300 seconds ends up lasting 299.25 seconds….

    And this is assuming that:
    – simulations or models are exact (= no bugs)
    – simulations or models represent reality (= false as soon as you have adds, brez/innervate, etc.)
    – you are able to perform as well as the simlation/model (= keep dreaming, or start botting)

    Conclusion: pick the enchant you like from the list (choose on visual effect if you want), it’ll make no difference whatsoever.

    • So, you prefer SC over Rawr because it’s more accurate, but then don’t care which enchantment you use? :) If you’ve modified SC’s code to correct Mongoose’s behavior, I’d be interested in seeing what you did. I took a look, but unlike Rawr where data is exposed in XML files, it seems to be all hard-coded.

      Yes, on single fights, the fluctuation of RNG will obscure the DPS advanage of one enchant over another, but that doesn’t mean you can’t optimize for the best long-term performance. (Or the best glow, if you want.) I mean, the difference between an epic and a rare gem is only about 10-15 DPS, which is also not visible due to RNG; would you advocate sticking with rare gems?

      • Yes, I prefer simulationcraft because I find optimization to be a game itself and have fun using it. I’m also very well aware that however optimal my dps set is, if half of the group gets gibbed by some boss mechanics then it doesn’t matter at all. If you want to see the light :P keep track of the causes of your group’s wipes: you’ll see that lack of 0.1% dps is very rarely one :) This is why after the first down the combats become a lot easier. It’s not like everyone’s dps went up, it’s that people learned how to deal with the mechanics.

        For mongoose: you have to change the code itself: the function is in sc_player.cpp, lines 1264 and 1265, the 1.0 + 0.02 should be replaced with 1.0 + 0.0119 (at least if the data on EJ is true).

        As for sticking with rare gems: if it’s ONE gem, it won’t make a difference. If it’s all of them and reverting all your equipment to ilvl 213 then you can be assured that the difference is a lot more than the RNG. Upgrades are incremental: one of them will be negligible, but add 20 of them and the result won’t be negligible (expecially because some stats stack non-linearly).

        A more direct comparison: one shred is (on average) 15k damage. Divided by a 300s fight, that’s 50 dps. So one single more OOC proc is already three times more than the difference you report between mongoose and berserking…… or, to put it differently, wasting an OOC proc is three times more than choosing your enchant…..

  7. That’s interesting, I’ve been using Massacre for a long time based on Rawr’s suggestions. I’m betting that may have more to do with my shortage of hit more than anything. More misses = less chances on procs?

  8. @Heli-

    I think what you’re saying is to keep changes in perspective, based on how much of a DPS hit it is (i.e. don’t agonize over ArP gemming vs Agi gemming, for example, but do agonize over bad SR uptime.) I fully agree. I think we do get a bit carried away with optimization sometimes; however, like you said, it IS fun. I want to provide info to help people play better; what I DON’T want to do is say “gem X, enchant Y, and equip Z, or you’re not optimal” and cause somebody to get rejected on a guildapp because they’re not “optimal.”

    Also, thanks for the mongoose info, I’d been looking through enchant.cpp.

    • Yes, that’s the point I want to make, but actually going a bit deeper. Optimization is an answer to a question, the problem is that often people forget the question completely and only focus on the answer.

      If the question is “what is the way to obtain maximum dps on a training dummy”, then information which provides even +10 dps is important. But more often than not the question is “what should be done to down this boss”. In this case, discussing +10 dps is irrelevant, as there are so many other variables which are involved. Unless statistics on raid wipes clearly shows that the +10 DPS will help in downing the boss, then caring about +10 DPS is irrelevant. This is why a good raid leader does not care about individual class details, unless it’s affecting wipes. And if you look at causes of wipe in ICC, DPS is rarely an issue: survivability and movement is a lot more important. This is why a non-dps-optimal Tuskarr’s Vitality enchant can help a mage, for example. On our first down of Blood Queen, if I remember right, I was using PvP Trinket + Corroded Key instead of DBW + Herkuml. Is it a dps loss? Sure. But we had the dps, the problem was people dying in the air phase. So “optimization” in this case was not increasing my dps, it was increasing my survival.

Leave a Reply