Gear optimization – at your own risk

24 Jul

Don’t get me wrong. Programs like Rawr, AskMrRobot, Reforgenator, and the like can be wonderful helpers if you need some guidance, but I want to strongly urge caution to people who use those. They are not the be-all, end-all for everything. Use your own common sense and decide on what your particular needs are, as well as your groups needs before blindly taking the advice of one of these optimizers.


We will take my own character as a reference point. Just a few days ago I decided to plug my character into AskMrRobot (AMR), and see what it came up with. Not because I was unhappy with my performance, but I always believe there’s room for improvement. If you look at my socketed items, you’ll see that I have a lot of pure agility gems in my gear, whereas before, all of those used to be agi/stam gems. There is a big part of me that still believes stamina is important as a tank, but I DO know better than to stack Stam to the exclusion of all else. When the optimizer suggested I replace those 6 hybrid gems with pure agi gems, I was a little caught offguard at first… then I thought about it for a while. Is it worth getting an extra 120 agility, at the loss of 180 stamina? As a bear, agility is our highest rated stat, because it affects both our survival via avoidance, as well as our threat generation via attack power. I kind of agonized over this for a while, but when I boiled it down, 180 stam equals out to ~3500 health in bear form. Maybe closer to 4500 with raid buffs. In the grand scheme of things? 4500 health really isn’t all that much. I can’t honestly think of a fight i’ve been in where 4500 health would’ve made the difference between life and death… or to perhaps phrase it better, if I was down to 4500 health, I was dead anyway. So, I said “why the hell not” and decided to regem. When we raid this coming week, i’ll let you guys know if I felt any difference, and whether or not I decide to keep these gems. I’m not expecting any mad desire to change back, so unless things go HORRIBLY awry, i’ll probably stick with this for a while.


After feeling pretty good about my re-gemming decision, I decided to look through some of their other suggestions to optimize my gear. It suggested that I replace some enchants (it wanted me to switch out agility to cloak for armor to cloak, for instance), or reforge a couple pieces differently (since Mastery is our 2nd highest rated “green stat”, it suggests I go with the lower stat-amount items to reforge. It would have only made me lose about 30 dodge rating… MAYBE .1% dodge, so meh… nothing i’ll lose any sleep over). One thing that really got me was it’s BiS list. Or maybe not so much the actual BiS listed items, but the items this program specifies as direct upgrades over what I have. I’ll give you an example.

I was lucky enough to pick up the first Ranseur of Hatred trash drop to upgrade from my 359 weapon. That’s a damn nice polearm… even if haste is pretty ugly for bears, the extra agility and stam on it make it so much of an upgrade, that green stats really aren’t that important. It’s an upgrade just for the base stats alone. Now then… AMR suggested that the Witch-Hunter’s Harvester is a direct upgrade over my current weapon.

Wait… what?

I’m sorry Mr Robot, I must strongly disagree. I think I would rather suffer the 252 worthless haste on the Ranseur than lose the 47 agi, and 70 stam. Not to mention skipping the plethora of ANNOYING ASS crafting materials i’d need to gather to make that bloody polearm. Why does it say the Harvester is better? Because mastery is rated higher than haste. Haste is worthless for bears, and mastery is our 2nd best green stat (behind dodge, which doesn’t come baseline on ANY of our gear, we just have to reforge into it). BUT… let’s look at it this way.

If I reforge the haste out of the Ranseur (which I did), it gives me 100 dodge rating. If I reforge the crit out of the Harvester, it gives me 96 dodge rating. I also get more dodge from the 47 agi on the Ranseur, so my overall dodge is STILL going to be higher, even if my mastery is a decent chunk lower. Is 241 mastery really worth 47 agi and 4 dodge rating? Not to me. You might disagree with me. I can definitely see, to a point, where it COULD be considered an upgrade. My whole point of this post is to take these gear optimizers under advisement, but DON’T just blindly trust everything they have to say.

After all… They put 130 agi on my weapon over Windwalk. /boggle.


12 Responses to “Gear optimization – at your own risk”

  1. caelcraft July 24, 2011 at 1:17 pm #

    Good piece of advice that I actually haven’t seen around before. Simcrafting and stat priorities are certainly good to have, but should definitely be taken with a pinch of salt and a huge chunk of common sense. More people need to be aware of this.

  2. Yellowfive July 24, 2011 at 2:13 pm #

    Hey — I am the main developer for Mr. Robot and the author of our bear stat weights. I think that I can shed some light on why the optimizer is making some of these choices:

    For Witch-Hunter’s Harvester vs. Ranseur of Hatred: if you look at the scores, they are VERY close for these two items. If I rank them absolutely, they are within 1 point; if I rank them relatively for my own feral druid, they are within 6 points. For all practical purposes, the items are identical according to the default stat weights. The list has to put them in some kind of order, so of course it will put the one that is 1 point higher on top, but I would not recommend replacing your Ranseur with an expensive crafted item to gain 1 point!

    The default weights consider only total damage reduction — threat is ignored because it is irrelevant for most raiding content. That is why the higher-damage Ranseur doesn’t get any special consideration — weapon damage is ignored for tanking.

    The score is just a math problem based on the stat weights. We automatically apply diminishing returns, so stats like agility and dodge will drop in value as you get more, whereas stats like mastery, crit, and expertise will become relatively stronger.

    As for the weapon enchant: we have simulated the up time on windwalk using our own proc simulator. It does many iterations, and averages the results over fights of durations from 4-8 minutes to give you a true average up time. It also uses parameters specific to each class and spec, thus giving an different and more appropriate estimate for each player. Our estimates are far more accurate than most of the generalized calculations used at places like EJ in their trinket lists.

    That said, Windwalk comes out to an average 245.42 Dodge Rating for bears. With the default stat weights on a tier12 character, the score for Windwalk is ~130, and Agility is ~107. So yeah, we should probably add Windwalk as an option… I think we just forgot to put it in the list for bears.

    Note that the value of the two enchants is very close though — this is because of diminishing returns. A 600 dodge proc suffers a lot more from DR than a 130 agility enchant because you get a bigger chunk all at once, and for agility, only a fraction of it is diminished: the bonus AP and crit do not diminish (and contribute to improving Savage Defense).

    We have a 2nd preset called “off-tank/hybrid” weights. They favor mastery/expertise/crit over dodge rating. With those weights, 130 agility is the superior tanking enchant.

    As your gear level improves, reforging to Dodge Rating becomes less and less effective, and agility, mastery, expertise, and crit become more effective for damage reduction, unless the boss is hitting VERY hard, which you usually don’t see unless you are doing heroic modes. Try out with the Gear Comparison fight style to see if reforging for Dodge or for Mastery/Crit would reduce your total damage taken on your particular character by more. You can adjust the settings for how hard the boss is hitting to match your current content (that is the most important parameter).

    (Note that we wrote the simulator to do mastery/crit for a SD-focused gear stategy, but we should probably change it to mastery/expertise, since expertise is very slightly better than crit for improving your SD at higher gear levels. The difference is small enough that the simulation results should be about the same for the purposes of deciding which gear strategy to use though.)

    • achloryn July 24, 2011 at 2:49 pm #

      First off, let me say that i’m absolutely flattered (and a little boggled) that you took the time to reply to my blogpost. Holy shit yo. Seriously LOL.

      Secondly, I didn’t mean any offense to you or any of the other optimization tools that are out there. You guys all do amazing jobs doing far more math than I could ever wrap my brain around, and I hope you didn’t take what the point of this post as a “bash” against you at all, cause that was definitely not my intent.

      My point in writing this was merely that, doing ANYTHING blindly, simply because someone tells you so, can be bad. Understanding why things are setup the way they are is equally important. For bears, mastery may be our 2nd best green stat, but mastery is still relatively meh for us, especially when compared to the other 3 tank classes. It’s nowhere near as effective, per point, for damage mitigation as “blocking” ends up being, and it doesn’t scale nearly as well as gear levels improve.

      I’ve known that some optimizers to be more accurate for some specs, and less so for others. I’m not doubting your math by any means, but at this point in Cata things seems far less definitive nowadays than they were before. There may be a “most effective” route mathematically, but there seems to be less disparity between “most effective” and “viable” than before.

      • Yellowfive July 24, 2011 at 3:22 pm #

        Yeah I didn’t take your article as “bashing” — it’s pretty tough to make a program or an algorithm that can handle all the nuances of the game. We try to get close, but in the end… all you can expect to do is get close!

        Mastery is definitely not as strong for feral as prot warriors and paladins, but that is offset by the fact that our primary stat (agility) is far better than strength is for them.

        I just wanted to give some clarification for why Mr. Robot was making some of his choices that might seem a bit counter-intuitive. (btw, in our next update we will have Windwalk as an option for feral tanks, and it will probably show up as the preferred enchant with the default weights and “exclude expensive enchants” unchecked.)

        I play a feral tank in-game, and I put a little extra effort into it (as evinced by RoboMaul, the only class that we have a simulator for atm 😉 ). I’ve found that as your gear levels improve, Savage Defense gets better and better, and the relative value of Mastery, Expertise, and Crit goes up. At the T12 and better gear levels, Mastery Rating is pretty damn close to Dodge Rating point-for-point. Unless blizzard changes something, gearing for Dodge Rating at the T12 heroic and higher gear levels will be a thing of the past: everyone will gear for Mastery, then Expertise, then Crit.

        Note that if you do normal-mode bosses, even at lower gear levels you’ll get a lot more out of Mastery. It’s only when the bosses start hitting really hard that Dodge has a significant edge. I prefer a Mastery/Expertise setup right now for normal-mode firelands 10, because it doesn’t significantly impact my survivability, and I do a bit more damage. Killing bosses more quickly (and having less attacks avoided) is more to my tastes than an extra 1-2% damage reduction.

        So yeah as you said — there is definitely room for personal preferences now, which is very cool to me. I don’t really like having one way that you “have” to gear.

      • achloryn July 24, 2011 at 3:45 pm #

        Have you tried single-tanking Baleroc yet? That’s a good time to have as much dodge as possible, since our self-healing from LotP is gimped after decimating strike. Dodging over 50% of his swings is just… effing godly lol. Also I was under the impression (and I could be wrong) that even at t12 heroic gear levels we’re not at the point of reaching dodge DR just yet. I’m not sure what point that steps in though, so make of that what you will.

        I can see what you mean about killing bosses quicker, but the onus of that shouldn’t fall on us as tanks. If a boss is hitting enrage timers, it’s more cause the dedicated DPS aren’t putting out enough damage, not cause the tanks need to deal more damage. That said, in our shannox kill I put out 23k dps while tanking… 😉

      • Yellowfive July 24, 2011 at 4:34 pm #

        Yeah that’s the other tricky thing for tank gearing — particular fight mechanics can favor particular gear setups, or even particular items. I haven’t tried Baleroc yet — just getting started on Firelands with a new guild, so only 2 bosses in, spending too much time making the website and not enough time punching dragons.

        Diminishing returns is a continuous function — it is applied no matter how much or how little dodge you have. For example, in my current setup, I have 25.84% dodge from gear (before diminishing returns). Bears get an extra 10% dodge from talents that cannot be diminished, plus a “base” 4.95% dodge that cannot be diminished.

        So total, I should have about 40.79% dodge, if there were no DR. But once DR kicks in, my actual dodge chance is 37%. So I lose about 3.8% dodge due to DR, or in other words, I lose about 15% of the dodge gained from my gear after DR reduces it. This ratio will continue to rise as you stack more agility and dodge rating.

        If I were to reoptimize and gear for dodge rating instead of mastery/expertise, after DR and all that, I would gain 3% chance to dodge. So that’s another point to keep in mind: the different gearing strategies aren’t going to make massive differences in your final stats. That’s why I tend to go with a mastery/expertise/crit setup for tanking: I only lose about 3% dodge, which is offset for the most part by larger and more often SD shields, and then I’m left with a set of gear that has a lot more cross-over with a cat spec for those single-tank situations. I find that to be one of the greatest strengths of feral druids: the ability to be a top-notch tank and DPSer with basically the exact same gear.

        For that same reason, I tend to ignore the armor enchant suggestions and use agility enchants instead. I do keep two helmets around though, one with tank meta and enchant, the other with dps meta and enchant.

  3. Pazupati July 24, 2011 at 3:57 pm #

    I think this was a well thought-out post. Working with a team of casual players, gear optimization is often something you walk on eggshells about. I’ve dealt with all sorts of people (so of my favorites are the ones that stubbornly refuse to take any advice on their class. Those are interesting >..<

    That said, however, I still strongly appreciate and support those that do all this work. I wouldn't know my class half as well without someone taking the time to explain it to me. I just know what works for me, and what doesn't. XD

    • Pazupati July 24, 2011 at 3:59 pm #

      OK, for some reason WP cut out half of my comment, I don’t know why. But, yea, good post! 😀

    • achloryn July 25, 2011 at 6:44 am #

      Thanks for the comment Pazu :D. I’ve no idea what WP decided to eat half of your comment.. silly blog.

      Anyway, like I told Yellowfive, I didn’t mean this post in any way as a complete warning against AMR, Rawr, or any of the others out there. More of a “buyer beware” if you will. Make sure you know what you’re getting yourself into, and why the program says the things it does. I clearly don’t have the kind of brainpower it takes to create all this mathematical stuff, so I have to rely on others to do it, and I am exceedingly grateful to them for it. ;D

  4. Ceraphus July 25, 2011 at 8:15 am #

    Its amazing as how much common sense is lost on folks, good point for folks I use ask mr robot, but at the end when it tells me to reforge and have me go way over the hit cap, I have to use common sense and say I rather have .01% chance to miss, than be over the hit cap by 50 points (yes this has happened)

  5. theerivs July 29, 2011 at 3:28 pm #

    Holy cow someone from Mr. Robot replied to your accusations. I’m stunned.


  1. Weekly Lagout « A High Latency Life - July 29, 2011

    […] a guildie of mine who goes by the name Fuzzbutt put an awesome post about gear optimization, so check him out at Warcraft of the […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: