So... how about all those enhancement geared totems they put in that are available from raiding? Guys? Guys?

Today is gripe day and one useful totem in all of the new TBC content is garbage itemization.

I don't forsee getting rid of the Pandemonius totem until the next expansion and I suspect it'll last a good deal longer than it has any right to even then.

My roommate came in and laughed at my pitiful strength last week. I was pretty depressed to discover I had as more +AP from gear than I had Strength, period. I 'lose' about 50 AP whenever I get BoK. Hooray for hunter itemization!

Also, Windfury cooldown, etc.

Gah sorry. I had a typo in there that made my whole post uber confusing.

You said that 75.6 ap = 5.4 dps. You got that number by using the rough conversion that says 14 ap adds 1 dps to your weapon. But what you fail to account for is that 75.6 ap, in addition to adding 14 dps to your main hand, will also add 7 dps to your offhand weapon, add dps to your stormstrikes and windfuries, and scales with flurry.

If your formula, that 14 ap = 1 dps, held true, then shaman with 1401 ap would only be doing 100 dps more than shaman with 1 ap.

So this is the real debate. Disquette's numbers definitely show a low value of Crit and Hit in relation to AP. He tested across values of AP from 800 to 2500 - of which my AP falls squarely in the middle. The numbers I use are directly from Disquette's results.

Now, we have a situation where no one completely trusts those results. To us, the value of hit and crit all feel low... so we all fudge them upwards. Of course, then what's the point of the modelling? I think the modelling tells us that a shaman with unbuffed 1700 ap, but only 20% crit and 20% hit would probably do more DPS than I do... but I don't know anyone who is really willing to abandon all that hit and crit and find out.

Of course, you eventually have to stop looking at this in a vacuum. The itemization is not there for us to really have as much say as we'd like. My guild is fairly rogue heavy, so I've decided that I'm going to stick to mostly mail, unless items are going to get DE'd. My second concession is that I'm a whore for looking good - so t5 is a must. Very quickly it becomes easy to see which slots you're going to have to use to populate your hit rating and crit rating: Neck, Cloak, Belt, Bracer (Lol good luck with this), Boots and Rings. With my unwillingness to drop below 200 hit rating, my looting goals are more or less set for me.

Running the numbers, it looks like in the next few weeks I'll be gaining about 200 ap, but losing an alarming amount of crit (4%). I suppose that I've been meaning to switch to mongoose x2 anyway, right? bleh.

Az

P.S. - (Yeah, I assume kings roughly 80% of the time in raids. That's how I ended up at 2.17 for STR)

I've got a java-based sim that matches up pretty well with Disquette's. As Disquette has been pointing out lately, our understanding of certain mechanics, such as flurry, isn't exact yet. That said, my results agreed with his in that the benefit from crit and AP are much more than that from hit, even well below the hit cap.

Starting at 2200/25/22 or so to get a baseline, I tested 6 gear combos: [+]and [-] a fixed amount of item points in [AP], [crit], and [hit]. Adding the points in hit increased DPS by the least of the three, and subtracting the points from hit reduced DPS by the least of the three.

Time to add a GUI and make it downloadable Pater

I agree with the methods you guys have been using, and I trust your models quite a bit. I'm just not entirely sure that I'm ready to spend 200 dkp on items without hit rating yet. I'm waiting for some other sucker to take that plunge first, and tell me how it's worked out, heh!

If you ever put a frontend on to the sim, I'd be interested to download it and plug in my own numbers to do some gear comparisons.

Maybe I've missed a subsequent update, but I believe my numbers are accurate. On page 16 ([Shaman] Itemizing Enhancement at 70, disquette gives the following numbers:

You can extrapolate that to be:
0.5 Str = 1 ap = 1.08 HR = 1.51 Cr = 1.57 Agi

Now, lets assume kings 50% of the time, making AGI and STR more valuable:
0.45 Str = 1 ap = 1.08 HR = 1.51 Cr = 1.43 agi

Now, these are equivalences, not weights. We want weights, so we divide 1 by all of these, and get the following weights:

Str - 222
Agi - 70
HR - 93
CR - 66
AP - 100

So, unless I've missed an update, those are the exact weights obtained by using the Sim results, but assuming Kings 50% of the time. I've followed this thread very closely, and I believe those numbers to be accurate. However, the Sim makes a few assumptions. It assumes a minimum of 20% hit, 20% crit and at least 800 attack power. If you're looking at values above those thresholds, then the above weights should be accurate for selecting gear (according to the sim).

Unfortunately, we don't have any concrete empirical testing to prove that these numbers are accurate. The SIM numbers value crit extremely low, but flurry also behaves unexpectedly, so your mileage may vary. I'm not shocked to find out that AP is valued so highly - our combat model with Flurry and Windfury means we have extremely good AP compression.

Edit: This should be the appropriate link, http://www.lootzor.com/index.php?c=7...6hedu93heeu100 - but does not value stamina. Add your own Stamina / Int / Mp5 and Armor weights if you'd like. Keep in mind that those weights assume some minimums... If you just go for the top thing on that list in every category, you'll end up with like 75 hit rating

I haven't looked at this stuff for a while, but I believe that what you were looking at actually was a weighting. I'm embarrassed that I have to ask for confirm, but I believe it was the pawn string that you were looking at. That is, I was lax in my writing, making that bit you quoted appear to be an equivalency, whereas when you value stats, you should value them at that ratio.

a pawn string should multiply the item's stats by the constants given. Here's a quote a little below the one you got. The equivalency formula is that second line starting with 34AP =...

these were older numbers, which is why they're off the 1.51CR, 1.08HR, but I wanted to pull from the same post you did.

The way Disquette phrased it, it does sound like those are part of an AP Equivalency formula, but they are actually weights for an item valuation formula. It should read AP = 1, Hit = 1.08, etc. Check a couple of his later posts for verification of this.

Holy Mass Confusion Batman!

God, if I've had the whole damn thing backwards, I think I'm gonna cry... I'll look over math.

Okay, Disquette and Rob - thanks for the clarification. Everyone can disregard all the math I'd cited in the previous meta gem discussions. So your current equilivencies are:

AP 1
CR 1.51
HR 1.08

Is that correct?
Sigh. Good thing it's a slow day at work

So as a new Shaman gearing up for endgame soon, should I focus my gear/gems around more +crit than anything else? I'm going for the Ebon Netherscale set as a strong base for my gear, if that helps clarify.

Based off the post you quoted, you should actually be focusing on +STR gems. However, most of us agree that a certain "baseline" of crit/hit is optimal to achieve before you stack AP; this figure varies depending on who you talk to, but I don't think anyone will argue with a low estimate of aiming for 20% crit and 6% hit from gear (in addition to the 9% you get from talents). Ebon Netherscale is definitely a nice start; however, if you don't have access to arena weapons, I'd have to say that Blacksmithing is the way to go, to get either MH; the Ebon Netherscale can mostly be replaced with Karazhan gear, but the Blacksmithing MHs cannot be. If you have access to Arena gear though, this concern mostly goes away. 2x Gladiator's Cleaver is the best easily attainable combo at the moment.

Ack yes ofcourse, I fail at math horribly. Thanks for the explanation. No more shaman based math for me, I'll leave that to the experts

This is a tad off-topic but, am I the only one who actually prefers the maces over the axes? Or is there some difference that I am missing.

Stoppokingme: No worries. I think your conclusion is still accurate, that the Relentless Earthstorm is awesome. I was just curious if you had some firm WWS numbers or not. My current helm doesn't have a Meta Gem slot, but I'll do some solid before-and-after once my t5 drops.

Negative: If you want to plan on having Blessing of Kings, then the following are some pretty good weights. (Plug these into www.lootzor.com). I also value stamina at around 50.

AP 100
CR 151
HR 108
STR 220
AGI 159

Myrrdon: I think axes have better graphics. As far as "best", I think Dragonstrike / Merciless Gladiator's (Axe or Mace) is the way to go. I wonder if Arena Season 2 goes live tomorrow...

I currently don't have access to nether vortex (or tier 4/5 atm), and I am currently wearing Ebon Netherscale and using Drakefist Hammer and Runic Hammer, I'm thinking maybe I should get season 2 gear, while working to upgrade to Dragonmaw, and then work on 2 season 2 weapons. Any thoughts? (And yes, I'm one of those crazy people who picks up LW and BS cause its got huge benefits, regardless of all the hours/money i spent leveling them :P )


Oh, and according to the devs, so far so good on the timeline for tomorrow being patch day.

There's no difference in stats; personally, I plan to take the axes just because I like the graphic better: the mace looks kind of cheezy. My team got our 5v5 rating up to 1650 this week (some idiots played with a crappy healer and sabotaged it 50 points in 4 games ), so I'll be the proud owner of a Merciless Gladiator's Cleaver on Tuesday. (Another reason to take the axe: if a Decapitator drops in the next four weeks, it will become my new MH until I get a 2nd Gladiator's Cleaver, and I think two axes looks cooler than two maces.)

Just as an interesting aside, I'm thinking that it will become my new OH rather than MH, bear with me on this math.
Fool's Bane = 87.5 DPS
Boggspine Knuckles = 71.7 DPS
Merciless Gladiator's Cleaver = 97.5 DPS

MH upgrade = 97.5 - 87.5 = 10 DPS
OH upgrade = (97.5 - 71.7)/2 = 12.9 DPS

This doesn't account for WF/SS damage being halved on the offhand but I think when you add in special attacks the calculations only get better in favor of the OH, right? (If WF/SS combined is 40% of damage and autoattack is 50%, then MH upgrade = 10 + .8*10 and OH upgrade = 12.9 + .8*12.9... so OH is even better than it looks.)

OK, I've uploaded my java-based simulator. It's still in development, and almost surely has errors. At a minimum, we don't yet know how flurry really works, so it's only our best guess so far. It also doesn't account for +weapon skill. It doesn't account for any weapon procs yet. There are other things it doesn't account for. That said, the goal is to model enh shaman melee combat as accurately as possible, and I believe it does a good enough job to draw some conclusions about what stats matter the most.

http://www.discofiend.com/pater/ - file is ESCS.java (the only file in there)

You're welcome to play around with it--let me know if you spot logic errors or have suggestions. Probably the biggest pain at this point is the I/O--you need to edit the java file, recompile, and rerun every time you change stats. I know zilch about building a web front-end, so if someone wants to collaborate with that, I'd love to work with you.



As far as what I've taken from it:

My starting point is 2200 AP, 25% crit, 22% hit. Along with my standard gear, mob level, etc., this gives a DPS of 752.


Now I do 6 other runs, adding or subtracting a fixed amount of itemization points to AP, hit, and crit. I added 50 ilvl points, or 100 AP, 50 CritRating (2.26% crit), or 50 HitRating (3.16% hit).

Results
+100AP = +23 dps / -100AP = -21 dps
+50CR = +23 dps / -50CR = -22 dps
+50HR = +17 dps / -50HR = -15 dps

Based on this, I've been using weights something like:

AP = 0.5
CR = 1.0
HR = 0.7

STR and AGI can be extrapolated from these direcly. I assume kings is up always so use STR=1.1 and AGI=1.0.


I think my java program is easy enough to manipulate to use your own weapons and gear starting point.

That should be Str=1.0 AP=0.5 right?

Yes, I'm sorry, brain fart from setting it up for equal ilvls. Editing original.
The reason I set up the 6 test cases with equal ilvls is so that I'm exploring an input domain of roughly the same size in each dimension, to help ensure the local linearity assumption isn't too far off.

So i'm trying to plan out my future weapon upgrades, and the way im heading is to be duel wielding this http://www.wowhead.com/?item=32262

Now im wondering if using 2 x 2.8 speed weapons would be better, then same or slightly worse than using http://www.wowhead.com/?item=28439 MH and http://www.wowhead.com/?item=32262 OH

Any thoughts on this, and how much increase I would see from the 0.1 speed by using Nathrezims, over the 2.7 speed with the haste effect.