Thanks Boevis. I keep running 10 man Maly trying to get the Surge Needle Ring but it won't drop sadly. My choice of rings at the moment is very limited. But I will start to make a more concerted effort at getting those replaced. I will be getting the badge ring as soon as I can afford it (I can't remember what I bought, but it was a waste sadly).

Anyway, thank you for the constructive words sir. Oh, I use Quartz right now for debuffs, but I should probably switch to Dot timer anwyay. I just looked voer NeedToKnow and Power Auras and that seems to be exactly what I need. Thanks again for the help.

DR% for Lvl80 player attacks = Armor * 100 / (Armor + 15232.5)

Unbuffed: 13000 * 100 / (13000 + 15232.5) = 46% reduced = 54% damage done
10% ArPen: 11700 * 100/(11700 + 15232.5) = 43.4% reduced = 56.6% damage done

56.6/54 = 1.048 = 4.8% damage increase

Debuffed: 8000 * 100 / (8000 + 15232.5) = 34.4% reduced = 65.6% damage done
10% ArPen: 7200 * 100 / (7200 + 15232.5) = 32.1% reduced = 67.9% damage done

67.9/65.6 = 3.5% damage increase

800/1300 = 61% less armor
3.5/4.8 = 73% effective

Not a HUGE difference, but significant enough to correct.

Well I don't know If I have understand.

Are you saying that eating a food or things like that changes a lot the relative stats weight because rawr uses rounding and adding for example 1 agility after the rounding will result in an "effective" 0.5 agility so it value it as half?

It seems buggy... I understand the problem, the way to avoid it is simple. Take a 100-200 stats up/down step.

I think what Allev means is that the stats on rawr is accurate, but only accurate if you follow the rotation and abilities which it indicates as your optimal rotation. If you have a lower percentage damage on shred and higher damage percentage on rip, then to you personally Strength is going to be better than the "normal value" since dot's can't crit, and Agility will be better "than normal value" on the opposite case because shred can crit.

That's one of the things I'm saying, yes. The other thing I was trying to say is what he was confused about, though, so let me give some examples.

The misunderstanding is, simply, stat rounding errors. Say you have 0 expertise and it takes 41 expertise rating for a point of expertise. In-game, it rounds down, so points 0-40 are worthless until you get the 41st point. If you eat 40 expertise rating food (in this example), that's weighted as 0 in Rawr because it's worth 0 in the game.

These kind of errors lead to strength-less-than-AP errors. Imagine you have 0 strength. You add a 14 Str gem, which gets multiplied by 1.03(SotF), then 1.10(BoK). That's 15.862 post-buff. But the game rounds it to 15, so it adds 30 AP. That gets multiplied by 1.1 (HotW), which gives you 33 AP. If you skip rounding in the first step, you actually get 34.8964 AP. Not a big difference-- only about 6%-- but this is one simple rounding of one stat. More rounding will cause more error.

Yes, the correction is to take X points of stats and figure it out. But, adding sufficient stats to eliminate the rounding errors can also cause other problems, like going over soft-caps or hard-caps of stats. Which is why lots of theorycraft tools, like Toskk's, skip the rounding steps to tell you average-case what those points will be worth. Which can differ from what the game actually gives you.

It's a design choice of Rawr to mirror the game world, but it makes it hard to say "stat X is better than stat Y". Which is why you should rarely if ever take Rawr's stat estimations as accurate except for exactly what they're measuring-- 1 point, 20 points, or 100 points.

The worst part about all this is, the rounding approximations are worst in Rawr when doing comparisons of small values of individual stats, because it only happens collectively at the end after all the stats have been added up. Only the last points have rounding error. So every time you look at a specific slot in Rawr, you get all the rounding errors, even though you don't get them when you look collectively. So, for instance, you can overlook 2 pieces of gear which the game would collectively add up as 6 <statpoints> each individually, but end up being worth 7 <statpoints> collectively, which makes the difference between both being best-in-slot and being in second place. (Not that this is ACTUALLY the case anywhere, it's just hypothetically possible.)

Finally-- every theorycrafting program/spreadsheet/webpage will have to make the decision to follow the rounding, or average out the rounding error to all of your gear. Comparing stat values from the rounding theory is a weak point; similarly, comparing exact combinations of items will never be exactly accurate if you average out stat values.

As an aside, I have no idea what Rawr is actually using to generate the average stats, which is one reason I don't use or dispute Rawr's relative values unless I see something orders of magnitude different from what I'm expecting.

Ok I'll tell you the order of magnitude of the error:

I'm at ease when you tell me that my extimate of agility vs. strenght goes for agility while rawr tell me that str is better because I use FB a lot in my cycle (so more crit damage -> agility is better). But really rounding error can't halves the value of agility vs. strenght.

First of all with all the possible rounding error on a 1 vs 1 base (worst case scenario) I should have:

1 agi = 1 ap + 0.01% crit (the game round on second % digit the crit isn't it?)
1 str = 2 ap
1 crit = 0.02% crit

So 2 agi = 1 str + 1 crit with maximum rounding errors.

Now let's see what rawr tell me.

I have my blue cloack with hit on.

Str = 1.23
Agi = 0.98
Crit = 0.57
AP = 0.45

Now I put on the vendor cloak

Hit = 1.04
Str = 0.93
Agi = 0.68
Crit = 0.57
AP = 0.45

Crit and ap value stayed the same while str and agi change A lot

Sustituting Aged winter cloak and badge cloak doesn't change anything.

Yes it's been fixed.

Well, that's not exactly how the rounding works. It doesn't take the individual stat itself, it takes the difference between having 1 more of the stat and having 1 less of the stat.

The rounding error is that when you add 1 Strength point, it actually adds 1 * 1.03 * 1.1 Strength, which for any given point will add either 1 or 2 strength to your final stat point total. With kings for instance-- for the first 9 points of a stat, you get 1 point; at the 10th point, you get 2. If it adds 1, then it's worth approximately twice your AP. If it adds 2, it's worth 4 times a point of AP.

The actual value is 2.7 times AP in your first example, which means it's taking a value greater than 1 and averaging stats together. My guess is that it's taking the average of 3 stat points. In your first example, the 3 strength points round up to be worth 4, which produce 8 or so AP from 3 stat points, which makes it valued at 8/3rds, or 2.66 AP per strength. Or perhaps it's 4 stat points, and it's rounding both strength (4 to 5) and AP (10 to 11), which is 11/4 = 2.75.

Either crit and AP aren't rounding differently in the examples you gave, or Rawr uses their exact values instead of rounding them (because the game doesn't). I'm not sure which is the case.

Does this make sense?

And does it make sense now to NOT use Rawr's relative stat weights unless you know what you're doing?

Edit: So you know this whole discussion? It looks like it's resolved in 2.1.9-- Rawr looks to be reliably and precisely calculating accurate averaged-out stat values now (probably as part of the generic rollout for all classes). So just update to the newest version.

I understand what you're saying Allev, the problem is understanding exactly what method is being used for stat values

With SotF and Kings, 1 strength is actually worth 2 due to Blizzard's rounding roughly every 6 points. Being such a small variable (less than 1 gem) I just don't see why Str and Agi aren't just defaulted to 2.332*AP and (Str+Crit)/2. Of if you're insistent on having values accurate to your specific gear, it should be averaged around +16, ie. gem value.

But this is a discussion for the Rawr thread, not really here.

Sorry if this has been brought up before (looked, couldnt find it) - but i was wondering which of the rip and shred idols is a greater dps boost? I'm currently using the rip idol - mainly because i know its 1050 dmg every rip, and i'm not sure how the dmg boost from the shred idol is applied.

So now, [Idol of Worship] beats out [Idol of the Ravenous Beast] ... Assuming you arent going nuts with strength for stronger shreds?

As pointed out by moof previously in this thread ( http://elitistjerks.com/1068454-post372.html ) crit has a soft cap at 75% if you are hit + expertise capped.

Strength doesn't scale the damage from either idol. Idol of Worship will generally win.

I see no reason to make use of both. Take Maly for example. Phase one use Worship to get good rip dmg while being tossed in the air and you'll get more dmg out of it anyway because it's one target sustained dps and you can keep Rip uptime high. Then, phase 2 switch to IotRB when on the scions since you're not going to get as much rip time on them Shred will probably give you better dps.

A question about the Shred Idol, does idol bonus dmg increase when you crit or is it completely static 203dmg?

Rawr tells me that the Shred idol is currently better for me, however I am not sure how accurate the idol calculations are in Rawr. Does anyone have any good theorycraft on this matter? Or are people just assuming the Rip idol is better?

This is actually a good point. With the duration of a glyphed, 2 peice T7 Rip being 20 seconds, it probably would be noticeably beneficial to swap out idols.

This makes a lot of sense. There is a short period of time before Phase 2 begins in Malygos where we can swap our Idols.
Perhaps Idol Swapping should be considered on certain fights to min-max dps.

Which is completely and totally irrelevant when, as on Malygos, the boss moves out of melee range when he's suggesting making the switch.

If you have a 20 sec RIP:

If you use both RIP and FB, shred idol wins.
If you use only RIP (100% uptime), rip idol wins.

It depends on your rotations. Also: Shred idol scales with crit (it's also increased with talents that increase shred damage), RIP idol doesn't scale at all.

^ This would be why my dps was slightly worse while trying to idol twist during naxx25 tonight. I was wondering about it, since I picked up Idol of Worship about a week and a half ago, and I hadn't seen anything about idol twisting on these forums yet. I tested out on a target dummy whether or not I could pull off a twist in my "rotation" since I knew it reset the GCD, but at low energy levels with TF on cooldown that doesn't matter since you can't do anything for a short period of time except autoattack anyway--I saved swapping out idols for when I had <10 energy and managed to pull of a neat little rotation for a couple minutes, thinking how great it would be to test this out in a raid setting. Unfortunately my results in Naxx were not as good as I hoped. With a roughly .9 second swing timer and swapping out idols maybe 10 or 15 times during a fight...would that completely account for the dps loss I saw? I apologize for not having hard numbers but it does seem like I will be discarding my twisting for a simple phase1-2 swapping for Maly.

A second question: for roughly 52% crit raid-buffed, obviously(?) the shred idol is better for trash, but have we seen an approximate threshold for bosses yet on shred vs rip idols with a certain amount of crit, assuming tier bonus and glyph? Testing is on my to-do list--I am one of the only kittys I have seen raiding in wotlk so there are not many people to compare numbers with.


Oh, one last thing...14k dps on spider packs from cat swipe = pure win, but I think you all have experienced that already!

We're doing Sarth3D now. Second drake is going down more reliably now, but various wonky things have kept us from a kill yesterday (such as the add tank and his healer d/cing at the same time...)

In the WWS, I noticed my DPS is flailing like mad - from 3.5k down to 2k. I know it'll always be somewhat iffy on such burst fights which are stop n go, but I was hoping to find some hints specific to that fight. It's awkward to try to keep a rotation going on a moving drake and firewalls don't make life any easier. Won't ask you to look over the WWS (there's bound to be things like a CR and even a Flame Wall in there), but a couple of pointers would be great.

I finally broke 3800 the other day and it's only because immediately after the first drake dies and immediately after the second drake dies we kill adds, and that tends to be where we can make up the difference, otherwise you will always see subpar numbers (imo). We just aren't designed to be able to switch easily between targets. Our CP's have to be spent on the current target, which if not spent absolutely correctly (i.e.- you may refresh SR too early when you normally would have ripped to switch targets or only did a 2 CP FB) but in any case it will never be optimal for cat dps (and if your guild positions such as we do... none of the adds are in your frontal cone other then when either of the drakes die so cat (swipe) is out of the question. I find though that other then fury warriors I tend to be toward the top of dps/damage for melee with what I described above.