To Nytkin and similar posters: discussing +hit here is all good and nice, but please read all the cat dps related threads on EJ before posting misinformation like "hit is vital for good dps".

Maybe I'm an odd breed, but I maintain dodge/hit cap to the point of complete regems every time item loadout changes. It may be an inferior stat for paperdoll/dummy DPS but there are many, many situations where stuff not landing is a severe blow to DPS, especially when you have a limited number of GCDs over a span of time on the boss along with i.e. automatically hitting rake->shred and not reacting fast enough to the fact that rake missed.

I'm content to drop a little bit of efficiency just to be sure that as long as I'm in shred territory (i.e. no parries) any button that I press is a landed ability. There's more than enough complexity in the DPS cycle itself along with various encounter mechanics that eliminating at least that bit of variability is subjectively worth it.

Yes, it's generally accepted that hit/expertise is point for point an inferior paperdoll stat.

Yes, it's also generally accepted that getting hit/dodge capped is a matter of personal preference, and can still improve overall DPS depending on player.

This discussion doesn't really need to be re-hashed every five pages in multiple threads.

It all boils down to:

Is paying attention to dodges/misses in addition to the standard DPS cycle and boss mechanics too complex for you? Try to reach hit/dodge cap.

Can you handle all this tasks? Will you adjust your next move immediately after missing an attack? Then Agi or ArP will grant you more DPS than Hit/Expertise even though some critical debuffs miss from time to time.

Exceptions are fights like XT, where there are short effects which increase damage and you don't want to miss here.

Every time this discussion comes up there is one big factor people miss which is events which require movement. If you miss your finisher when moving out for an AoE effect, or on adds that must be bursted down or to finish a cycle before innervating/CR, you don't get another chance at that, you max out on energy and you start losing damage.

I don't really try gem for hit but I do my best to maintain a decent figure. I think I'll likely use [Blood of the Old God] over DM:G if I ever get it, I've put off replacing Grim Toll with my Runestone for the time being because I didn't really want to lose that big chunk of hit and the reliability it provides.

Also I find your claim that "If you can't pay attention, cap hit, if you can don't" to be downright absurd.

The "But you may miss in a critical situation!" argument regarding hit has always seemed odd to me. Sure, if you're 2% under the hit cap, then one in every 50 times you'll miss that ferocious bite when running out of AoE instant death X. The other 49 times, you'll connect and do slightly more damage as a result of your higher ArP.

Arguing that hit is better because of a situation like this seems akin to arguing that crit is better than ArP since "Think how much damage you are losing if your FB fails to crit at a crucial moment!" In the end, it all balances out, and unless undercapping hit leads to the problems described above (you routinely shred without debuffs up etc due to reaction time) then the simulator result will generally reflect how you'll do on average.

Obviously the one instance where a critical move misses at exactly the wrong time will look bad for Joe UnderHitCap - but all the times he doesn't miss, he's doing slightly more, and in the end, it evens out (or in many cases, comes out with ArP ahead).

Well, the value of hit is low because when you miss a special attack, you lose only 1/5 of the energy and none of the combo points. If you regularly encounter situations where a miss costs you all of the energy (say you have a long enough break that your energy overflows) or even combo points (perhaps you have to switch targets), then its value would go up.

The effect is small given a reasonable frequency of such events, however. According to one estimate I made a while ago (in 3.0x), if we apply a very generous rate of once per minute for such events, such that every 60 seconds a special attack will cost its full energy cost if it misses, then this increases the value of hit by a factor of 1.03 or so - not nearly enough to push it over agility or armor penetration.

No, you're completely missing the point. Say on Freya. A lot of the time you have rapid target swaps and need to get everything up on each one, decide what finisher to use on each (if at all). Other times you dodge in and out of range due to fight mechanics and end up in front of the mob a lot (Hence expertise becomes twice as valuable, though I'm not condoning going over 6.5). Other times you -need- mobs killed ASAP or people die.

THAT is the conclusion you're saying compares to arguing crit is better than ArP. And even then the above point is just topping on the cake.

The main thing is if you can react fast enough to every single miss, while paying attention to the overall DPS cycle, WHILE paying attention to boss mechanics. I'm sorry but as much as I'd like to say I'm an amazing person with cat-like reflexes, this isn't my case. Playing consistently with 300-600ms doesn't help. Guaranteeing that every ability lands is the workaround.

The main issue isn't a 2% chance of a yellow attack missing and losing 1/5th of that energy. The main issue is having to pay more attention -constantly- to the possibilty of a miss and a much more severe DPE/DPS loss from not pressing the right button in the event it happens.

Actually I don't like to play to far from hit/expertise cap, but I'm not gemmin' specifically for it. If we look at BiS gear (T7.25-T8.25-T8.25 hard mode) are all very very near hit/exp cap so it's not a really big problem (my main problem is usually to not over cap them).

As for attack from the front / changing target, I don't see fights where it happens so often. The only fight were you should attack from the front is Kologarn and some part of Yog. Freya if you know how to move and your tanks know it, you'll basically always attack from behind, the same with aruya. For the miss problem, in wich fight you switch target so often?
Razor = AoE add + Kill boss, not so much change.
XT = Single target
Ignis = Single Target
Aru = single target + AoE but without target switch
Kolo = switch target about every 60 seconds
Freya = switch target around every 60 seconds, I use to build a 5cp sr (with 4t8) between phases on freya so I usually don't have to refresh it while doing adds, the little plant = build cp on her and aoe swipe = high arpen value.
Vezax = you don't switch target
hodir = you don't switch target (we use casters to down iced)
thorim = into arena built 2-3 cp sr + swipe, and then single target him.

Hello guys,

First I wanna say that I am not druid player my character is warrior, but I want to get some info about the ferals and that is because I want to clear something in my mind when I distribute loot for our guild raids.

So I read here (remember I am druid noob) that agility > strength for feral dps. Ok, but 1 strength is roughly 2 AP and 1 agility is roughly 1 AP, and yes agility gives also some crit. Our feral druid has 55% crit raid buffed, so I wonder does he really need more crit? He is trying strength gemming atm and he has that crit chance with strength. So could you please tell me why agility is better, when strength gives more AP and I don’t see the need for more than 55% crit?

Really sorry if it is that much noob question, but I want to understand what is the logic for considering agility > strength for ferals.

Thanks in advance all!

That's because crit is important for us.

First, we've got a talent which increases our critdmg by 10%, additionally to the meta gem.
Second, and this is the important point, for every yellow crit, we get 2 combopoints instead of one.

Actually, Strength is not far behind Agility and I consider it the third best stat (ArP > Agi > Str), but Agility is clearly better.

Actually str=2*1.1*1.06 = 2.332 AP
agi = 1.1*1.06*( 1 AP + 0.012% crit) = 1.166 AP + 0.01272% crit

as said by murna we do 2*1.1*1.03 = 222,6% normal damage with a crit and we also generate 2 combo points. Due to max combo points being 5, with an high crit you have an higher chance to waste a combo point. If you graph it vs. crit % you'll se a change in slope at around 60% crit, but the dps value doesn't change a lot.

Basically if everything your feral druid should gem arpen (we are probably the class that benefit more from it) or at least agility, feral druids doesn't gem str since the last patch.

Yes but rip is buffed by strength AND agility as well. Rip crits doesn't give you any additional bonuses to favor agility over strength.
Actually you could even argue, that strength could be better than agility, because one portion of our damage, namely rake, is not affected by the crit portion of Agility.

The reason, that Agi > Str is not rip being able to crit. It's extra crit damage and bonus Combopoints. These two advantages even outweights the advantage of strength: Buffing rake more than Agi.

Of course you are right, when you say that Primal Gore was one of two reasons that our focus shifted from Str back to Agi when Patch 3.1 was released.

I was looking at it more from the standpoint that, as far as my understanding went, that gemming strength was the way to go in patch 3.0. The only thing that changed from 3.0 to 3.1 (besides gear) was that we got critting dots and SR swapped from AP buff to damage buff, and that shifted the weighting in favor of agility over strength. The extra crit damage and bonus combo points existed in 3.0, and that didn't change from 3.0 to 3.1. I could be wrong about the pre-3.1 gemming, but I didn't DPS as much in Naxx and consequently didn't do as much research as I do now.

This discussion of hit reminds me of a really interesting test a shadow priest from my guild tried on the PTR. He stacked nearly 1000 haste so he could fit 3 mind flays in between his mind blast cooldown (and have like 3.8 sec mindsears), the problem was that without him stacking SP and crit into the slots like priest Sims tell you you should that extra mindflay couldnt make up for the extra damage done throught the full rotation or the ticks off MS.

For Druids a similiar thing happens, hit is important but no matter how much you want your attacks to hit at the right times, your total damage done (and as a dps thats the ONLY stat that matters) will come out greater stacking armour pen and agil over hit, almost 100% of the time.

@Cliffjumper:
IIRC The change to SR was at least as important in the shift back to Agility over Str for Cats as was the introduction of Primal Gore. Btw I appreciate all the work so many people on these forums have put into contributing to the theorycraft that has helped my Druid out so much.

Please don`t forget that Arpen is our best stat in simulations.

Yes hit, str, agi, and expertise are good stats, they will improve your dps in whatever way.

You are never able to say which stat is the best stat in a real fight. If you get very very unlucky in the heartphase of XT and you miss/dodge every attack. Then it would have been bettet to gem for expertise/hit. But if you hit everything when his heart is active arpen is the best stat.

Arpen gives you 3-5dps per gem over agi in a simulation. If you have 10 Arpen gems that would be 50dps. Could be more, could be less.

Now you are in a fight, a healer needs your innervate and another one needs your BR. The improve you gain from your gemming is lost when you are done with both.

So if a simulation shows us Arpen is the best stat in that simulation this stat will probably be the best stat for this Tier, for all bosses not just for one maybe two bosses.

If you have enough gems then you could simulate each bossfight in Ulduar an regem for every Boss. If not then just trust people like night who put alot of effort in those calculation.

Well this is the problem the cat community suffers at the moment. RAWR used to be our reference. We used to trust RAWR.

Now there are two different concepts - FBN and RAWR have some difference results here and there.
The author of RAWR, Astrylian, says that the FBN simulator has heavy flaws. Though, nightcrowler did some incredible work there and his code has been checked by many other druids and I can't see mistakes being made.
On the other hand, the theory behind the RAWR dps calculator is unclear.
Remember - the FBN suggestions to stats are only true, if you follow the FBN Addon Move suggestions.

The fact is: Cat DPS is very very complex. Slight adjustments will lead to different results. There's so much that can change the outcome. There is no rule for every playstyle. It depends on your rip and rake uptime, on your shred and FB usage and many other things.

My experience is, that RAWR will start to suggest ArP gemming a little later, at around 350-400 ArP. But then again there seems to be some hidden "magic thresholds" where the stat value table changes completely, when only one item is changed. I noticed some very odd behaviour. Sometimes Haste Rating jumps from fifth to first place in the RAWR stat value table and Str getting better than Agi, when only slight changes are made to my gear.


At the moment my personal decision is to trust Nightcrowler. Since nobody knows the theory behind RAWR we can't check which of the two models displays real bossfights more realistic. Perhaps they both work but assume different Cycles.
However - RAWR has had a great history and for a very long time could be trusted completely. It is still a very good tool and has by far the best user friendly interface.
So I tend to follow the RAWR item suggestions, but have in mind, that ArP is probably a little more worth, than RAWR suggests.

@Aangus: I used to have the same problem with rawr. Now it says Arpen, but only if I have trauma and Sunder Armor.
I had the problem that there were no Itemsuggestion with Arpen as gems, once I added those Arpen again became the best stat


Like Murna said if you are using the FBN addon then gem for Arpen and ignore Rawr. If not then follow Rawr until it tells you to gem Arpen.

Setboni: Again nightcrowler (gotta love this guy) ran simulations with 2t7, 2t7/2t8, 2t8, 4t8. Use 2T7/2t8 until you get 4T8.

You should ask yourself:
Do you have Problems with the Priority Rota?
-If yes try to use FBN and gem for Arpen.

Do you run 10 or 25, is trauma and Sunder always on the target?
-If yes arpen will be a better stat.

Do you use some of the gear for tanking?
-If yes gem for Agi. ( I had that problem since I am ST in my guild.)

Is there a chance you will get an Arpenproctrinket in the near future?
Do you have a good OoC management?
Do you use some gear for PVP?
You don`t have a manglebot, and you have problems with Mangel/rip uptime?



The gem question is only interesting once you have no more issues with your Rota or use FBN.

@Aangus: First of all, there's no Aangus on CC on the armory. Regardless, I don't feel like answering a "this is what I have, help me not suck" post by telling you what to do. You can see that the resources are there (Rawr, Toskk, FbN) If you choose not to believe them, then don't. I really don't get why people feel the need to ask why the generic blanket-statement advice in the "for Dummies" thread is telling them something different from a personalized and specific program.

Rawr is telling you to gem agi because your hit/expertise/crit is pretty low, thus your CP per Energy is low. Rawr is trying to increase your CP Generation so you can generate enough CP for Rawrs cycles. Trust Rawr, or don't.

From posting of damage breakdowns I've seen white attacks do roughly 30% of your damage. Druid dps specials are instant and not affected by haste, so x haste % will increase your damage by ~ x/3, right? I find myself waiting for energy more often than the global cooldown, so seeing as hit & expertise will linearly increase ALL of your damage until you reach the cap, and 1 point of haste, hit & expertise rating all use the same factor (1pt = 0.03%) shouldn't hit and expertise be prioritized above haste until capped, at which point haste would of course become better? I apologize if this is wrong.

Do haste, hit and expertise scale differently with level so that the 0.03% per point factor @ 80 is different for each stat at lower levels?

More haste - more OoC proccs.

Oh right, I knew I was forgetting something! Would haste also increase weapon/weapon enchant procs by haste% then? Probably not a discussion for this thread though.