Fury DPS: 4.0 and Cataclysm - Page 33

Fphut · 03/01/11, 10:36 AM

TG question.
I recall someone a few pages earlier suggesting the use of slam to get rid of rage, and only using heroic strike at very high thresholds outside of CS. This may have been a suggestion meant for SMF, but I'm curious if anyone has tried sims considering this approach at different gear levels with TG.
Lately during raids I've been using slam to avoid rage capping, and using heroic strike almost exclusively during CS, which seems like it works, but may be a DPS loss compared to just using heroic strike to avoid rage capping and during CS, and never to slam without a bloodsurge.
I've tried 2 small sims that allow use of slam at a threshold ~10 rage below the HS threshold, but these were a DPS loss from the current default priority. It seems like blizzard really likes slam, but I still don't see its use being good for DPS, outside of bloodsurge, even during heavy aoe damage (increased rage generation) or bloodlust/heroism.

Examples:

(Patchwerk)
# action,conditions
0 flask,type=titanic_strength
1 food,type=beer_basted_crocolisk
2 snapshot_stats
3 golemblood_potion,if=!in_combat|buff.bloodlust.react
4 auto_attack
5 stance,choose=berserker
6 recklessness
7 death_wish
8 cleave,if=target.adds>0
9 whirlwind,if=target.adds>0
A heroic_strike,if=((rage>90&target.health_pct>=20)|buff.battle_trance.up|((buff.incite .up|buff.colossus_smash.up)&((rage>=50&target.health_pct>=20)|(rage>=75&target.health _pct<20))))
B execute,if=buff.executioner_talent.remains<1.5
C colossus_smash
D execute,if=buff.executioner_talent.stack<5
E bloodthirst
F berserker_rage,if=!(buff.death_wish.up|buff.enrage.up|buff.unholy_frenzy.up)&rage>15& cooldown.raging_blow.remains<1
G raging_blow
H slam,if=buff.bloodsurge.react|!(buff.incite.up|buff.colossus_smash.up)&rage>77
I execute,if=rage>=50
J battle_shout,if=rage<70

Results in ~300 dps loss from my default patchwerk.

Adding more later.

Kurg · 03/01/11, 12:49 PM

Although the changes to Flurry amount to a substantial damage buff, both via extra white damage and further heroic strike rage conversions, it only exacerbates the weakness of haste as a secondary stat. The haste rating into haste % is already abysmal for us, and now the massive amount of haste% acquired through Flurry will make whatever haste rating we keep on our gear even more inconsequential. In contrast the flurry buff drastically increases the value of critical strike to the point that gear without critical strike rating will be seen as weak, and perhaps even to the point where we consider using a lower item level piece that does have critical strike in favor of a higher ilevel item that has haste. I'll take the damage buff, but I am not sure I like being pigeonholed, gear-wise, the way arms currently is.

Kaniption · 03/01/11, 4:01 PM

Will the boost to flurry increase the importance of a higher than 8% hit rating?

Deathwing · 03/01/11, 4:26 PM

No, if anything it will lower it because all hit after 8% does is stabilize and increase rage generation. Doubling flurry increases rage generation and stability by itself, so hit, and especially haste, are devalued in comparison to crit.

Add on top of that:

Unshackled Fury now Increases the benefit of abilities that cause or require you to be enraged by 11.2%, down from 44,8%.

This devalues mastery, specifically because it changes the priority in execute phase. And I bet for TG warriors, it puts RB below Slam. It's quite obviously a buff-stacking nerf in pvp, but I thought the CS nerf accomplished that. Or if it didn't accomplish it enough, nerf it is some more. At least that's more stat-neutral. This nerf makes the itemization guy look bad if he doesn't put any crit on the item.

ComMcNeil · 03/01/11, 5:05 PM

Paired with the Flurry buff, will this now promote HS again in our rotation, as we will surely produce more rage now?
Nevertheless, is the nerf to the base Mastery value really that huge on Mastery stat weight? I mean it scales exponentially, but the increased return was not that big if I remember correctly.

Deathwing · 03/01/11, 5:14 PM

Yes, but it will just be more random HS, as in avoiding the rage cap. And since it will devalue hit, your rps might not change that much. Hard to say without an update simulator.

Calling mastery exponentially scaling is technically true, it's a bit misleading. For the value we'll see in this expansion, it's probably closer to linear than exponential. It might not be a big change(stat weights), but it's hard to tell where the priorities will shake out.

• zeidrich · 03/01/11, 7:45 PM

Yeah, I'm not feeling these changes. The flurry change was ill conceived in my opinion. Crit is already far and away the premier rating and this pushes it further.

The reduction in base mastery also devalues mastery, which was already tenuous as a secondary stat. I honestly don't think it will put RB below Slam in priority for TG because slam still only hits with main hand for TG, but it will further alter stat weights and probably lower SMFs mastery rating to below white hit level.

Divergent stat weights is primarily going to manifest in annoying gear choices, meaning boneheaded itemization is going to be more frustrating than if stat weights were more clustered. Otherwise, I don't think these changes will make any difference to actual play-style, only numbers.

Mastery isn't exponentially scaling. It's cubic.

Narkanius · 03/01/11, 8:48 PM

Originally Posted by landsoul

4 swings theory - The synced weapons scenario
If you are able to sync your weapons by resetting your autoswings on a target, in theory the game will not be able to differentiate between the 3rd or 4th attack when consuming the Flurry charge. Your 4th attack occurs at the same time as your 3rd attack, effectively giving you 4 swings per Flurry charge. This only changes the Tconsume formula.

{TConsume} = 3/(1.25*{IASeffects}*(1/{MHWSpd}+1/{OHWSpd}))
Becones...
{TConsume} = 4/(1.25*{IASeffects}*(1/{MHWSpd}+1/{OHWSpd}))

Since Tconsume increases, the opportunity to crit an instant attack and refresh the Flurry lengthens. This slightly increases your Flurry uptime.

This has seen minimal debate and/or discussion on any of the warrior threads. I think it's time to assess the validity of this theory, as it has some important implications. I've tried testing this theory in-game, but am dissatisfied with the results. I first synchronized my swing timer on two 3.6 weapons. After watching my combat log for 10 minutes, I verified that the whites were indeed occurring at the same moment (if they were to have differed by any margin greater 2 milliseconds, the melee swings would have eventually occurred at different time-stamped moments (20.06:14 and 20.06:15 for instance, this did not occur over the full 10 minute span).

So I had confirmed over the course of 10 minutes, that my melee swings were synchronized (assuming the lag between strikes wasn't greater than zero but less than .002 seconds, this is a major assumption). What I found on flurry procs was disheartening and in fact seemed to imply that the swings were not in fact synced.

Spans like this imply that flurry is actually applied earlier than stated in the combat log, somewhere between my "synchronized swings" as the second swing at 20:06.07 is actually consuming the first charge of the Flurry supposedly applied at 20:06.08. This is evidenced by the fact that the combat log only lists two melee swings between the gaining and fading of flurry (these are the synchronized swings at 20.06:10).

So my hypothesis is that my swings weren't actually ever synced and were in fact apart by less than 2 milliseconds. But this whole process did leave me with one parting question. If melee swings were to be synchronized, how would they show in the combat log? Which would show first and why?

If anyone has any research or resources on the 4 swings theory, I'd love to see it. I'd especially like to know if I'm sacrificing more than a negligible amount of DPS using a 3.8 MH and 3.6 OH (Ashkandi and Shalug'doom for instance) as opposed to dual-wielding 3.6 or 3.8 speeds.

cbgoding · 03/01/11, 9:07 PM

Originally Posted by Narkanius

I had confirmed over the course of 10 minutes, that my melee swings were synchronized (assuming the lag between strikes wasn't greater than zero but less than .002 seconds, this is a major assumption). What I found on flurry procs was disheartening and in fact seemed to imply that the swings were not in fact synced.

Spans like this imply that flurry is actually applied earlier than stated in the combat log, somewhere between my "synchronized swings" as the second swing at 20:06.07 is actually consuming the first charge of the Flurry supposedly applied at 20:06.08. This is evidenced by the fact that the combat log only lists two melee swings between the gaining and fading of flurry (these are the synchronized swings at 20.06:10).

So my hypothesis is that my swings weren't actually ever synced and were in fact apart by less than 2 milliseconds. But this whole process did leave me with one parting question. If melee swings were to be synchronized, how would they show in the combat log? Which would show first and why?

If anyone has any research or resources on the 4 swings theory, I'd love to see it. I'd especially like to know if I'm sacrificing more than a negligible amount of DPS using a 3.8 MH and 3.6 OH (Ashkandi and Shalug'doom for instance) as opposed to dual-wielding 3.6 or 3.8 speeds.

Quite simply, if you have equal speed weapons and they're misaligned by less than 2 milliseconds, then the one that's behind will be hasted up until that 2 millisecond mark. Assuming a pair of 3.6 speed weapons and a difference of 2ms, then it is hasted for 99.9444% of its swingtimer. It is, for all intents and purposes, synchronized.

As for the amount of dps you'd lose, try modeling yourself in simcraft with a pair of Ashkandis or a pair of Kalugdooms.

EDIT: nevermind, didn't realize that you said flurry was being applied in between your swings, that pretty much nullifies the above.

Deathwing · 03/01/11, 9:37 PM

Narkanius, did you stand next to the dummy and initiate combat? Or did you charge it?

There's some weird bug where if you initiate combat right next to the target, your weapon swings will desync. If you charge in or start off with heroic throw(resets your swing timer), your swings will resync.

This is important for fights like Chimaeron.

Narkanius · 03/02/11, 1:35 AM

Originally Posted by Deathwing

Narkanius, did you stand next to the dummy and initiate combat? Or did you charge it?

There's some weird bug where if you initiate combat right next to the target, your weapon swings will desync. If you charge in or start off with heroic throw(resets your swing timer), your swings will resync.

This is important for fights like Chimaeron.

I initiated combat with white attacks, backed out for about 4 seconds and reengaged to sync the swings. It worked, but I decided to try your method of synchronization with heroic throw instead. It worked, at least as well as my method did, possibly even better as I've been at this for a while now and they are still hitting with the same timestamp. That being said, there are two cases that seem related to me.

Case #1
This is what I referenced earlier. The first melee strike (which I believe to my ilvl 359 MH strike because it does more damage than second melee strike, which I would imagine is my ilvl 346 OH) crits, but flurry does not appear on the combat log until after the second "OH" melee strike (which occurs at the same moment as the crit). However, the next set of synced melee strikes consumes the flurry proc. In other words, critmelee-flurry-meleemelee-flurry fades. I posted an image of such a log in my first go, but here's another:

Case #2
I just now noticed the case where the OH crits and procs flurry a second later. The next set of synced melee strikes reduces the charges to one. Finally, a second set of synced melee strikes occurs, but the combat log does not recognize flurry as fading until both attacks are posted. In other words, crit-flurry-meleemelee-meleemelee-flurry fades. Image:

I think all I've exposed is that the combat log may have a bit of lag to it. But can someone with more technical game knowledge approach this subject? Landsoul, being the one that first proposed the theory (at least I haven't seen it elsewhere) may know something I don't.

Gruntle · 03/02/11, 4:37 AM

Originally Posted by nca

I do not understand how the Flurry change is a nerf to Fury. This seems like a buff to me, however I should mention that this is a gut feeling. Is there any mathematical explanation for the change being a nerf?

Who's saying that the Flurry change is a nerf? That's not at all what's being discussed here. People are concerned that the change will bring problems in itemization and that stacking crit will be the so much better than the alternatives that little choice remains. Some people feel this is a bit boring.

On a separate note, does anyone have a working model (with detail on how things are calculated, if not the exact calculation methods) for stat weights including Str? The ones that I have seen around don't make much sense to me, as they seem to be grossly oversimplified.

Stat weight are simplifications by nature. The stat weights that have been discussed in this thread are mostly from SimulationCraft, I'm not sure you can calculate stat weights in a more accurate way than with this tool (unless, of course, there are errors in the code). These stat weights are based on running thousands of dps simulation runs and comparing what happens when you increase the different stats. Landsouls excel sheet uses analytical approximations to do the same (at least it used to do so, maybe there's some iterations being done in that sheet too), but he's still working on the Cata version of the sheet.

About synched weapon speed and flurry. It seems quite evident to me that flurry is applied and lost as it should. If you crit with the first of two synched swings flurry will be eaten up by the three next swings, thus giving you the flurry haste for one "between-swings-period". If you crit with the second of the synched swings it will still last for three swings, but will in this case give you haste for two "between-swings-periods". Never mind the combat log, it will not always show the exact time and order that things happen (I've seen similar stuff many times before in my logs).

What Narkanius testing shows is that synched weapons are not synched with infinite precision, the swings are not done in the exact same moment. Good work on the testing, maybe it has some implications for the simulation tools. I'm not sure how this is modeled in SimulationCraft, and it could maybe also change how flurry is modeled in spreadsheets. Taking this into account synched weapons will only give longer flurry uptimes in about half of the crits (only looking at white crits).

Urun� ag · 03/02/11, 11:02 AM

Just read the postings above. your gut-feeling trys to hide some facts for you to sleep well

.

enrage and deathwish both requires mastery.

base mastery -> 44,8 (8*5,6) => 11,2 (8*2)

thats a heavy nerf not only a "slight nerf" in my opinion.

with base-mastery your enrage moves from 14% dmg-boni (10 +44,8%) to 11% (10 + 11,2%).
with mastery of 15 (now, after the nerf 9) it would be 18% (10 +84%) vs 15% (10 +50,4%)
with mastery of 18 (now, after the nerf 12) it would be 20% (10+100,8%) vs 16% (10 +67,2%)

an overall 3% dmg-loss to enrage alone, not mentioned the same effect works on DeathWish (only with slightly higher numbers

).

jeah, with RB up to 120% you are close to the "live" stats of RB (some minor loss in weapon-dmg-%) but only if you use RB triggered with bloodlust. if you take the reduced effect of enrage (to trigger RB) into account or you trigger RB via DW the overall dmg-reduce should be quite obvious.

and the flurry-increase may help with higher "white-hit-dmg" and rage-buildup ... but it doesnt account for the overall-dmg-loss (again...). it only helps "damn i cant do dmg without smashing HS every 1,5 / 3,0 sec"-players... but sadly ... no one else.

Knova · 03/02/11, 11:09 AM

Will the CS nerf affect us much in a raid setting (assuming prot warrior, feral/balance, etc..)? I assume it will stack with Sunder additively and that it was largely meant as a PvP nerf and at worst we lose two GCDs up front before the initial CS.

It seems as though the Flurry buff would push hit to our next secondary stat after crit (given that haste is terrible and getting worse and the nerf to mastery). Rage gen should go up handily if I'm not mistaken and it seems viable that HS will move up and we might spec back into 3/3 Incite. I guess at this point it is all speculation until we can get some sims, but I think we could see HS moving up on the priority list a decent bit unless we end up with massive amounts of mastery to keep RB high.

(edit: I did forget that they buffed RB dmg% at the time of posting, so the difference might not be as bad as I thought but the post above makes some good points.)

Overall, I like the idea of the Flurry buff and smoother dps for PvE, but like a few have stated it is dangerously close to pigeon-holing us.

Ecatombe · 03/02/11, 11:10 AM

You guys are forgetting the CS reduction to 70% in your calculations of the chages between Live damage and PTR damage.
Atleast, you didn't list it in the "nerf" side.