Druid Changes Patch 3.1 - Page 31

• Adoriele · 03/06/09, 11:10 PM

Originally Posted by mallhoof

what about the talent trees? Does the spec change any due to IS buff, FF and FFF increased duration (picking up imp FF for the crit bonus), and starfall being buffed from the cooldown? And with the changes to mana regen is intensity or dreamstate the clear front runner? Is owlkin frenzy that much better by adding magic damage to it?

You should already be specced for iFF, you should already be casting IS, you should already be specced for Starfall, the Owlkin Frenzy change is just reverting the previous change on an earlier PTR push which made it prof on physical damage only, so no change from live. Intensity also has no net change, the buff to Intensity compensates exactly for the nerf to spirit regen.

[edit] Hamlet, the mechanics of NG are now such that almost all Wraths will be affected. Hence either all wraths are above 1s, or all are below. Hence the hard break.

leithkin · 03/06/09, 11:52 PM

Originally Posted by foxglove

While this sounds like a nice change overall, it seems like it would worsen the "Boomkin Annoyance" (not being able to cast Bear Form FFF on a target when a Balance druid's Improved FF is on the target.) If FFF could apply its damage and threat without overwriting the debuff, it wouldn't be a problem. I'm going to submit that feedback on the PTR.

How important is FFF to bear tanking threat generation? I've only ever played the other two specs, and then mostly balance, and I have to say that I'm personally looking forward to a minimized loss of DPS casting time with this increased duration.

Deathwing · 03/07/09, 12:58 AM

Originally Posted by leithkin

How important is FFF to bear tanking threat generation? I've only ever played the other two specs, and then mostly balance, and I have to say that I'm personally looking forward to a minimized loss of DPS casting time with this increased duration.

Not very much overall. The refresh bug annoys me in two situations:

1. You see ~5s left on FF and would like to refresh it now because you have the free GCD, but won't when it runs out. Not a biggie really.

2. A dodge streak or zealous mashing of the maul button(or both!) leads you to be rage starved and to keep your threat going while your rage rebuilds, FF is a nice ability to use. However, if you depleted your rage from mauling too much, your threat is likely fine. It's still annoying in times where you run out of rage, you know you COULD have used FF had it not been for a stupid bug that's lasted how many patches now?

Hmm, I didn't really answer your question. FF is not used in any bear threat rotations. Mangle > Maul > Lacerate to 5 > Swipe. Switch lacerate and swipe if two or more targets. I believe FF generates threat a bit under lacerate.

Darwexn · 03/07/09, 2:17 AM

I noticed earlier the discussion about the value of hit and exp for feral dps differing depending on which spreadsheet/calculation you use. It just seemed odd that Nightcrowler (I think it's spelled) came up with a value that if memory serves was more than 50% higher than str, yet, rawr for example has hit at only 75% the value. While I understand the slight difference in rotation is FB, the difference in value for the stat seems disproportionate since FB is only 6% of total damage. Personally, I hate how a miss can screw up a rotation a bit and cause human error on top of that from not noticing the miss/dodge and losing a few seconds of mangle/bleed time; but I generally follow a more FB free cycle. How do most people view this (yes or no FB cycle), because bleed crit seems to push even more towards higher rip over occasional FB. Currently, both rawr 1.9.3 and the latest beta suggest I not hit cap (I also tested best in slot gear).

foxglove · 03/07/09, 2:25 AM

Originally Posted by leithkin

How important is FFF to bear tanking threat generation? I've only ever played the other two specs, and then mostly balance, and I have to say that I'm personally looking forward to a minimized loss of DPS casting time with this increased duration.

Its importance in threat generation is nonexistent--for a mob being tanked normally in melee range. I get annoyed when I can't refresh it, but no big deal when the boomkin is attentive. (Kinda sorta big deal when the boomkin is not, but that's the boomkin's fault. I have one in mind.) I also tend to use it as filler when I'm empty on rage, as Deathwing mentions, and having a balance druid in the raid prevents that.

It's very valuable to me, however, for picking up mobs at range that spawn or aggro. Most druids are in the habit of casting FF on a mob as soon as it floats on in (and I admit I have the same habit in cat form) because it's so little threat, right? Which is fine until it also happens to prevent the tank's ranged pull.

Originally Posted by Mijae

I'm more concerned about this in PvP. While it's great for my FF to last longer, it will be ridiculous to be on the receiving end in BGs. 2 minute Hunter's Mark is bad enough.

And you can actually stealth with Hunter's Mark; you just have a big red arrow on your head. As usual, I didn't even consider the PVP implication. Ironically, it's a nerf to druids... rogues can just cloak the damn thing off. 5 minutes of no stealth without someone to come dispell your kitty tail seems really unreasonable. Edit: of course, it's not likely to be a whole 5 minutes when you can't stealth, because death clears your debuffs nicely. :P

Edit: So folks are telling me that only the PVE duration is changing and the PVP duration remains 40s. That makes a lot more sense.

nightcrowler · 03/07/09, 2:41 AM

Night's Cat Simulation

I've redone my simulation with the latest PTR push:

+30% haste from items
+25% ArP from items
100% critting dots

I've also redone the cycle completly. Now I've found a decisional algoritm based on a formula that try to find the optimal time to use FB/RIP/SR and it seems to work in a large array of stats value, that way the cycle adjust itself while upgrading the gear.

New stats value and cycle:

Expertise 40,0

Agility 26,8

Armor Pen 26,8

Feral AP 12,0

Strengh 23,8

Crit 22,8

AP 10,0

Haste 18,1

For sake of knowledge -> 20 points = +1 DPS.

Cat (crit% vs. a boss (-4.8% added)): 51.9659%
Cat (AP): 10745
Expertise: 26
Hit: 7.01434%
Armor Penetration: 13.9653%
Haste (from equipment): 10.6648%

Boss base Armor (before penetration and debuff): 11900
Boss modified Armor (after penetration and debuff): 7678.6
Boss Damage Reduction (after debuff and penetration): 33.5148%
Weapon Speed Reduction % (Total without Bloodlust): 27%

DPS: 6664.1 TPS: 4731.87

White: 28.1877%
Shred: 27.3967% (Avg dmg: 7469.59, #/cycle: 7.31439),
RIP: 21.9695% (Avg dmg: 42921.4, #/cycle: 1.02075)
Rake: 17.4662% (Avg dmg: 12971, #/cycle: 2.68535),
Ferocious Bite: 4.97984% (Avg dmg: 15141.9, #/cycle: 0.65586, Average Energy Usage: 40)
Savage Roar uptime: 99.057%

Average Cycle Time: 29.922

The cycle is about 3SR/5RIP/5FB. The decision about RIP & FB & SR is made via a formula that take care of your stats, ability damage, CP generation and so on. I've a mangle bot and bleeds always up in my simulation (that's what I usually have in raid). Berserk used every 3 minutes, Heroism every 10 minutes.

Here is the SR/RIP/FB formula

Fb_damage = (1+RA*0.02)*(fb.hit*(fb.adamage+fb.bdamage)+fb.crit*(fb.acritdamage+fb.bcritdamage));
						fbextra_damage=(1+sr*src)*((1-red)*(1+RA*0.02)*(fb.hit*((ap*0.002439)+(9.4))+fb.crit*(((ap*0.002439)+(9.4)))*2*(1+pistinct_m)*(1+META*0.03))*(1+naturalist_m)*(1+ferocity_m));
						rip_damagextick = (1-wcrit)*(1+sr*src)*((1+RA*0.02)*((ap*0.01*cp)+(39+21.0*IDOL_WS*cp+99*cp))*(1+naturalist_m)*(1+mdebuff*0.3))+wcrit*(1+sr*src)*((1+RA*0.02)*((ap*0.01*cp)+(39+21.0*IDOL_WS*cp+99*cp))*(1+naturalist_m)*(1+mdebuff*0.3))*2.0*(1+pistinct_m)*(1+META*0.03);
						
						avg_cp[5]=(1/(shred.crit+shred.hit))*((5*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)+4*4*shred.crit*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)+5*shred.crit*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)+3*3*shred.crit*shred.crit*(1-shred.crit)+3*shred.crit*shred.crit*shred.crit+3*4*shred.crit*shred.crit*(1-shred.crit)*(1-shred.crit)));
						avg_cp[4]=(1/(shred.crit+shred.hit))*(4*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)+3*3*(1-shred.crit)*(1-shred.crit)*shred.crit+3*2*(1-shred.crit)*(shred.crit)*shred.crit+2*(shred.crit)*shred.crit);
						avg_cp[3]=(1/(shred.crit+shred.hit))*(3*(1-shred.crit)*(1-shred.crit)*(1-shred.crit)+2*2*shred.crit*(1-shred.crit)+2*shred.crit*shred.crit+3*(1-shred.crit)*(1-shred.crit)*(shred.crit));
						avg_cp[2]=(1/(shred.crit+shred.hit))*(2*(1-shred.crit)*(1-shred.crit)+2*(1-shred.crit)*shred.crit+shred.crit);
						avg_cp[1]=(1/(shred.crit+shred.hit))*(1);
						
						
						 tick_num_FB=(Fb_damage+5*fbextra_damage)/rip_damagextick;
						
						
						 cp_to_cycle=avg_cp[sr_t]+avg_cp[rip_t]; 
						 max_rip_downtime= tick_num_FB*2;
						 energy_x_sec= (10*30)/30+ooc_rate*30/(speed/100.0);
						ripvsfb_time[0]=((0.55+0.18*manglebot)*shred.energy+(0.225+0.045*manglebot)*rake.energy+0.225*mangle.energy*(1-manglebot))*(cp_to_cycle/energy_x_sec)-max_rip_downtime+(sr_energy+rip.energy)/energy_x_sec;
						ripvsfb_time[1]=((0.55+0.18*manglebot)*shred.energy+(0.225+0.045*manglebot)*rake.energy+0.225*mangle.energy*(1-manglebot))*(cp_to_cycle/energy_x_sec)-max_rip_downtime+(sr_energy+rip.energy)/energy_x_sec-koj_m1/10.0;
						srvsfb_time[0]=((0.55+0.18*manglebot)*shred.energy+(0.225+0.045*manglebot)*rake.energy+0.225*mangle.energy*(1-manglebot))*(avg_cp[sr_t]/energy_x_sec);
						srvsfb_time[1]=((0.55+0.18*manglebot)*shred.energy+(0.225+0.045*manglebot)*rake.energy+0.225*mangle.energy*(1-manglebot))*(avg_cp[sr_t]/energy_x_sec)-koj_m1/10.0;

And here the decision making:

SR?

					if((free_sr==1)&&(((sr==0)&&(en>=sr_energy)&&(cp>=sr_t))||((sr_timer<=(srvsfb_time[0]*100))&&(en>=sr_energy)&&(cp>=sr_t)))){

FB?

					if((en>=fb.energy)&&(cp>=fb_t)&&(((bdebuff==1)&&((rip_timer>=(ripvsfb_time[0]*100))||((koj_timer<=ripvsfb_time[0]*100)&&(rip_timer>=(ripvsfb_time[1]*100)))))||(rip_t>5))){

RIP?

if(((en>=(rip.energy+shred.energy+rake.energy-20-((int)(lag/10.0))))&&(cp>=rip_t))||((((rip_expire==0)&&(((en>=rip.energy)&&(cp>=rip_t)&&(fb_t>5))||((en>=rip.energy)&&(cp>=rip_t)&&((rip_timer<=(ripvsfb_time[0]*100))||((koj_timer<=ripvsfb_time[1]*100)&&(rip_timer<=(ripvsfb_time[1]*100))))&&(doublecycle_counter==1)&&(fb_t<=5))||((en>=(rip.energy+shred.energy+rake.energy-20-2*(int)(lag/10)))&&(doublecycle_counter==0)&&(cp>=rip_t)&&(fb_t<=5))||((en>=rip.energy)&&(cp>=rip_t)&&(fb_t<=5)&&(bdebuff==0))))||((rip_expire==1)&&(rip_timer<0)&&(((en>=rip.energy)&&(cp>=rip_t)&&(fb_t>5))||((en>=rip.energy)&&(cp>=rip_t)&&(doublecycle_counter==1)&&(fb_t<=5))||((en>=(rip.energy+shred.energy+rake.energy-20-2*(int)(lag/10)))&&(doublecycle_counter==0)&&(cp>=rip_t)&&(fb_t<=5))||((en>=rip.energy)&&(cp>=rip_t)&&(fb_t<=5)&&(bdebuff==0))))))){

Basically:

- Compute average hit done to put:
1, 2, 3, 4, 5 CP based on your hit/exp/crit/haste

- Compute the average energy usage based on ability cost (it change with berserk) and ooc procs (it change with haste).

- Based on energy cost and #of hits to generate cp you can compute average time to generate x cp.

- Based on that you will have TF off CD before that time end -> subtract 60 energy.

- Then I've computed the average lost on losing rip uptime based on a 40 energy FB / RIP ticks = number of tick loss before RIP > FB. I've not taken into account the 10 extra energy cost for FB because it average out with the gain of not doing shred without cps gain.

So now each time you have 5 cps you can choce to do SR, RIP or FB. You will do SR if SR will expire before you'll be able to reapply it, you will do FB if the RIP uptime lost will be small enough or otherwise you'll wait rip to expire and then reapply it.

3 SR is an empirical value but it have a meaning: with about 5% crit, 3 sr minimum will give you = 9+5*3 = 24 seconds or 9+5*4=29 seconds time that's about the cycle time and on average SR will last the same time as RIP+1 GCD, exactly what we need.

Best in slot items with those new values:

Best in slot dps (exp, hit):

chest: chestguard of the reclusive (83,0)
boots: dawnwalker (0,50)
Hands: frosted androit (0, 58)
head: Valorous Dreamwalker Headguard
legs: Sarth 2d (0,66)
shoulder: Valorous Dreamwalker Shoulderpads
belt: Belt of the tortured (0,50)
bracers: Sinner's Bindings
cloak: Drape of the Deadly Foe
ring1: Strong-Handed Ring (0,28)
ring2: Ruthlessness (32,0)
neck: Favor of the Dragon Queen

83+32 = 115 exp. (24 expertise = 16+8 = 24 = 1 gemma con 16 expertise (Precise Scarlet Ruby
) con 8exp/12 stam (Guardian's Twilight Opal)
50+58+66+50+28 =252 hit (11hit = 12 hit / Icewalker sugli stivali)

Altre gemme: + 1 gemma con 8str/8crit per averne una gialla (Inscribed Monarch Topaz ) + XXX x Delicate Scarlet Ruby (16 agi)
Cibo: 40agi/40stam (Blackened Dragonfin)

Enchant:

chest: powerful stats (+10 all)
boots: icewalker (12 hit/12 crit)
Hands: 20agi / Scroll of Enchant Gloves - Major Agility
head: reputazione Ebon Blade / Arcanum of Torment
legs: patch epico / Icescale Leg Armor
shoulder: reputazione Son's of hodir / Greater Inscription of the Axe
bracers: Sinner's Bindings / greater assault
cloak: (22 agi) / Scroll of Enchant Cloak - Major Agility
weapon: berserking.

foxglove · 03/07/09, 3:14 AM

Originally Posted by Darwexn

I noticed earlier the discussion about the value of hit and exp for feral dps differing depending on which spreadsheet/calculation you use. It just seemed odd that Nightcrowler (I think it's spelled) came up with a value that if memory serves was more than 50% higher than str, yet, rawr for example has hit at only 75% the value. While I understand the slight difference in rotation is FB, the difference in value for the stat seems disproportionate since FB is only 6% of total damage. Personally, I hate how a miss can screw up a rotation a bit and cause human error on top of that from not noticing the miss/dodge and losing a few seconds of mangle/bleed time; but I generally follow a more FB free cycle. How do most people view this (yes or no FB cycle), because bleed crit seems to push even more towards higher rip over occasional FB. Currently, both rawr 1.9.3 and the latest beta suggest I not hit cap (I also tested best in slot gear).

Yes; I notice that Night's latest simualtion just above has really, really high numbers for hit and expertise compared to what I'm used to seeing from other calculators.

Hit is something I've gone back and forth about a lot. I've gone through phases where I go out of my way to get +hit pieces and gemmed a bit excessively toward the +hit side of things. (I used to play a rogue). Right now, while I wish that hit capping were a good DPS choice, I gem strength everywhere on live (which would become agility in 3.1). If I'm given two pieces of comparable DPS gear, I will usually choose the one with more hit or expertise, but that's the extent of it.

One thing that helped me a great deal with the "human error" aspect of running well below the hit cap is an addon called Mikk's Scrolling Battle Text. It's highly customizable, and you can configure triggers for all sorts of notifications, including sound effects. Now, when Mangle, Rip, or Rake misses, I hear a very distinct noise (the one the crystals in BEM make when they zap you for messing up Simon) and I get a giant message with red text and an icon, e.g., "RIP MISSED." (I also have triggers to announce when Mangle, Rip, or Savage Roar fade, so that I'm less likely to completely blow my DPS rotation if I get distracted from my timers.)

My DPS is quite Rip-centric, because I still use 4pT6 even in a full cat spec. (I don't have all the BIS DPS gear since my main role is as a tank.) I still work in FBs when I Berserk or have a good run of crits, but on nights when I'm "off my game" and make bad decisions about FB, my DPS plummets. Prioritizing Rip with very conservative use of FB certainly seems more reliable.

dukes · 03/07/09, 5:33 AM

To the FF duration impact on PvP: MMO-Champion BlueTracker - Faerie Fire 5 mins duration? Not PvP, right?

5 min PvE. 40 sec in PvP.

IS testing (all without E+M) but in Moonkin Form so Master Shapeshifter and E+M (3% spelldamage personal bonus) apply:

42 spellpower (naked, spellpower bonus from int/spirit only): 240 per tick.
502 spellpower: 338 per tick.
1032 spellpower: 451 per tick.
1513 spellpower: 554 per tick.
2013 spellpower (+2t7 bonus): 728 per tick.

More testing, this is with just 21 points in Balance for IS so no extra multipliers get in the way:

0 spellpower: 215 per tick
499 spellpower: 315 per tick
1023 spellpower: 420 per tick
1480 spellpower: 511 per tick
1843 spellpower: 642 per tick

As far as I can tell, IS now has a 0.2 coefficient (per tick) from Spellpower, or a 1.2x coefficient over 6 ticks (making it 1.4x over 7 ticks). Hopefully someone who actually knows what they're doing can confirm this. Going off Wrathcalcs latest version, the current IS coefficient is 0.13/tick I think, so it's a pretty significant boost to IS. Between this change, the Starfall changes, the NG changes and the eclipse change, it'll be interesting to see which glyphs are actually of most benefit now.

nightcrowler · 03/07/09, 6:01 AM

About high value of hit/exp in Night's Sim

First of all, I'm 99.99% sure that the simulations are ok. I've tested cycles on live and they work fine.
Second. If you don't use FB, hit=exp=38, still high.

Now, let's try some math (i'm simplifying value for readibility).

How hit works?

-white damage (30%): assuming for simplification 50% crit and 25% glance = 25% normal hit, with 1-roll system only those hits are affected by hit.

-1% hit: 0.01*0.25*0.3 = 0.08% less dps (I'm avoiding ooc proc math)

- Rip+rake (40%): assuming no clipping, if you miss a rip or a rake you are missing 1 GCD of their dps, so

-1% hit: 0.01*0.4 = 0.4% less dps

- All yellow attacks (70%): if you miss you are wasting energy. If we assume an average energy consumption of 38 energy and an energy regain x secods of about 17 energy. We have 0.2*38= 7.6 energy wasted. 7.6/16 = 0.47 = 48% dps loss due to energy. So:
-1% hit = 0.01*0.7*0.48 = 0.34%

I've ignored in my math cylce screw up and ooc procs.

Sum it up = 0.34%+0.4%+0.075% = 0.82%

Now: 0.82% of 6665 = 55 dps

1% hit is about 32 hit rating so 1 hit rating = 55/32 =1.72 dps

In my prevois table as I said 20 points = 1 dps

So: 1 hit = 20*1.72 = 34.4 points.

As you can see the difference is marginal and the value is still high, the difference is made up by cycle screw up and ooc missed procs.

Candela · 03/07/09, 6:17 AM

So, here are my single-target boss tanking calculations for SD:

--------------------------------

base values:

attp = your attack power
crit = your crit chance
avoidance = your avoidance (dodge + miss)
armor_live = your armor on live
armor_ptr = your armor on ptr
swing_mob = mob swing timer, in seconds
swing_you = your swing timer, in seconds (usually 2.5 seconds)

--------------------------------

intermediate values:

-> damage after mitigation change (against lvl 83)
dmg_ratio_live = PotP * 16635/(armor_live+16635)
dmg_ratio_live = 0.88 * 16635/(armor_live+16635)

dmg_ratio_ptr = PotP * 16635/(armor_ptr+16635)
dmg_ratio_ptr = 0.88 * 16635/(armor_ptr+16635)

dmg_ratio = dmg_percent_ptr/dmg_percent_live
dmg_ratio = [0.88 * 16635/(armor_ptr+16635)] / [0.88 * 16635/(armor_live+16635)]
dmg_ratio = (armor_live+16635)/(armor_ptr+16635)

-> attack ratio (during a swing of the target mob, how many attacks do you do on average in return)
attacks_mob = timeframe/swing_mob
attacks_you = timeframe/swing_you + timeframe/1.5

attack_ratio = attacks_you/attacks_mob
attack_ratio = (timeframe/swing_you + timeframe/1.5)/(timeframe/swing_mob)
attack_ratio = swing_mob*(1/swing_you + 1/1.5)

--------------------------------

Probability for SD to be up on an incoming landing hit:

SD will only be up if you had a crit before you get hit, this means only when:
(1). Target hits you
(2). Target misses you X times (X can be 0)
(3). You hit the target while it misses you, and at least one of your hits is critical
(4). Target hits you

Going with the premise that the Target hits you on (1) and (4).

The chance for you to dodge X times is:
prob_dodge(X) = (1-avoidance) * avoidance^X

You hit the target (X+1)*attack_ratio times on average for any given X.
The chance that at least one of those hits is critical is:
prob_crit(X) = 1 - (1-crit)^((X+1)*attack_ratio)

So, the chance for SD to be up on the incoming hit on (4) is:
SD_ratio = sum over X (for X = 0 to infity) of prob_dodge(X)*prob_crit(X)

Since SD has only a 10 second duration (tested on PTR right now), the highest X u have to account for is:
max_X = floor(10/swing_mob)
Any X higher than this is not relevant, since a SD that was up for a miss before these max_X misses procced more than 10 seconds before the hit on (4) and therefore expired already.

--------------------------------

With this, we can calculate the breakeven point at which SD gets worse than the armor which we have on live:

BEP:
dmg__taken*dmg_ratio - SD_ratio*(attp*0.25) = dmg_taken
dmg_taken*(dmg_ratio-1) = SD_ratio*(attp*0.25)
dmg_taken = SD_ratio*(attp*0.25)/(dmg_ratio-1)

================================

giving an example with my own stats from offtanking Patchwerk.

armor_live = 34032 (selfbuffed)
armor_ptr = 28491 (selfbuffed)
attp = 5670 (selfbuffed + BoM)
dmg_ratio = (34032+16635)/(28491+16635) = ~1.1228
crit = 0.3458 (char info from paperdoll sheet in bear form)
avoidance = 0.527 (read from WWS)
attack_ratio = 212/129*(1/2.38+1/1.5) = ~1.786 (129 hatefuls thrown at me during a 212 second fight, my autoattack speed is 2.38)

max_X = floor(10/[212/129]) = 6

prob_dodge(0)*prob_crit(0) = 0.4730*0.5313 = 0.2513
prob_dodge(1)*prob_crit(1) = 0.2493*0.7804 = 0.1946
prob_dodge(2)*prob_crit(2) = 0.1314*0.8971 = 0.1179
prob_dodge(3)*prob_crit(3) = 0.0692*0.9518 = 0.0659
prob_dodge(4)*prob_crit(4) = 0.0365*0.9774 = 0.0357
prob_dodge(5)*prob_crit(5) = 0.0192*0.9894 = 0.0190
prob_dodge(6)*prob_crit(6) = 0.0101*0.9950 = 0.0100

-> SD_ratio = 0.6944

dmg_taken = 0.6944*(5670*0.25)/(1.1228-1)
dmg_taken = 0.6944*1417.5/0.1228
dmg_taken = ~8016

so if i take more than 8016 dmg on average from the hatefuls SD will be a nerf for offtanking Patchwerk.

Originally Posted by Soultrigger

Even though bleed is affected by resilience, it is only reduced dmg dot applyed if it doesn´t crit. If it crits, it is the crit reduction that is applyed. There is no double resilience penalty.

Tested on PTR yesterday, 5-stack lacerate:

Target with 0 resilience: normal ticks for 551, critical ticks for 1101
Target with 459 resilience: normal ticks for 520, critical ticks for 912

459 resilience -> -5.6% dmg from DoT, -12.32% dmg from critical hits

520/551 = 0.9437 = 1 - 0.0563 =~ 100% - 5.6% (equals 5.6% DoT dmg reduction)
912/1101 = 0.8283 = 1 - 0.1717 =~ 100% - 17.17% (equals 5.6% AND 12.32% reduction, so both DoT and critical hit)

To be exact, Resilience applies a number of penalties:
#1 reduces periodic damage
#2 reduces chance to critical hit
#3 reduces the effect of mana drains
#4 reduces the damage from critical strikes

Normal hits don't receive any penalty
Normal periodic ticks receive 1 penalty (#1)
Critical hits receive 2 penalties (#2 #4)
Critical periodic ticks receive 3 penalties (#1 #2 #4), as you can see from the test

Edit: These big lines weren't intended, seems the forum formats = signs at the start of a line in an odd way.

tr33hugger · 03/07/09, 7:14 AM

I have to admit i really like the change to Nature's Grace and Eclipse. I have 516 haste rating which leads to 2.44 SF and 1.22 wrath casts. The new NG reduces this to 2.04 / 1.02. Now i don't have my exact numbers from live since i'm feral atm, but i think they are about the same for SF and less for Wrath ofc. The real benefit is the duration of NG. If you crit once your next two SF or next three Wraths are hasted, which is huge. I put on my wrath idol and tested a little with SF (only to trigger eclipse) / wrathspam while keeping up dots and managed to sustain roughly 4,2k dps on the dummy (compared to ~3.5k on live iirc). GCD clipping with wrath is gone as long as you stay below a certain amount of haste (ignoring heroism for now), that amount is somewhere between 400 and 500 if you count in WoA totem.

edit : found a friendly shaman and with WoA i had .998 NGed wraths with 407 haste, so it's more like 400.

Taudark · 03/07/09, 7:20 AM

Crits on bleeds don't proc Savage Defence on PTR at the moment.

Starfox · 03/07/09, 7:55 AM

Originally Posted by dukes

42 spellpower (naked, spellpower bonus from int/spirit only): 240 per tick.
502 spellpower: 338 per tick.
1032 spellpower: 451 per tick.
1513 spellpower: 554 per tick.
2013 spellpower (+2t7 bonus): 728 per tick.

(553-240)/(1513-42)/1.03/1.04 = 0.1986
(728-240*1.1)/(2013-42)/1.03/1.04/1.1 = 0.1998
So yea, 0.2 it seems to be

Originally Posted by dukes

More testing, this is with just 21 points in Balance for IS so no extra multipliers get in the way:

0 spellpower: 215 per tick
499 spellpower: 315 per tick
1023 spellpower: 420 per tick
1480 spellpower: 511 per tick
1843 spellpower: 642 per tick

Your value for 1843 spellpower is way off here: (642-215)/1843 = 0.231687 perhaps a trinket procced and you didn't notice, because with a coeff of 0.2, it would need 2135 sp to get 642ish ticks.
rest matches nicely

(511-215)/1480 = 0.2

Wulfmanjack · 03/07/09, 8:10 AM

On PTR it seems we get more out of expertise.
I have 25 on live and 29 on ptr (from 6,25% to 7,25%)
On Char Screen its says on both 125 expertiserating.

in my case from Gear only its a gain from 3,75% to 4,75% (26,67%more out of expertiserating)

So with max expertisegear on i have on live 304 expertise Rating 9,25% and on PTR 11,5% (24,32% more out of expertiserating)

I guess its 25% and can vary with rounding

I also testet 4t6 on test dummy. Unbuffed the damadge was a bit lower than with with my normal gear (4300 dps with 4t6 and with normal gear 4400)

Rip damadge with this Set is insane (over 50k unbuffed) Raibuffed and on the Boss Loatheb its maybe possible to do 100k Rips... Scaling is to high atm on Ptr even without 4T6.

Valerian · 03/07/09, 8:35 AM

Originally Posted by nightcrowler

About high value of hit/exp in Night's Sim

First of all, I'm 99.99% sure that the simulations are ok. I've tested cycles on live and they work fine.
Second. If you don't use FB, hit=exp=38, still high.

Now, let's try some math (i'm simplifying value for readibility).

How hit works?

-white damage (30%): assuming for simplification 50% crit and 25% glance = 25% normal hit, with 1-roll system only those hits are affected by hit.

-1% hit: 0.01*0.25*0.3 = 0.08% less dps (I'm avoiding ooc proc math)

- Rip+rake (40%): assuming no clipping, if you miss a rip or a rake you are missing 1 GCD of their dps, so

-1% hit: 0.01*0.4 = 0.4% less dps

- All yellow attacks (70%): if you miss you are wasting energy. If we assume an average energy consumption of 38 energy and an energy regain x secods of about 17 energy. We have 0.2*38= 7.6 energy wasted. 7.6/16 = 0.47 = 48% dps loss due to energy. So:
-1% hit = 0.01*0.7*0.48 = 0.34%

I've ignored in my math cylce screw up and ooc procs.

Sum it up = 0.34%+0.4%+0.075% = 0.82%

Now: 0.82% of 6665 = 55 dps

1% hit is about 32 hit rating so 1 hit rating = 55/32 =1.72 dps

In my prevois table as I said 20 points = 1 dps

So: 1 hit = 20*1.72 = 34.4 points.

As you can see the difference is marginal and the value is still high, the difference is made up by cycle screw up and ooc missed procs.

This looks fine at a glance to me. The real question is then why does both the new Rawr (2.2.0b3) AND Tossk's 3.1 simulator have Hit and Expertise as significantly lower than Agi/Str?

Rawr (Fully raid buffed and specced):
Agi=1.19->2.53/AP
Str=1.12->2.38/AP
ArP=0.97->2.06/AP
Crit rating=0.94->2/AP
Hit rating/Expertise=0.85->1.81/AP
Haste=0.66 (though I don't know if this includes the new 30% more Haste)->1.40/AP
AP=0.47 (0.94 for equivalent stat values)->1/AP (clearly)

Tossk
Str:AP = 2.3786
Agi:AP = 2.4404
Crit Rating:AP = 1.9094
Hit Rating:AP = 1.3433
Expertise Rating:AP = 1.3433
Haste Rating:AP = 0.9768
Armor Pen Rating:AP = 1.6062

Its possible again Tossk's hasn't included the buffs to ArP or Haste but I can't imagine that would effect the other stats rankings that much.