FeralbyNight (Cat & Bear simulation tool) - Page 19

Smartie · 05/12/09, 8:46 AM

Originally Posted by nightcrowler

Yes. It show tiger's fury on the left part of the suggester frame, it doesn't mean that you should use it but only that it's "reloaded". If you feel strange about it I'll try to change it in a more intuitive way. As for the FFF bug, do player with that bug have a different language (not english?), it could be a localization error.

It was in the center suggestion bit. Not on the sides where it says if you have TF and Berserk ready.

I'll check if I have the lastest version when I get home.

Iwiperaids · 05/12/09, 9:32 AM

Hello everyone, I am new here. I am reading this amazing thread for some time now and using the FBN addon and it's amazing.

However one thing I cannot find is the simulation tool, where do I download that? On FBN's google site there is a simulator I believe but I cannot use it as it's extension is "ccp" or something.

If someone could direct me on how and where to get the simulator I would appreciate.

Thanks and keep up the good work.

KENisFIS · 05/12/09, 9:39 AM

I will look into it when I'm tanking tonight. Ill see if i can get a video of it for you to look over. It may have just been my poor threat building then, since FFF is on the GCD. i may have been missing the timing and then burning another GCD on something else thinking the addon was bugging.

KENisFIS · 05/12/09, 9:42 AM

Originally Posted by Smartie

It was in the center suggestion bit. Not on the sides where it says if you have TF and Berserk ready.

I'll check if I have the lastest version when I get home.

I noticed this too in the older one. But it works correctly in the most recent download. So that might be the issue.

nightcrowler · 05/12/09, 10:21 AM

I've runned an updated bear simulation.
I've also updated the simulator (you can found the link on the main page)

Bear simulation (last update May, 12 - 2009)

Tank Build used: 0/60/11

The simulation takes all tanking talents + the new Primal Gore talent and Improved Mark (you can do that).
I'm using best in slot item except hard modes and pvp gear
I'm using Noble Deck agility, Heart of Iron and Mongoose.
I'm using the new mangle idol and 4T8 (legs, chest, shoulders and gloves).
I'm also using all possible raid buff + flask + food.
I'm not using priest/paladin shields.
I'm using Ispiration (+25% armor buff), barkskin and survival instinct.

I'm also using 5/5 ferocity over 2/3 King of the jungle and 3/3 Improved Mangle

------

Priority List
There isn't a real rotation, it's more a priority list.

1) Use mangle when not in CD
2) Use Maul always
3) Keep 5 stack of lacerate up (you will use lacerate every 13.5 second to minimize lacerate stack loss due to miss/parry/dodge
4) Use FFF to fill in and not in CD.
5) Use swipe to fill in.

In the following report I've assumed 100% mangle debuff uptime (manglebot) and 100% bleed effect (if you have a manglebot it's likely that there is a bleed on the target).

BEAR SIM report:

5 minute fights.
100,000 Simulations.
Time step: 0.01 seconds.
Latency: 90ms.

Boss Data
_______________________

Swing timer (before Infected wounds): 1.7 seconds
Swing damage (before Demoralizing Roar and Improved DemoR): 70000 damage
Special attacks: none.
Parry Gibbs: enabled.

Aggressive Stats after all the buffs (full stacked raid + Bloodlust used 1 time in a 5 min fight)
_______________________

Bear (crit% vs. a boss (-4.8% added)): 42.1106%
Bear (AP): 7530
Expertise: 50
Hit: 6.22141%
Armor Penetration: 7.46981%
Haste (from equipment): 6.30375%
+weapon damage: 0
Shield value = 1882

Boss base Armor (before penetration and debuff): 10643
Boss modified Armor (after penetration and debuff): 7508
Boss Damage Reduction (after debuff and penetration): 33.016%
Weapon Speed Reduction % (Total without Bloodlust): 24%
Missed % (Total): 1.77859%
Dodged %: 0%
Not normalized attacks parried by the boss: 1.25297%
_____________________________________

DPS: 3610.69 +/- 0.260738, (Min-Max variance: 21.962%, Lower dps sample: 3181.25, Upper dps sample: 3974.23) , Cat: 0%, Bear: 100%, TPS: 7950.88

White bear: 0, Yellow bear: 3610.69
White bear: 0%
Mangle bear: 10.0825%, Average Damage: 2736.78,
Maul: 64.9659%, Average Damage: 4390.38,
Lacerate: 15.3455%, Average Damage: 5416.44, (for application)
Swipe: 4.75186% Average Damage: 578.137,
FeralFairieFire: 4.85428% Average Damage: 1475.71

Boss crit: 0%, Boss hit: 36.6546%, Boss miss: 63.3454%, Boss white dps: 2450.12, Boss average weapon speed: 2.02798 sec., Boss average parried attacks x sec: 1.37913%, Average Boss White DPS increase due to parry: 0.589146%
Chance to die: 0.0160833% without heals in: 5 seconds

_____________________________________

Defensive Stats after the buffs (vs. a raid boss, full stacked raid + Devotion Aura)
_______________________
Bear armor (before inspiration): 31707
Physical Damage Reduction Bear (before inspiration): 70.6267%
Defense: 433
Chance to be critical hit: 0%
Dodge (bear): 49.9781%
Miss: 10.0518%
Total Avoidance (bear): 60.0298%
Total HP (bear): 49455
HP soft cap (bear): 42352.3
Shielded hits: 90.7053%
Inspiration uptime: 99.5267%
Tank Points (hp equivalent with 0% avoidance, 0% Damage Reduction and without parry from the boss) (bear): 699780 +/- 243.94
_____________________________________

Stats Value

Stats	Tank Points	STAM-EQ	DPS	APEQ DPS	TPS	APEQ TPS	NIGHT's CHOCE
AP	12.5	5.5	0.37	10.0	0.73	10.0	10.5
Feral AP	15.0	6.6	0.44	12.0	0.88	12.0	12.6
Agility	282.6	124.4	0.34	9.1	0.66	8.9	128.9
Strengh	29.6	13.1	0.87	23.8	1.75	23.8	24.9
Crit	11.1	4.9	0.53	14.3	1.03	14.0	11.9
Haste	-9.5	-4.2	0.85	23.0	1.93	26.2	8.9
Hit	7.3	3.2	0.94	25.5	2.18	29.7	18.1
Expertise	151.5	66.7	0.94	25.6	2.18	29.6	81.5
Armor Pen	0.0	0.0	0.85	23.1	1.85	25.2	12.6
Defense	209.7	92.3	0.00	0.0	0.00	0.0	92.3
Dodge	222.7	98.0	0.00	0.0	0.00	0.0	98.0
Armor	118.9	52.3	0.00	0.0	0.00	0.0	52.3
Bonus Armor	16.8	7.4	0.00	0.0	0.00	0.0	7.4
Stamina	227.1	100.0	0.00	0.0	0.00	0.0	100.0
Health	13.8	6.1	0.00	0.0	0.00	0.0	6.1

Tank Points are tank points: Health * (boss expected damage with 0 mitigation and 0 avoidance)/(damage taken)
STAM-EQ: are stamina equivalent tank points with stamina value fixed to 100.
DPS: is the dps increase for 1 point of that stat.
TPS: is the threat per second increase for 1 point of that stat.
APEQ DPS and APEQ TPS: are the tps and dps equivalent with attack power value fixed to 100.
NIGHT's CHOCE: (STAM-EQ)+0.5*(APEQ TPS) (for the average lether gear this mean about 7.5% of gear value based on threat)

Expertise value is based on the following assumption:
you are already at the soft cap (26 exp), otherwise you should add hit value to expertise value.
Mobs have parry haste enabled, for bosses without parry haste you should use hit value instead of the expertise one (if you are over the soft cap, otherwise you should use hit*2)

You can found "Night's choice" value of all stats and item rankin here:
Lether
Weapons
Back, fingers, trinkets, neck

BiS Ulduar (without hard modes) gear

Slot	Item
Head	Garona's Guise
Neck	Titanstone Pendant
Shoulders	Conqueror's Nightsong Shoulderpads
Chest	Conqueror's Nightsong Rainment
Waist	Belt of the Twilight Assassin
Legs	Conqueror's Nightsong Legguards
Feet	Runed Ironhide Boots
Wrist	Thrusting Bands
Hands	Conqueror's Nightsong Handgrips
Finger1	The Leviathan's Coil
Finger2	Platinum Band of the Aesir
Trinket1	Heart of Iron
Trinket2	Darkmoon Card: Greatness (Agility)
Back	Cloak of the Shadowed Sun
MainHand	Twisted Visage + mongoose
Idol	Idol of the Corruptor
Sockets	Solid Sky Saphire

Professions: JC+Enchanting (3x41 stam gems and 2x24 stam enchants).

Discussion
About capping value: hit caps at 8%, spell hit caps at 17% (used for demo roar and FFF). Expertise soft cap (dodge) is at 26 expertise, hard cap (parry) is at 56 expertise (-14% chance to be parried). Armor soft cap is at 40k.
HP soft cap: I wanted to put an indication of a good hp mark. My choice was to pick the hardest hitting boss in Ulduar and consider how much hp we need (without inspiration and parry gibbs) to survive 5 seconds without heals.
Parry gibbs: It's the main reason of wiping for a tank, they are the new crushing blow, that's why EXPERTISE is really really important, removing the chance to be parry gibbed means having more stable health, avoidance is a random help, expertise instead remove the randomness.
Savage Defense: I've taken as an example the hardest hitting boss, obviosly when a boss hits for less, the shield is better, so the value of ap, crit, etc.. increase.
Talents: With dual spec, boss hitting harder and our armor nerf in ulduar, I think that we can easly avoid improved mangle and king of the jungle to take Ferocity. Demoralizing Roar decrease damage from melee by 10.4%. 5/5 Improved demo roar reduces damage by another 4.6%. It doesn't seems to much but is the way to put the HP soft cap at around 42k Hp instead of 44.5k HP...

Boevis · 05/12/09, 10:51 AM

Originally Posted by nightcrowler

Well 7k is something you are not likely to see for very different reasons. First of all basically every fight in Ulduar has movement or buff/debuff to damage. Second (at least for me) you're unlikely to have all best buffs and then there is the human factor, missing a key ability can lower the dps by a lot.

Ignis doesn't.

I will never be out of range, I will never not be behind my target, none of my buffers will be out of range to provide their buffs. Last go at it, I didn't have Windfury at all but everything else was still at the same uptime (or higher) as I had on the boss except of course sunder which took 7.5 seconds to 5 stack, and I still made the leaderboards with 6492 dps (6689 actual). Rawr is telling me I should do 7308 DPS, are you telling me that I lost 619 dps because of human error? My rotation is not perfect, but even going with "worst rotation" in Rawr, I still was several hundred short of where you guys are telling me I should be, so is the top druid on that fight (who outgears me by a bit and had a reliable mangle bot)

I'm not going to say you guys are at fault here, but something's wrong with current calculations whether it's some unknown mechanic, a bug causing our abilities to not be performing as listed, or a calculation error. I, and probably others, simply haven't been bringing it up because Rawr and FbN have been such a big help that it just seems rude when I don't actually have a clue what's causing the disparity, I just know it's not me.

psuman99 · 05/12/09, 11:14 AM

More discussion on armor pen from Ghostcrawler

Armor Penetration Cap
There are caps to armor pen, but the total amount of armor that can be reduced is not one of them. The numbers you need to really to get armor pen to the point where it feels broken cannot really be done with available gear.

[...] There is a cap per point. You can't ignore more than a certain percentage of the armor per point. This is what keeps armor pen from being better vs. targets with lower armor.

There is no cap on the amount of armor pen from which you can benefit. More points of armor pen always give you more damage (in the way more AP always gives you more damage, but more hit does not).

We are familiar with some of the theorycrafting for how great armor pen might be for fully Ulduar-geared melee. We're not worried about it yet because it's a big improvement over armor pen being a junk stat, which is where it was for many until recently.

We're not too worried about the double trinket proc in PvP. It's impossible to use on demand, you have to give your PvP trinket to do it, and it still isn't better vs. lightly armored targets, which is what we wanted to fix.

nightcrowler · 05/12/09, 11:27 AM

@psuman99:
that's cool. Basically our Arpen rating is fine then and actually Blizzard is not going to nerf double-trinket / negative armor effects.

@Boevis: I don't know if you are following FBN addon or not and I don't know if you have BiS pre-ulduar gear. With BiS pre-ulduar gear FBN predict 7k dps. 6689 actual dps (this is what we predict) is only 300 dps lower the predicted one and it's easly do with windfury, 100% 10% ap buff uptime (for istance if you have an enh shaman and not a MM hunter you will have something like an 80-90% uptime), some seconds for sunders and other debuffs, some downtime of DKs buff or shouts and so on (like ignis that toss you into the air or into his furnace). Also remember that there is a 20% dispersion on an average fight so I don't see that great discrepacy. As for RAWR, the assumption made by that model are far from real (it still remain a great tool indead).

The only real way to find out if there is any error should be making a 25 men raid with all possible buff (actually you only need 12 people or something like that), wait for all the buff to go up, go to a training dummy and do at least 100, 5 minute fight, average out and see the results.

• moz · 05/12/09, 11:57 AM

Originally Posted by nightcrowler

BiS Ulduar (without hard modes) gear

Slot	Item
Head	Garona's Guise
Neck	Titanstone Pendant
Shoulders	Conqueror's Nightsong Shoulderpads
Chest	Conqueror's Nightsong Rainment
Waist	Belt of the Twilight Assassin
Legs	Conqueror's Nightsong Legguards
Feet	Runed Ironhide Boots
Wrist	Thrusting Bands
Hands	Conqueror's Nightsong Handgrips
Finger1	The Leviathan's Coil
Finger2	Platinum Band of the Aesir
Trinket1	Heart of Iron
Trinket2	Darkmoon Card: Greatness (Agility)
Back	Cloak of the Shadowed Sun
MainHand	Twisted Visage + mongoose
Idol	Idol of the Ravenous Beast
Sockets	Solid Sky Saphire

Professions: JC+Enchanting (3x41 stam gems and 2x24 stam enchants).

Discussion
About capping value: hit caps at 8%, spell hit caps at 17% (used for demo roar and FFF). Expertise soft cap (dodge) is at 26 expertise, hard cap (parry) is at 56 expertise (-14% chance to be parried). Armor soft cap is at 40k.
HP soft cap: I wanted to put an indication of a good hp mark. My choice was to pick the hardest hitting boss in Ulduar and consider how much hp we need (without inspiration and parry gibbs) to survive 5 seconds without heals.
Parry gibbs: It's the main reason of wiping for a tank, they are the new crushing blow, that's why EXPERTISE is really really important, removing the chance to be parry gibbed means having more stable health, avoidance is a random help, expertise instead remove the randomness.
Savage Defense: I've taken as an example the hardest hitting boss, obviosly when a boss hits for less, the shield is better, so the value of ap, crit, etc.. increase.
Talents: With dual spec, boss hitting harder and our armor nerf in ulduar, I think that we can easly avoid improved mangle and king of the jungle to take Ferocity. Demoralizing Roar decrease damage from melee by 10.4%. 5/5 Improved demo roar reduces damage by another 4.6%. It doesn't seems to much but is the way to put the HP soft cap at around 42k Hp instead of 44.5k HP...

Pretty sure you'd want Idol of the Corruptor there instead of the Shred idol for bear.

Inviktus · 05/12/09, 12:02 PM

I've been reading this forum for a long while now and the information on it has been incredibly valuable. Like Boevis I have never wanted to post information that may not agree with this thread because of how helpful it has been to me and all of the hard work that has been put in by others far more in tune with the mathematical mechanics this class than myself, but I'm not seeing the same top end numbers that even pre BiS Ulduar gear should be seeing.

Granted, I know my gear is not the best in the world but I would consider it pretty decent. I also am aware that I am losing 6% on FB due to taking SI and the extra point in Swipe (I used to run with all 5 points). I'm fully gemmed for ArPen and above the Shred Idol threshhold. I am also accidentally at the Exp and Hit caps (unless I got an upgrade as of writting this). Yet, I still can not obtain close to the ~7.1k numbers that "should" be possible. I've tried it as both fully gemmed for AGI and ArPen. I've even tried it with 2pT7/2pT8 and 4pT8 (I feel like I get better results from the mix of T7/8 although they are very close). Finally, I've tried running with ever scenario of SR CP cycles and FB cycles. My best results on non-damage buffed fights (ie Ignis burn down strat, Gen V, and Mimiron P1) where there is little movement range in the realm of 6.3-6.7k. For comparisions sake I do ~4.2-4.4k on a target dummy over an extended period of time (I usually run 15-30m tests and stop 10s before Berserk refreshes).

Not being a numbers genius myself I would say that my gut feeling is the high end number caps aren't possible strictly due to RNG. My higher numbers always come from when my OoC and 2pT8 CC proc at just the right moment to allow me to weave in an extra FB without energy or CP waste. Also, even with our massive amount of crit there are deffinate crit strings that occur that can either leave you short or over on CPs at the wrong time. My best numbers have always come when the stars line up so to speak.

As additional points of reference I've used both FBN and FM. In the end I've found that my best DPS comes when I don't follow either strictly but rather use them as a point of reference based on what is actually going on with the mob in general. We also have a Bear MT for mangle bot and a MS warrior as well. In addition to that I'm the primary raid leader so I'm sure that lowers my DPS by a margin as well.

Lastly, this post is simply meant to give feed back and to share my personal experience and feelings about the current state of Kitty DPS. It is not meant to be the final word on anything at all but hopefully it sheds some light for those that are number crunchers to ponder over.

Allev · 05/12/09, 12:17 PM

The damage point here isn't just that our average numbers aren't equaling the theorycraft's average numbers-- it's that the max value of all the players who got parsed by WMO isn't even up to the average of the simulator. SimulationCraft outputs a graph of results distributions after a parse, and it shows +-600-800 DPS. While slow time-to-debuff or low uptimes on buffs/debuffs might be the case, it's really hard to believe that only 90% uptimes on certain buffs are causing consistent 10+% damage loss.

I'm guessing there's an "upstream" problem-- I'm pretty sure FBN, SimulationCraft, Toskk's, Rawr, and Mijae have all independently come up with very similar final DPS numbers-- in that, they're all overestimating DPS. There is a lot to test, although the first would be to check to see if we're hitting as hard as we're supposed to be and as often as we're supposed to be. It could very well be that our tooltips are wrong, or the wowhead/wowwiki/game file spell info is wrong.

Tonight's Ulduar clear for my guild should provide some decent numbers for both Ignis and Vezax (melee is tank/spank on Vezax, and I don't draw interrupt duty). My recent Ignis/Vezax parses aren't any good though (for Ignis, I was potted, for Vezax I was shifting bear to demo roar, or battle rez).

ithecho84 · 05/12/09, 1:07 PM

Originally Posted by Iwiperaids

Hello everyone, I am new here. I am reading this amazing thread for some time now and using the FBN addon and it's amazing.

However one thing I cannot find is the simulation tool, where do I download that? On FBN's google site there is a simulator I believe but I cannot use it as it's extension is "ccp" or something.

If someone could direct me on how and where to get the simulator I would appreciate.

Thanks and keep up the good work.

I believe it's written in C++ so something that will run that will do the job for you.

nightcrowler · 05/12/09, 2:01 PM

Actually Simcraft, FBN, Toskk's and Rawr are all great (I don't want to celebrate myself). But we should distinguish.

Rawr and Toskk's are theorycraft - theorical tools, they have statistical fomulas inside and so they made assumption and can't be 100% matching real value in the same way physical formulas can't match exactly complex, everyday reality without adjustment.
FBN and Simcraft are simulation tools, they actually replicate world of warcraft mechanics and simulate a fight, in this way the "can" be 100% accurate given the right conditions.

So we are down 2 possible things:
- are we sure that we have the "right" conditions? I'd like to see parses from "tank & spank" fights, with buffs uptime and so on and the druid armory, so I can run "specific" simulation to see the output difference (actually FBN basic condition never exist in a real fight, the equip is slightly different and so on).
- game mechanics and/or cycle could be wrong. Actually the second case can't really exist because FBN addon is simply a copy & paste of FBN simulator, so if you follow it the results should be around the same.

As for game mechanic I've tested with an high degree of accuracy the a lot of the FBN mechanic, in particular I've tested:
- 2-roll system
- crit suppression
- glancing blows
- hit cap
- expertise cap
- OOC proc rate
- Trinkets proc rate

I've not tested single abilities dps, that's true.

Actually FBN and Simcraft have about 200 dps difference (7k for FBN and 7.2k for SimCraft in best pre-ulduar gear), but reading in Simcraft post there should be problem with glancing and crit suppression (I suppose), otherwise the output is very similar.

The only mechanic I had problem testing are weapon procs (does haste affect them? I've supposed so but the difference is less than 20-30 dps.. so it doesn't justify the expirienced difference).

I'd like the previous posts, keep finding out problem and help me solving them.

Talking about myself, I know that many times in a fight I'm called to add/c-res, I have lag, I push the wrong button, the shaman totems wear off, the warrior doesn't shout, the dks doesn't refresh their buff, etc.. so I was never worried about "not doing simulated dps", a try I can do is to run a "target dummy / no buff sim". I'll tell you the results.

The simulation is for 1000 fight of 15 minutes on a boss target dummy. The feral druid is in BiS pre-ulduar gear and only self-buffed.

Aggressive Stats after all the buffs (full stacked raid + Bloodlust used 1 time in a 5 min fight)
_______________________

Cat (crit% vs. a boss (-4.8% added)): 40.2917%
Cat (AP): 7606
Expertise: 23
Hit: 7.65478%
Armor Penetration: 34.1013%
Haste (from equipment): 12.7264%

Boss base Armor (before penetration and debuff): 10643
Boss modified Armor (after penetration and debuff): 7230.04
Boss Damage Reduction (after debuff and penetration): 32.1871%
Weapon Speed Reduction % (Total without Bloodlust): 12%
Missed % (Total): 0.345225%
Dodged %: 0.553826%
Not normalized attacks parried by the boss: 0%
_____________________________________

DPS: 4235 +/- 2.5, (Min-Max variance: 13.2218%, Lower dps sample: 3928.86, Upper dps sample: 4488.79)

**************************************
CAT

White cat: 1061.91, Yellow cat: 3173.01
White: 25.0751%
Mangle: 6.35883%, Avg dmg: 3456.88, #/cycle: 2.23643),
Shred: 25.2978%, Avg dmg: 5886.79, #/cycle: 5.22479),
RIP: 25.7041%, Avg dmg: 33791.8, #/cycle: 0.924814)
Rake: 13.0992%, Avg dmg: 6782.12, #/cycle: 2.34824),
Ferocious Bite: 4.46499%, Avg dmg: 9946.8, #/cycle: 0.545759, Average Energy Usage: 43)
Savage Roar uptime: 99%

Bleed debuff uptime: 92%, Mangle debuff uptime: 93%, Mangle-RIP debuff downtime: 5.2%
Average Cycle Time: 28.7

The target dummy test seems to mach data, so it's probably a "buff" problem, or we don't have all the buff or I've wrongly taken into account some buffs. If someone wants to review buffs stacking into the code will be welcome.

Inviktus · 05/12/09, 2:21 PM

A no buff Target Dummy sim would be really helpful to me. And again I can't stress enough that I'm no where near perfect nor do I believe there is 100% uptime on all raid buffs. I just think the gap is pretty large (~500dps) which just could be due to me being a comeplete tool and/or noob.

syntax53 · 05/12/09, 2:24 PM

Originally Posted by nightcrowler

I'd like to see parses from "tank & spank" fights, with buffs uptime and so on and the druid armory, so I can run "specific" simulation to see the output difference (actually FBN basic condition never exist in a real fight, the equip is slightly different and so on).

Here is my Ignis kill from last week where I hit 6k dps: Wow Web Stats

I didn't get slag potted and my rotation went pretty well. I only use BadKitty for timers.

Here is the gear list (my armory currently has mostly tanking gear on atm)--

Valorous Dreamwalker Headguard (Relentless, +16 ArP) +Torment
Pendant of the Dragonsworn (+8 hit, +12 stam) ...wtb 25 man version, terrible luck
Valorous Dreamwalker Shoulderpads (+16 Agil) +Great Axe
Chestguard of the Recluse (+16 ArP) +Super Stats
Stalk-Skin Belt (+16 hit)
Leggings of the Honored (2x +16 Agil) +Nerubian
Runed Ironhide Boots (+16 Agil) +Icewalker
Wristwraps of the Cutthroat (+114 AP) +Fur AP
Valorous Dreamwalker Handgrips (+16 Agil) +Major Agility
Cindershard Ring
Power Enhancing Loop
Mirror of Truth
Grim Toll
Hammerhead Sharkskin Cloak +Major Agility
Journey's End +Berserking
Idol of Worship

371 ArP before Grim Toll. The gear with +agility gems are also used for tanking.

Tonight I'm going to regem my legs with 2x ArP since I got new tanking legs and use the shred idol.

Originally Posted by nightcrowler

a try I can do is to run a "target dummy / no buff sim". I'll tell you the results.

That would be awesome.

ithecho84 · 05/12/09, 2:37 PM

Originally Posted by nightcrowler

Aggressive Stats after all the buffs (full stacked raid + Bloodlust used 1 time in a 5 min fight)
_______________________

Cat (crit% vs. a boss (-4.8% added)): 40.2917%
Cat (AP): 7606
Expertise: 23
Hit: 7.65478%
Armor Penetration: 34.1013%
Haste (from equipment): 12.7264%

Boss base Armor (before penetration and debuff): 10643
Boss modified Armor (after penetration and debuff): 7230.04
Boss Damage Reduction (after debuff and penetration): 32.1871%
Weapon Speed Reduction % (Total without Bloodlust): 12%
Missed % (Total): 0.345225%
Dodged %: 0.553826%
Not normalized attacks parried by the boss: 0%
_____________________________________

DPS: 4235 +/- 2.5, (Min-Max variance: 13.2218%, Lower dps sample: 3928.86, Upper dps sample: 4488.79)

**************************************
CAT

White cat: 1061.91, Yellow cat: 3173.01
White: 25.0751%
Mangle: 6.35883%, Avg dmg: 3456.88, #/cycle: 2.23643),
Shred: 25.2978%, Avg dmg: 5886.79, #/cycle: 5.22479),
RIP: 25.7041%, Avg dmg: 33791.8, #/cycle: 0.924814)
Rake: 13.0992%, Avg dmg: 6782.12, #/cycle: 2.34824),
Ferocious Bite: 4.46499%, Avg dmg: 9946.8, #/cycle: 0.545759, Average Energy Usage: 43)
Savage Roar uptime: 99%

Bleed debuff uptime: 92%, Mangle debuff uptime: 93%, Mangle-RIP debuff downtime: 5.2%
Average Cycle Time: 28.7

The target dummy test seems to mach data, so it's probably a "buff" problem, or we don't have all the buff or I've wrongly taken into account some buffs. If someone wants to review buffs stacking into the code will be welcome.

I come close to the target dummy simulation by about 200 dps, though I am in Ulduar gear, but I'm thinking there's something going on with the buffs as well.

charriu · 05/12/09, 2:45 PM

Originally Posted by ithecho84

I believe it's written in C++ so something that will run that will do the job for you.

To run the simulator, you need to compile it first. If you don't know how to do that, search on the net (there's enough info on that stuff, really) or use some more userfriendly tool like Rawr or Toskk's Calculator.

Inviktus · 05/12/09, 3:19 PM

I'm at or slightly above the Target Dummy sim using FBN and FM. Although I am using 2pT8 and some other Ulduar pieces. Maybe just human error on my part.

Allev · 05/12/09, 5:15 PM

Originally Posted by syntax53

Here is my Ignis kill from last week where I hit 6k dps: Wow Web Stats

Looking at this:

- The gear looks approximately equal to
- It's hard to tell exactly how long you were in combat-- it looks like you pulled the boss, while you ran in you applied FF, starting your personal combat, and then another few seconds before you started attacking. Anyways, it looks like 10-15 seconds were clipped off the full 5 minute fight-- does that sound right? (I'm going to use 285 seconds as the length of the fight for estimations).
- You had 30 seconds of downtime on Battle Shout, so we have to assume your damage was a little low from that.
- You got pretty lucky with crits on shred/rip, so if this is a personal high performance, it may have been RNG-related.
- So, first indicator of incorrect sim DPS: autoattacks. Average white damage in the first post is 2200+, while you're grabbing 1861 (530,451/285). The number of attacks also seems low-- .783 attacks per swing. That seems pretty high if you've got a haste buff going on-- on a 300 second fight with WF/Icy Talons/Bloodlust or Heroism, you get .80 or so just from those alone. Can you verify whether you were in melee range and attacking the entire time? Something definitely isn't right here-- you missed the equivalent of 30 swings.

syntax53 · 05/12/09, 6:06 PM

Originally Posted by Allev

Can you verify whether you were in melee range and attacking the entire time? Something definitely isn't right here-- you missed the equivalent of 30 swings.

We do kill the first two golems to relieve a bit of pressure on the healing towards the end. It's possible through the movement at the beginning I missed some swings. 30 sounds like a lot though. And yes, I usually start out throwing FF on the boss while the tank gets a few hits in before I start my rotation.

RareBeast · 05/12/09, 8:07 PM

Another possible reason for the discrepancy between the sim and real life is latency. This is an issue that important to me as I regularly play with 300+ms and sometimes up to 500ms due to playing from Australia. Even in fights like Patch, using FBN I can't seem to approach the numbers I should be getting.

I feel (hope) latency is a big factor as I am fairly confident in my own abilities. Certainly, latency seems to have a much greater effect now than it did in our BC powershifting macro days.

Deathwing · 05/12/09, 11:40 PM

I got ~4100 dps on a few 15m boss dummy trials. Some of my gear(should be on armory) is not BiS pre-ulduar and some of it is from ulduar. Considering I don't have feral aggression, and as you keep adding more and more variables, not to mention more human interaction that can affect those variables, your variance is going to increase. To average out that variance, you need A LOT of time. One boss fight doesn't come anywhere near close enough. I think the sim is working pretty well.

Night: on your bear sim, why heart of iron and sta gemming when you're way over hp soft cap? Why not agi gemming and furnace stone?

nightcrowler · 05/13/09, 1:58 AM

@Allev: 2200 is the average white DPS not damage x hit.

Syntax was far from being fully buffed: Wow Web Stats

- Battle shout was up 4:20 / 5:00 minutes.
- You missed hearth of the crusader or similar effects (3% more crit)
- mangle was up only 4:04/ 5:00
- you missed 10 sec of sunders and 5 sec of FFF
- you missed blood poison/blood frenzy effect (4% more damage)
- you missed windfury? I can't find it on report.
- you missed ret pally haste aura.

Also Syntax was applying mangle so no mangle bot for him and a mangle bot itself value around 300 dps.

@Deathwing: our healers loves stamina. Actually I don't find a single reason for avoidance. Healers don't go oom and you can live more time without heals with stamina. The soft cap is: average TTL without heals for a 1.7 swing time, 70k damage boss. Actually Ignis when he is going to die, or final phase of thorim can hit harder.

a civilian · 05/13/09, 2:28 AM

He had a combat rogue and a mutilate rogue, which should cover blood poisoning and heart of the crusader respectively. (These are applied as side effects of the rogues' poisons, rather than separate debuffs.) The log doesn't show when players gain or lose totem effects, but the enhancement shaman was dropping windfury (link).

nightcrowler · 05/13/09, 3:08 AM

@A civilian, ok sorry for mi noobness I'm not too much skilled at reading wws reports. The numbers of white attacks seems really low considering WF+Ret haste (20% and 3%).

Important code fragment for white damage:

	double base_paw=54.857+wep_dmg; /*Base Paw Damage*/

	double weapon=base_paw; /*Weapon*/


	int speed=(int)(100.0/((1+wf*wf_haste*0.01)*(1+bloodlust_up*bloodlust*bloodlust_haste*0.01)*(1+moonkin*moonkin_haste*0.01)*(1+wpn_enchant.speed*wpn_enchant.up)*(1+haste+trinket1.haste*trinket1.up+trinket2.haste*trinket2.up+wpn_enchant.haste*wpn_enchant.up))); //autoattack speed


...

	agi=(int)((agi+SOE*178+MOW*52)*(1+BOK*0.1+imow_m)*(1+sof_m)); // Blessing of King + Strengh of Earth + Gift of the wild + Survival of the fitness.
	str=(int)((str+SOE*178+MOW*52)*(1+BOK*0.1+imow_m)*(1+sof_m)); // Blessing of King + Strengh of Earth + Gift of the wild + Survival of the fitness.

----
	double base_crit = 0.054+0.05+0.06; // Base crit chance + Leader of the Pack + Sharpened Claws.

....

	int apbase=(120+140+(attack_power+feral_ap*1.2+688*BS)+(agi+2*str))*(1+hotw_m1)*(1+TSA*0.1); /*Attack Power*/

.....

double armor_pen=1.25*(armor_pen_rating/(100*15.3952999));

	double arpen_cap=(((1-FF*0.05)*(1-SA*0.2))*barmor+15232.5)/3;
	if ((((1-FF*0.05)*(1-SA*0.2))*barmor) < arpen_cap){
		arpen_cap=(((1-FF*0.05)*(1-SA*0.2))*barmor);
	}
	double armor=barmor*((1-FF*0.05)*(1-SA*0.2))-(armor_pen+trinket1.armor_pen*trinket1.up+trinket2.armor_pen*trinket2.up+wpn_enchant.armor_pen*wpn_enchant.up)*arpen_cap;
	
	double red=1-(15232.5 / (armor + 15232.5)); /*Boss damage reduction based on boss armor*/


....


	 bmiss=bmiss-Draneo*0.01;
	bcrit-=bcrit_red;
.....

	double wmiss=miss+dodge+parry;
	double wnhit=1-bcrit-wmiss-wglance;
	if(wnhit<=0){
		wnhit=0;
	}

.....


	attack paw, mangle, shred, rake, rip, fb;
	/*paw*/
	paw.energy=0;
	paw.adamage=(1-red)*(ap/14)*(1+naturalist_m);
	paw.bdamage=(1-red)*(weapon)*(1+naturalist_m);
	paw.acritdamage=paw.adamage*2*(1+pistinct_m)*(1+META*0.03);
	paw.bcritdamage=paw.bdamage*2*(1+pistinct_m)*(1+META*0.03);
	paw.miss=wmiss;
	paw.hit=wnhit;
	paw.crit=wcrit;
	paw.glance=wglance;

......

if(((swinger%speed)==0)&&(catbear==1)&&(swinger!=0)){
				att=randnum();
				////cout << "Attack roll: " << att << "\n";
				
				if(att>(1-paw.hit)){
					is_hitted=1;
					if(ooc_rate>randnum()){
						ooc_proc=omen;
						
					}
					dmg=(1+sr*src)*(1+RA*0.03)*(1+BF*0.04)*(paw.adamage+paw.bdamage);
					damage+=dmg;
					wdamage+=dmg;
					//////cout << "\n" <<"Energy: " << en << ", Time: " << (1.0*timer)/100.0 << ", CP: " << cp <<"\n";
					//cout << "Paw (normal): " << dmg <<"\n";
					is_hitted=1;

				}
				else if(att>(paw.glance+paw.miss)){
					is_hitted=1;
					is_critted=1;
					if(ooc_rate>randnum()){
						ooc_proc=omen;
						
					}
					dmg=(1+sr*src)*(1+RA*0.03)*(1+BF*0.04)*((paw.acritdamage)+paw.bcritdamage);
					damage+=dmg;
					wdamage+=dmg;
					//////cout << "\n" <<"Energy: " << en << ", Time: " << (1.0*timer)/100.0 << ", CP: " << cp <<"\n";
					//cout << "Paw (crit): " << dmg <<"\n";
					is_hitted=1;
					is_critted=1;
				}
				
				else if(att>(paw.miss)){
					is_hitted=1;
					dmg=(1+sr*src)*glance_red*(1+RA*0.03)*(1+BF*0.04)*(paw.adamage+paw.bdamage);
					damage+=dmg;
					wdamage+=dmg;
					//////cout << "\n" <<"Energy: " << en << ", Time: " << (1.0*timer)/100.0 << ", CP: " << cp <<"\n";
					//cout << "Paw (glancing): " << dmg <<"\n";
					if(ooc_rate>randnum()){
						ooc_proc=omen;
					}
				}	
				else{
					//////cout << "\n" <<"Energy: " << en << ", Time: " << (1.0*timer)/100.0 << ", CP: " << cp <<"\n";
					//cout << "Paw (miss)" <<"\n";
				}
				
			}