Remove Advertisements

[Ret] Rawr.exe

Ret, Holy, PVP, etc

Moderators: Fridmarr, Worldie, Aergis, Baelor

[Ret] Rawr.exe

Postby Joanadark » Thu Jul 02, 2009 1:35 am

So I was playing around with Rawr yesturday, and I noticed a few interesting things.
First off, no matter the gear set, our optimal fight length is apparently 4.4 minutes. cool to know.

Secondly, and the reason I'm making this thread, is that according to rawr, post-3.0 my full-BiS gear set list will enable me to do 8.3k DPS under optimal conditions, before even counting the buff to JCing-only gems.... HOLY HUH?

How reliable is rawr? Who uses it regularly? I tend to do my own math myself so I only just picked up the program but it's supposed to be fairly accurate for most classes. Most of my guild uses it at least.

8.3k DPS in ulduar gear as a support class? wow.
Arkham's Razor: a theory which states the simplest explaination tends to lead to Cthulu.
Joanadark
 
Posts: 3087
Joined: Fri May 11, 2007 7:09 pm

Re: [Ret] Rawr.exe

Postby Anorian » Thu Jul 02, 2009 1:39 am

From my experience so far 8k is really not possible atm. our regular ret pally maxes at about 5k on normal fights, im usually a bit below when going offspec.

About the optimal length thats not so strange, since gear doesnt affect our sustainability. 4.4 minutes means we just had enough time for 2 wings. I guess after that mana drops and you cant use your full rotation anymore.
Image
User avatar
Anorian
 
Posts: 361
Joined: Tue Jun 03, 2008 2:38 am

Re: [Ret] Rawr.exe

Postby loeni » Thu Jul 02, 2009 5:30 am

Its pretty good for comparing gearsets and finding the optimal set, but the raw numbers its spits out for dps are for ideal conditions like any spreadsheet, so its only reliable for comparing two gearsets to each other, not for figuring out how much damage you can expect to do. The 3.2 support is less than reliable at this point i think, since so much of the current ptr mechanics around SoV hadn't been nailed down when the last version of the ret module was released, although atm the fact that its showing a large dps increse in 3.2 doesn't seem to be too far off the mark of the current ptr at least.
Image
loeni
 
Posts: 38
Joined: Mon Jan 28, 2008 12:59 am

Re: [Ret] Rawr.exe

Postby Dorvan » Thu Jul 02, 2009 7:55 am

All the BiS lists I've seen posted for Rawr put are hovering shy of 7k DPS. If you could post a link to your character file I could maybe help you figure out where the settings are wrong. The most common error I see is people checking "Bloodlust" under the buffs tab....that checkbox actually enables perma-bloodlust (yes, it's stupid, and I don't know why they still have that in there). For bloodlust used as you'd actually see it in game, see options -> fight parameters tab.

Also keep in mind that as with virtually all DPS models, it assumes perfect FCFS execution and a completely stationary fight, which aren't exactly realistic assumptions for Ulduar.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Re: [Ret] Rawr.exe

Postby Dragonzbane » Thu Jul 02, 2009 11:00 am

Here's a pretty up to date list from EJ:
http://elitistjerks.com/f76/t57144-3_1_ ... ear_lists/

Under their fight parameters the BiS Alliance Paladin list looks like this:

Hard Mode + All Items:

Head: [Conqueror's Aegis Helm]
Amulet: [Frigid Strength of Hodir]
Shoulders: [Conqueror's Aegis Shoulderplates]
Cloak: [Drape of the Faceless General]
Chest: [Embrace of the Gladiator]
Wrist: Solar Bindings
Hands: [Conqueror's Aegis Gauntlets]
Waist: [Soul-Devouring Cinch]
Legs: [Conqueror's Aegis Legplates]
Feet: [Sabatons of Lifeless Night]
Ring1: [Seal of the Betrayed King]
Ring2: [Bladebearer's Signet]
Trinket1: [Darkmoon Card: Greatness]
Trinket2: [Comets Trail] 720 Haste for 10 sec/45 ICD, 15% chance to proc, 240 AP
Weapon: [Voldrethar, Dark Blade of Oblivion]
Libram: [Furious Gladiator's Libram of Fortitude]


They even provide downloads for each of the profiles on the list in the OP.
Image
Dragonzbane
 
Posts: 1673
Joined: Wed Mar 28, 2007 11:58 am
Location: In the Forums, stirring the pot.

Re: [Ret] Rawr.exe

Postby Joanadark » Thu Jul 02, 2009 2:10 pm

The most common error I see is people checking "Bloodlust" under the buffs tab....that checkbox actually enables perma-bloodlust (yes, it's stupid, and I don't know why they still have that in there). For bloodlust used as you'd actually see it in game, see options -> fight parameters tab.


Thanks for this, this turned out to be the error.

Correcting this dropped my projection down to 6510 DPS pre-3.2, and 7664 DPS post-3.2.

7.66k is still a pretty amazing number. And yes, I know all the assumptions that any spreadsheet style calculation has to make regarding optimal conditions, etc.

On a related note, why do everybody's ideal gearsets always include 4-peice t8.5 when running all off-set 239 peices grants such a higher return in AP? Granted, the 4-peice is powerful, its what I run with currently, but with all off-set BiS items you should be able to switch away from running 4-peice. At the level of raid buffed crit that gear set would be running, in my experience stacking more crit isn't really all that helpful, since stuff either crits, or it doesn't.
Arkham's Razor: a theory which states the simplest explaination tends to lead to Cthulu.
Joanadark
 
Posts: 3087
Joined: Fri May 11, 2007 7:09 pm

Re: [Ret] Rawr.exe

Postby fafhrd » Thu Jul 02, 2009 2:16 pm

Joanadark wrote:At the level of raid buffed crit that gear set would be running, in my experience stacking more crit isn't really all that helpful, since stuff either crits, or it doesn't.


*scratching head*
ImageImage
1/1 Lore pre-nerf.
User avatar
fafhrd
 
Posts: 5432
Joined: Fri Aug 31, 2007 2:31 pm

Re: [Ret] Rawr.exe

Postby Joanadark » Thu Jul 02, 2009 2:18 pm

Its the same reason why Holy Paladins stopped stacking crit in naxx, and started stacking int. More crit simply didn't continue doing much for them.

To elaborate on that, one of the major reasons why you can't just slave yourself to the spreadsheets is that they go by averages always working out. The reality of pretty much all encounters you'll face rigth now is that you don't actually use your abilities enough times each for the law of averages to make everything work out exactly the way it's supposed to. If I'm only Divine Storming something 25 times, it means that the value of crit is different than if I was Divine Storming something 1000 times, like in a spreadsheet.
Essentially what this means is that, in real practice, crit is extremely valuable when you have low amounts of it, but much less valuable once you are reaching 45%+ crit raid buffed or more. At that point, you have enough crit that you basically "have a good chance of critting", but adding more crit isn't really going to increase the number of crits you see significantly.
When that happens, you want to start heavily favoring AP (more than usual that is) over crit because its the onlything that is garenteed to affect ALL your attacks that you do make.
Arkham's Razor: a theory which states the simplest explaination tends to lead to Cthulu.
Joanadark
 
Posts: 3087
Joined: Fri May 11, 2007 7:09 pm

Re: [Ret] Rawr.exe

Postby Dorvan » Thu Jul 02, 2009 2:26 pm

Joanadark wrote:If I'm only Divine Storming something 25 times, it means that the value of crit is different than if I was Divine Storming something 1000 times, like in a spreadsheet.


This is incorrect. The variance as a percentage of your damage certainly changes, but the average damage contribution is no different based on the number of times you use the skill.

Essentially what this means is that, in real practice, crit is extremely valuable when you have low amounts of it, but much less valuable once you are reaching 45%+ crit raid buffed or more. At that point, you have enough crit that you basically "have a good chance of critting", but adding more crit isn't really going to increase the number of crits you see significantly.
When that happens, you want to start heavily favoring AP (more than usual that is) over crit because its the onlything that is garenteed to affect ALL your attacks that you do make.


This is also wrong. As long as you can keep up Vengeance (which requires a trivial amount of crit, you're probably fine naked with talents), the DPS contribution of going say 30-31% crit is exactly the same as going 40-41%. Now certainly as your crit goes up, the relative value of other stats goes up (and vice versa) but the modeling takes that into account when making gear evaluations, is not simplistically using static stat weights.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Re: [Ret] Rawr.exe

Postby Joanadark » Thu Jul 02, 2009 2:35 pm

But if I only hit something 25 times with Divine Storm, if I have 40% crit or if I have 41% crit, I'm still likely to have the same number of Divine Storms be critical hits.
It takes a greater quantity of crit the fewer hits with an ability you do to have the difference actually manifest its self on a DPS parse, beyond the normal variance of the RNG.

Technically you could run with no crit beyond baseline, stack 100% strength, and just keep pulling the boss until you get lucky. I know that's an exaggeration, but the fewer the ability uses you have in a fight, the more close to reasonable that example becomes.

(edit: By the way, the place I'm taking 25-DSs from is the 4.4 minute optimal fight length, in which you would, in fact, only be using DS 25 times. In a optimized DPS scenario, you're talking about a hard stand-and-burn encounter or phase of an encounter, which, by nature, is going to be short. Think 3-minute Hodir for example, or the 40-seconds of Yog Saron Brain DPS time you get each portal phase. Short means few ability uses, which means depriciating effect of adding greater and greater crit.)
Arkham's Razor: a theory which states the simplest explaination tends to lead to Cthulu.
Joanadark
 
Posts: 3087
Joined: Fri May 11, 2007 7:09 pm

Re: [Ret] Rawr.exe

Postby fafhrd » Thu Jul 02, 2009 2:42 pm

Joanadark wrote:Its the same reason why Holy Paladins stopped stacking crit in naxx, and started stacking int. More crit simply didn't continue doing much for them.

To elaborate on that, one of the major reasons why you can't just slave yourself to the spreadsheets is that they go by averages always working out. The reality of pretty much all encounters you'll face rigth now is that you don't actually use your abilities enough times each for the law of averages to make everything work out exactly the way it's supposed to. If I'm only Divine Storming something 25 times, it means that the value of crit is different than if I was Divine Storming something 1000 times, like in a spreadsheet.
Essentially what this means is that, in real practice, crit is extremely valuable when you have low amounts of it, but much less valuable once you are reaching 45%+ crit raid buffed or more. At that point, you have enough crit that you basically "have a good chance of critting", but adding more crit isn't really going to increase the number of crits you see significantly.
When that happens, you want to start heavily favoring AP (more than usual that is) over crit because its the onlything that is garenteed to affect ALL your attacks that you do make.


This makes no sense.

Take your "only 25 attacks" case.

Suppose you do it one time, crit 13 times and hit 12 times and do 13 x 2000 + 12 x 1000 = 38000 damage

Now you say "i'm critting so much, i'm going to increase my AP instead of my crit, since ap is *guaranteed* to increase my damage on every attack, and you go do your 25 attack boss again

But... crit is random. So guess what, you didn't get 13 crits again and do more damage. You got 12 crits and 13 hits for the following damage (note each hit and crit is bigger): 12 x 2050 + 13 x 1025 = 37925 damage. By saying "ap is guaranteed to increase my damage on all attacks" you're assuming you'll get the same distribution of crits, which you can't because it's random (and since you're specificially talking about small samples, the variance is huge - if you were instead talking about statistically large samples you could at least fall back on saying you'd on average get the same number of crits, but you can't in your 25 attack case.

The only case where you can say AP (or some other stat) is guaranteed to increase damage on all attacks if when you *are* talking about a large sample of attacks, and can then see that the amount of crit traded for AP is such that you'd still gain damage from the AP.
ImageImage
1/1 Lore pre-nerf.
User avatar
fafhrd
 
Posts: 5432
Joined: Fri Aug 31, 2007 2:31 pm

Re: [Ret] Rawr.exe

Postby Dorvan » Thu Jul 02, 2009 2:48 pm

Joanadark wrote:But if I only hit something 25 times with Divine Storm, if I have 40% crit or if I have 41% crit, I'm still likely to have the same number of Divine Storms be critical hits.


Actually, you've got a 40% chance of getting at least one more crit just from the DS (remember, it procs seals which can also crit). Factor in all our abilities and you easily will get extra crits with high probability. Yes, crit increases the variance of your DPS, but for any reasonable length fight things are going to average out decently.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Re: [Ret] Rawr.exe

Postby Joanadark » Thu Jul 02, 2009 2:52 pm

This makes no sense.

Take your "only 25 attacks" case.

Suppose you do it one time, crit 13 times and hit 12 times and do 13 x 2000 + 12 x 1000 = 38000 damage

Now you say "i'm critting so much, i'm going to increase my AP instead of my crit, since ap is *guaranteed* to increase my damage on every attack, and you go do your 25 attack boss again

But... crit is random. So guess what, you didn't get 13 crits again and do more damage. You got 12 crits and 13 hits for the following damage (note each hit and crit is bigger): 12 x 2050 + 13 x 1025 = 37925 damage. By saying "ap is guaranteed to increase my damage on all attacks" you're assuming you'll get the same distribution of crits, which you can't because it's random (and since you're specificially talking about small samples, the variance is huge - if you were instead talking about statistically large samples you could at least fall back on saying you'd on average get the same number of crits, but you can't in your 25 attack case.

The only case where you can say AP (or some other stat) is guaranteed to increase damage on all attacks if when you *are* talking about a large sample of attacks, and can then see that the amount of crit traded for AP is such that you'd still gain damage from the AP.

Thats my point though, the randomness of crit in a small sample size means that small gains in crit chance really don't matter much, they dissapear completely within the normal variance of the RNG. Gains in crit do not compensate for it's randomness, unless the amount of crit you stack is massive, which isn't really viable.
Arkham's Razor: a theory which states the simplest explaination tends to lead to Cthulu.
Joanadark
 
Posts: 3087
Joined: Fri May 11, 2007 7:09 pm

Re: [Ret] Rawr.exe

Postby fafhrd » Thu Jul 02, 2009 2:54 pm

Joanadark wrote:But if I only hit something 25 times with Divine Storm, if I have 40% crit or if I have 41% crit, I'm still likely to have the same number of Divine Storms be critical hits.


No, you aren't. Even with 40% crit again, you are not likely to have the same number of divine storms be critical hits on the second attempt. Nor the third, nor the 4th etc. With a sample as small as 25, your number of crits is vastly more likely to be different each time than the same.


Joanadark wrote:(edit: By the way, the place I'm taking 25-DSs from is the 4.4 minute optimal fight length, in which you would, in fact, only be using DS 25 times. In a optimized DPS scenario, you're talking about a hard stand-and-burn encounter or phase of an encounter, which, by nature, is going to be short. Think 3-minute Hodir for example, or the 40-seconds of Yog Saron Brain DPS time you get each portal phase. Short means few ability uses, which means depriciating effect of adding greater and greater crit.)


As noted above, this analysis doesn't make sense. However even if it did, your wow lifetime isn't 4 minutes. You might do 25 DS per pull of Yogg+0 hard, but you pull him 10 times a week for 3 weeks, and end up doing 750 divine storms, not 25. Also, the one time your gambling against crit paid off and crit for [LOTS] 85% of the time with only a 40% crit rate, the rogue in your group who tried the same thing only crit 25% of the time (with a 50% crit rate), so you wiped and your spectacular performance was wasted - that particular 4 minutes, it could have been the one that mattered, but it wasn't :(

It's better to bank on averages than say "i'm lucky enough already", trust the math models when it comes to how much of stat X to trade for stat Y.
Last edited by fafhrd on Thu Jul 02, 2009 2:59 pm, edited 1 time in total.
ImageImage
1/1 Lore pre-nerf.
User avatar
fafhrd
 
Posts: 5432
Joined: Fri Aug 31, 2007 2:31 pm

Re: [Ret] Rawr.exe

Postby Dorvan » Thu Jul 02, 2009 2:55 pm

Joanadark wrote:Thats my point though, the randomness of crit in a small sample size means that small gains in crit chance really don't matter much, they dissapear completely within the normal variance of the RNG. Gains in crit do not compensate for it's randomness, unless the amount of crit you stack is massive, which isn't really viable.


Setting aside that you actually *do* have a fairly large sample size between all your abilities over the course of the fight, you're also only considering the worst case. Larger variance means that sometimes you don't see any improvement, but other times you see significant improvement. Realistically, you're actually averaging over all the spell cast by you entire raid, which homogenizes the overall raid DPS numbers even more.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Next

Return to Off-specs & Other Classes

Who is online

Users browsing this forum: Bing [Bot] and 1 guest


Remove Advertisements

Who is online

In total there are 2 users online :: 1 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: Bing [Bot] and 1 guest