Jump to content


Photo

[Fury/Arms] Rawr.DPSWarr Thread


  • This topic is locked This topic is locked
125 replies to this topic

#41 Jothay

Jothay

    Suit Up

  •  Patrons
  • 105 posts

Posted 01 October 2009 - 02:01 AM

In Rawr.DPSWarr the only thing that Haste does is increase number of White Swings, which increases Total Rage, which allows for more Heroic Strikes, which reduces rage since they replace White Swings.

Judging the Items you named, the "problem" most likely isn't Haste at all, it's that you are overly dependent on the Hit being provided by these items. What is your Hit at with Crystal Heart?
Come see the Rawr Project for all World of Warcraft Classes and Specs!
- Astrylian: Project Coordinator, Bear & Cat Dev
- Jothay: Project Coordinator & Arms Warrior Dev

#42 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 01 October 2009 - 02:03 AM

I'll look in to the value of haste. I just checked in some changes that will be in the next version that settles special effects more intelligently, and both of the effects you're referencing would be affected. I'm not at my dev box right now so I can't verify, but I am pretty sure those items are weighed lower. I'll update you when I know for sure.

As a side note, I'm not surprised that the eng gloves are better than anything else (although they are probably a little over-estimated right now). An average of 60-some haste rating is gonna be better than crusher, that's for sure.

#43 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 01 October 2009 - 04:40 PM

I did just check with the current unreleased code, and Hyperspeed Accelerators are still up around 3x the value of Crusher for Fury. This still makes sense to me and is on par with what Landsoul's sheet shows. In Landsoul's sheet, I have 2AP=0.75SEP and 1haste=0.74SEP. The accelerator proc averages out to 68haste for 50.32SEP total. Crusher is 0.75*44/2=16.5SEP (roughly 3x weaker).

The Shard of the Crystal Heart is considerably weaker with the changes. It's a hair worse than Mark of Supremacy, and is below Wrathstone by ~20dps but above Mirror of Truth by ~30dps.

You say that you're seeing lower DPS with the Shard than with Mirror. Is it possible that you're finding yourself rage-capped while using the Shard and unable to spend all incoming rage? That would be something I'm hoping to fix in the future (capping rage based on how quickly you can spend it vs the bursts that come in), but should be a problem that exists in Landsoul's sheet as well.

One thing that I noticed while investigating is that Landsoul's sheet values hit far less than Rawr does (we value it on par with Haste after the cap, while Landsoul does not). One thing that contributes is that Landsoul's sheet uses a straight Hits-per-Second in his proc uptime. We use Triggers-per-Second with a Hit Chance. Net result is that we have a slightly higher value on hit. There is a difference between an attack landing every 0.86secs and always hitting, and an attack landing every 0.84secs but sometimes it misses. I'm exploring this more to be sure, but I'm fairly confident that Rawr's is more correct, and that Hit Rating is valued far more than landsoul's sheet calculates.

#44 Montegomery

Montegomery

    Aloof Aggravator

  •  Patrons
  • 3,634 posts

Posted 01 October 2009 - 05:32 PM

I've also seen Rawr appear to value Haste higher than it should. In version 2.2.19 it claims that the would be an upgrade over my current trinkets, and . This seems outright incorrect, and the most obvious explanation I could see was that Rawr overvalued the haste proc on the Whetstone.

... you could very well have a concerto, but the closest the average listener gets to hearing it is the interpretation as put on by a group of small children with those little rainbow-coloured xylophones.


Monte's LoL Blog
Monte's LoL Stream

#45 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 02 October 2009 - 03:19 PM

You're right, it did (see the last two posts right above yours). That's the same as the issue I was just talking about that has been fixed and will be in the next release.

With DMC:Greatness and Grim Toll equipped, I'm seeing the whetstone giving 40dps less than GT and 130dps less than Greatness.

For technical details, we were only doing one pass of special effects rather than iterating over ones that affect others (ie, haste will affect berserking uptime). This made haste-based special effects seem more valuable than they were in practice, and is fixed for when 2.2.20 comes out.

I'm diving into the GT/MR fix that everyone's been clamoring for, and it should also be included in the 2.2.20 release.

Edit: Some other changes that have gone in recently:
- Mongoose and Executioner are now modeled on each independent weapon resulting in lower value
- Rage Gains from Damage Taken have begun to be added
- "Hide Bad Items" has been split up to Tank gear, Spell Gear, and PvP gear. Spell gear doesn't filter out on Int due to mail getting this as a free stat, but will filter out all items with mp5, spirit, or spell power
- Reaction Time/Latency are now used more intelligently (ie, reaction penalty is only applied to abilities that are, well, reactive)
- Avoidance Streak has been modified a bit to be less penalizing on your DPS


#46 Flab

Flab

    Glass Joe

  • Members
  • 16 posts

Posted 03 October 2009 - 11:15 PM

You say that you're seeing lower DPS with the Shard than with Mirror. Is it possible that you're finding yourself rage-capped while using the Shard and unable to spend all incoming rage? That would be something I'm hoping to fix in the future (capping rage based on how quickly you can spend it vs the bursts that come in), but should be a problem that exists in Landsoul's sheet as well.


Yes, I probably am taking in a little more rage than I can spend while using the Shard. This actually makes perfect sense.

In Rawr.DPSWarr the only thing that Haste does is increase number of White Swings, which increases Total Rage, which allows for more Heroic Strikes, which reduces rage since they replace White Swings.

Judging the Items you named, the "problem" most likely isn't Haste at all, it's that you are overly dependent on the Hit being provided by these items. What is your Hit at with Crystal Heart?


I was at 285 hit before equipping the Shard, so I don't think this is the case... I could be wrong, though, as more white swings landing is never a bad thing.

#47 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 04 October 2009 - 04:48 AM

After re-doing a lot of places where hit and haste were indeed overvalued (particularly in trinkets), the shard is now (rightfully) lower than Mirror. So while you were losing rage due to the haste, even if there weren't a rage cap it would still be worse. You were right from the start :)

Some other changes added recently:
- Grim Toll/Mjolnir Runestone calcs working properly (averaged over damage reduction instead of rating given)
- Rage gains are granted on dodged white attacks (I'm honestly now sure how I missed the memo that this was occuring!)
- Max Armor Penetration (ie, how much arp you have with your trinket procced) has been included in the Armor Penetration tooltip. This number will actually exceed 100% when you're over-capped, so you know just how much over you are. When your trinket is procced, however, the calculation does auto-cap itself at 100%

All of these changes will be in 2.2.20, which should be coming Sunday or early next week.

#48 Jothay

Jothay

    Suit Up

  •  Patrons
  • 105 posts

Posted 04 October 2009 - 05:07 AM

Ebs had a couple bugs in that GT/MR proc thing (see previous post) when it came to Arms, but we have that fixed now. We also discovered one of the major bugs with Hit, and fixed that (had to do with 2 roll crits). We also just fixed Sword Spec activates and some double-dipping activates for Deep Wounds.

I'm gonna spend the next 24 hours comparing Rawr to Landsoul's current sheet to see if there's any new insights that could help Rawr.
Come see the Rawr Project for all World of Warcraft Classes and Specs!
- Astrylian: Project Coordinator, Bear & Cat Dev
- Jothay: Project Coordinator & Arms Warrior Dev

#49 BarbaMsk

BarbaMsk

    Glass Joe

  • Members
  • 5 posts

Posted 05 October 2009 - 08:20 AM

Pre3.2 patch i was an arms/fury, and after Execute's nerf i dropped arms for tank, 'cause my arms-dps becomes 3/4 of my-fury. But today, i switch my gear from fury-file in Rawr in arms-talents, and Rawr saya: "even with str-gems, your arms-dps is slightly better then fury". How it can be? ^)

Link to armory

#50 BWarner

BWarner

    Piston Honda

  • Members
  • 128 posts

Posted 05 October 2009 - 10:23 AM

Pre3.2 patch i was an arms/fury, and after Execute's nerf i dropped arms for tank, 'cause my arms-dps becomes 3/4 of my-fury. But today, i switch my gear from fury-file in Rawr in arms-talents, and Rawr saya: "even with str-gems, your arms-dps is slightly better then fury". How it can be? ^)

Link to armory


It's helpful to take a line from Landsoul's sheet on this one: "Please don't use the spreadsheet to determine if fury is better than arms or not. It's just a metric to determine what gear to acquire to get the most efficient output. You should try both specs if you want to see which is better for what fight." Same goes for Rawr. It depends on the fight. It depends on the buffs available, and the buffs you need to provide. The DPS output from a calculation is merely a metric, a means of comparison within the closed system (that system being your specific character setup). Don't try to over-emphasize the values you get between two totally different specs here.

Some fights lend themselves to Arms. Some fights lends themselves to Fury. If there's enough of a performance difference for you to switch for a certain fight or set of fights, go for it! For everything else - play what you have the most fun with. Really. They're close enough where you can do that for most of the content, and not be hurting anybody by doing so.

Let me say that again:

For the overwhelming majority of the current
and recent raiding content, you can
play whichever spec you enjoy more,
and still be pulling your fair share in the raid.


Really.

#51 BarbaMsk

BarbaMsk

    Glass Joe

  • Members
  • 5 posts

Posted 05 October 2009 - 10:49 AM

Thanks for reply, but it does not really needed to write so much. I know about fight for Arms and fight for Fury, I've just wondering, is it _average_ damage on 232-245 gear for Arms close to Fury enough to get arms spec back and can i trust in Rawr dps valuation (round to 0.5k) or not. And answer "try what you like more" not exactly that i want, because i cant simply change raid spec and go into raid :)

#52 Jothay

Jothay

    Suit Up

  •  Patrons
  • 105 posts

Posted 05 October 2009 - 02:47 PM

Thanks for reply, but it does not really needed to write so much. I know about fight for Arms and fight for Fury, I've just wondering, is it _average_ damage on 232-245 gear for Arms close to Fury enough to get arms spec back and can i trust in Rawr dps valuation (round to 0.5k) or not. And answer "try what you like more" not exactly that i want, because i cant simply change raid spec and go into raid)


If you can't jump into a Raid like that for an evening you need a better guild :P

What you can do instead is test yourself against the Heroic Training Dummy before your raid
Come see the Rawr Project for all World of Warcraft Classes and Specs!
- Astrylian: Project Coordinator, Bear & Cat Dev
- Jothay: Project Coordinator & Arms Warrior Dev

#53 Coeus

Coeus

    Von Kaiser

  • Members
  • 33 posts

Posted 05 October 2009 - 03:19 PM

What you can do instead is test yourself against the Heroic Training Dummy before your raid


Unless you have your whole raid working on the dummy with you, that won't be an accurate comparison, either. You'll be missing vital buffs and debuffs that affect the specs quite differently. You can't look at dummy DPS and say that one spec is better for you with that as your only data point.

#54 Jothay

Jothay

    Suit Up

  •  Patrons
  • 105 posts

Posted 05 October 2009 - 03:20 PM

Unless you have your whole raid working on the dummy with you, that won't be an accurate comparison, either. You'll be missing vital buffs and debuffs that affect the specs quite differently. You can't look at dummy DPS and say that one spec is better for you with that as your only data point.


Which is why you should do it in a raid but if his guild won't let him, he doesn't have much of a choice.
Come see the Rawr Project for all World of Warcraft Classes and Specs!
- Astrylian: Project Coordinator, Bear & Cat Dev
- Jothay: Project Coordinator & Arms Warrior Dev

#55 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 05 October 2009 - 04:38 PM

No, Arms and Fury DPS are not meant to be compared to one another. They use different methods to determine DPS, and are only meant to compare against the same spec.

In a perfect world, we'd like to make them equal. They do share a lot of code with one another (all the helper functions are the same). But the way we model the rotation is going to be intrinsically different and imperfect with real world situations. We do our best to make it a close approximation, but in the end they are just approximations.

Just for some examples why the DPS numbers don't carry over:
- Fury is using a rotation where you never delay your Whirlwind or Bloodthirst (giving you a lower DPS than you would get if you sometimes delay BT for a short time to squeeze in that Slam that's about to fall off)
- Arms uses a priority list that slightly under-estimates slam usage because it assumes that you're able to always Execute on Sudden Death, even if it delays another ability (giving you probably a higher DPS than you would get in the real world).

Both models rely on being able to play the spec perfectly. Nobody's perfect. Even being imperfect at Arms, Rawr's representation of Arms gearing should be pretty close for you. But you may just be "closer to perfect" as Fury, so you perform better at Fury.

In the end, the point of Rawr is to determine how to best gear yourself for each spec. It is not meant to help you determine if you will do more DPS as Fury or as Arms. Just like using Landsoul's sheet for Fury and Rawr for Arms -- "Landsoul's sheet says I do 8k DPS as Fury, but Rawr says I do 8100 as Arms" -- that's a silly reason to change specs, and so is "Rawr's Arms says I do X and Rawr's Fury says I do Y". They just don't carry over.

#56 Vulgrym

Vulgrym

    Your Huckleberry

  • Members
  • 68 posts

Posted 05 October 2009 - 07:41 PM

It's helpful to take a line from Landsoul's sheet on this one: "Please don't use the spreadsheet to determine if fury is better than arms or not. It's just a metric to determine what gear to acquire to get the most efficient output. You should try both specs if you want to see which is better for what fight." Same goes for Rawr. It depends on the fight. It depends on the buffs available, and the buffs you need to provide. The DPS output from a calculation is merely a metric, a means of comparison within the closed system (that system being your specific character setup). Don't try to over-emphasize the values you get between two totally different specs here.


For what it's worth, I've had decent success modeling performance in Rawr that turns out to be pretty close in raid situations, provided I make it as apples-to-apples as possbile (mostly Arms). No, you can't apply this to every fight and yes, it means extensive fiddling with Fight Info, Ability Maintenance and Buffs, but I've found it a useful predictor.

Some fights lend themselves to Arms. Some fights lends themselves to Fury. If there's enough of a performance difference for you to switch for a certain fight or set of fights, go for it! For everything else - play what you have the most fun with. Really. They're close enough where you can do that for most of the content, and not be hurting anybody by doing so.


I generally agree with the "play what you want" approach (provided buff coverage is a non issue), but looking at aggregated parses on sites such as WMO, there's a pretty big discrepancy in terms of which spec does better, pretty much on all fights. This probably comes down things that are difficult to model, such as 'burn phase' flexibility via Deathwish, Bloodlust/Heroism scaling between the specs, consistent AoE damage due to adds etc. If there's a fight that currently lends itself to Arms better than Fury in this raid tier or even Ulduar, please point it out (Faction Champs?).

#57 BWarner

BWarner

    Piston Honda

  • Members
  • 128 posts

Posted 05 October 2009 - 09:30 PM

For what it's worth, I've had decent success modeling performance in Rawr that turns out to be pretty close in raid situations, provided I make it as apples-to-apples as possbile (mostly Arms). No, you can't apply this to every fight and yes, it means extensive fiddling with Fight Info, Ability Maintenance and Buffs, but I've found it a useful predictor.



I generally agree with the "play what you want" approach (provided buff coverage is a non issue), but looking at aggregated parses on sites such as WMO, there's a pretty big discrepancy in terms of which spec does better, pretty much on all fights. This probably comes down things that are difficult to model, such as 'burn phase' flexibility via Deathwish, Bloodlust/Heroism scaling between the specs, consistent AoE damage due to adds etc. If there's a fight that currently lends itself to Arms better than Fury in this raid tier or even Ulduar, please point it out (Faction Champs?).


I play Arms on Faction Beasts. I play Arms on Jaraxxus most of the time. I play Arms on Faction Champs. I play Arms most of the time on Twins. I play Fury on Anub, absolutely.

Could I get more number performance by making sure that I switch to the absolutely most optimal spec for any given fight? Sure. But, right now, I just don't enjoy Fury as much as Arms. So, I stay Fury for most of our raiding content. And, in my mind and that of my guild's, that's okay. The reason being that I am STILL CONTRIBUTING as a "sub-optimal spec, and not being "carried" by any means. I'm okay with playing a slightly suboptimal spec on the content we have on farm, because I know that I will enjoy those 3-12 minutes more as one spec than I would using the "best" spec.

There's some content that just begs to be accommodated with a certain spec, though. Arms for Vezax. Because I know it's a miserably boring (and sometimes painful resource-wise) fight to play as Fury. And, when the content being approached is progression content, it is only logical to play the spec that suits the fight much better (Fury on Anub, at the moment).

For most of the content, though? Don't get so hung up on it. Unless the only/primary enjoyment you derive is putting out the absolute "best" numbers possible in any given fight (not to insult those who enjoy shooting for "scoreboards"), Arms and Fury are close enough in the current gear/encounters that it's possible to perform well in either spec on any given fight. They're even close enough on the whole, or on a single-target damage fight that this is even true in the "macro" sense - the aggregate of the current content.

Regardless, more to the point at hand (as all of that is rather tangential from the main topic here) - ideally, accounting for all variables accurately (rotation efficiency, maintained abilities, adds, movement, mechanics, etc.), and using similar enough approaches to the DPS metric, the results would have enough parity to be able to compare the two metrics against each other. But, because of the reasons Ebs mentioned above, it's just not workable to do so.

#58 BarbaMsk

BarbaMsk

    Glass Joe

  • Members
  • 5 posts

Posted 06 October 2009 - 06:26 AM

Thanks to all of you, folks :)
This is not another holywar thread, its a Rawr thread, and The Question was "Can i trust in Rawr dps valuation, when Rawr shows equal dps in both Arms and Fury". I don't need an numbers, only logicat True/False

/And yes, i'm think i'm ask GM about one-two full raid-day for trying Arms again, but this is different story for different thread.

#59 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 06 October 2009 - 08:53 PM

In that case, if you want a short answer, it is:

No, a higher number in Arms does not mean it will do more damage than a lower number in Fury. Arms numbers can only be compared to other Arms numbers, and Fury numbers can only be compared to other Fury numbers. This is true for both Rawr and Landsoul's spreadsheet.

#60 ebs2002

ebs2002

    Von Kaiser

  • Members
  • 83 posts

Posted 07 October 2009 - 02:43 AM

Rawr 2.2.20 has been released, and most of the problems discussed in this thread have been repaired.

Check it out!




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users