The Arcane thread - Page 51

Renkyu · 01/31/09, 10:11 AM

Originally Posted by galzohar

The above was mentioned quite a few times really, and beyond what you've seen in rawr there isn't really much to add. With very certain setups the extra mana results in higher dps, but mostly the spec is inferior. More mana may give more dps but more dps also gives more dps (which is also why molten armor is often the right choice as well, but not always).

Well thanks I had a feeling it was inferior. I actually now do recall some discussion I believe on the 51/0/20 spec or if I am mistaken it might have been 53/0/18 however after reading nearly 30 pages obviously I forgot some stuff.

I guess even with an excellent tool such as Rawr the only true way I would assume to test a spec is to go out to the dummy's and test it yourself however that would be assuming your rotation and all timing is perfect which of course will never happen on a dummy let alone an actual raid boss in a long fight (except maybe patch).

galzohar · 01/31/09, 1:52 PM

The problem is that the dummy does not simulate a lot of important factors that rawr does simulate. Basically the only good way to get your theoretical dps is by theory. Practical tests will have all kinds of things shifting results due to factors that have nothing to do with your spec choice (say, with one spec u had crit streak and not with other, or with one spec your shaman sucked and with the other he didn't, or fight durations changed between raids etc...).

♦ nathanbp · 01/31/09, 3:23 PM

Originally Posted by Tifordin

Don't want to start a big thread on this until it's polished and actually has a UI / configuration etc, but...

TimeToExecute-0.1.zip

The above is something I've been having a dabble with. In a nutshell, it keeps a record of the hitpoints of a boss over the last 30 seconds, and does a simple linear least squares regression on the data to determine how long before a mob reaches execute range, and how long before a mob will die. Currently it has no UI or proper configuration - it has several configuration options in the .lua file (which are commented to document their purpose), and to inform the user of results, a message is printed in the default chat frame at most once every 5 seconds (configurable).

Feel free to use it, hope it's helpful. It should be fairly accurate for what we're using it for - although it by no means uses the strongest method of estimation. I chose the algorithm for performance sake (it takes <1ms on my machine), and because I didn't want to have to link in a matrix library. I don't have a timeframe for completion of a UI and so on, as it's definitely a spare time thing for me currently.

The screenshot was taken with a regular mob on my pally, with ret aura on, not attacking. He dodged occasional attacks, which is why the estimate is not linear.

There is a mod called TimeToDie (Broker_TimeToDie - Addons - Curse or TimeToDie - Addons - Curse) that does something similar. It appears to be slightly out of date, although the second link there works fine for me (the first link claims to be an updated version that I haven't tried yet). No clue how the algorithms differ from yours, but it has a nicer display than printing to chat every x seconds (which you may be able to "borrow" easily?).

Edit: Removed pic to save space.

galzohar · 01/31/09, 3:31 PM

With both my bad gear and manly's gear current rawr version shows fireball as the superior spec... Is current rawr version missing something or were the nerfs and fixes actually bad enough to bring fireball on top? Or am I possibly doing something wrong on rawr?

The recent discussions about rawr are a total mess, so it's really hard to understand what's the current state of rawr with the recent changes.

Tifordin · 01/31/09, 4:21 PM

Originally Posted by nathanbp

There is a mod called TimeToDie (Broker_TimeToDie - Addons - Curse or TimeToDie - Addons - Curse) that does something similar. It appears to be slightly out of date, although the second link there works fine for me (the first link claims to be an updated version that I haven't tried yet). No clue how the algorithms differ from yours, but it has a nicer display than printing to chat every x seconds (which you may be able to "borrow" easily?).

Edit: Removed pic to save space.

Ahh cheers - I searched for something similar to what I was making and came up dry (don't reinvent the wheel and all that). TimeToDie seems to simply use current health and initial health and assume linear kill rate. This means you will get an estimate which won't jump around at all (constantly ticking down), but won't adjust to a change in raid DPS well (which could happen for any number of reasons). I'll look into how they did their UI though... First time writing an addon

- software engineer by trade though.

anchelios · 01/31/09, 6:52 PM

Hey, just wondering about spell pent. Been archane since 3.0.8 and have been noticing infrequent amounts of damage resisted. Seemingly random, the amount of damage resisted varies from 200-700 damage and occurs on AB, AM and A Barr but not constantly. In effect, I might cast three A Blasts, of which nothing gets resisted, and then on AM a few of the ticks will be dampend by 200-300 points. While that's obviously not huge, it's still an unacceptable damage loss. Keep getting to the top dps but second or third with damage done.

So again, should I be picking up some spell pent or ? Haven't seen something about these fractional resists mentioned here yet, so I'm assuming it's related to my raid composition/gear/spec -which I'll admit is altered from the first post's ideal.

Also still curious if 18/53/0 has pulled ahead of arcane since the AM clipping was corrected.

Xentropy · 01/31/09, 7:06 PM

Targets 3 levels above you (i.e. all bosses) have approximately 4-6% partial resists that CAN NOT BE AVOIDED with any amount of penetration. This shows up now as each spell either being unresisted, 10% resisted, or 20% resisted, in proportions which will average out to the appropriate damage loss. It showed up in Burning Crusade as 25% and 50% resists, so at least it's a bit less spiky in Wrath. As far as I know, you shouldn't see 30% or larger resists unless the target has some arcane resistance above 0. In all practical terms, penetration is a PvP stat only.

Omnia · 01/31/09, 7:09 PM

Originally Posted by galzohar

With both my bad gear and manly's gear current rawr version shows fireball as the superior spec... Is current rawr version missing something or were the nerfs and fixes actually bad enough to bring fireball on top? Or am I possibly doing something wrong on rawr?
The recent discussions about rawr are a total mess, so it's really hard to understand what's the current state of rawr with the recent changes.

If you have suboptimal mana regen, a very long fight, and spellhit wasted in Arcane beyond 210, it's possible for Fireball to pull ahead. But here's a Rawr version comparison with ~optimal (spec-specific) pre-Ulduar gear, optimal raid buffs (incl. 1 innervate, 1 manatide) except FI/SR**, and cooldowns averaged over a 3'10" fight:

Rawr2.1.8 Arcane: 7064 dps
Rawr2.1.8 Fireball: 6353 dps
Rawr2.1.8 FFB: 6360 dps

Rawr2.1.9 Arcane: 6838 dps
Rawr2.1.9 Fireball: 6353 dps
Rawr2.1.9 FFB: 6360 dps

Disclaimer: Take these numbers with a grain of salt.

** I recommend not enabling Ferocious Insiration or Sanctified Retribution (new in 2.1.9), as they increase the 2.1.9 numbers above by varying amounts, instead of the expected 2%. (Boosts Arcane by 0%, FFB by 0.9%, and Fireball by 3%.) Kavan/Astrylian?

anchelios · 01/31/09, 7:09 PM

Originally Posted by Xentropy

Targets 3 levels above you (i.e. all bosses) have approximately 4-6% partial resists that CAN NOT BE AVOIDED with any amount of penetration. This shows up now as each spell either being unresisted, 10% resisted, or 20% resisted, in proportions which will average out to the appropriate damage loss. It showed up in Burning Crusade as 25% and 50% resists, so at least it's a bit less spiky in Wrath. As far as I know, you shouldn't see 30% or larger resists unless the target has some arcane resistance above 0. In all practical terms, penetration is a PvP stat only.

Ok, thanks. Completely forgot about that factor.

radikal · 01/31/09, 7:10 PM

How horrible is 20/51/0 assuming infinite mana and AB+ LB rotation? (Assuming bugged burnout obviously) Granted you end up with a LOT of wasted talent points...

manly · 01/31/09, 7:16 PM

Originally Posted by radikal

How horrible is 20/51/0 assuming infinite mana and AB+ LB rotation? (Assuming bugged burnout obviously) Granted you end up with a LOT of wasted talent points...

Bugged Burnout is the same as spell power. I see what you're trying to do but this simple fact will make it not work.

Gohei27 · 01/31/09, 8:01 PM

I was thinking the other day about Ember SD and how I wouldn't need to use two blue gems, so I would gain +45 dmg just in gems. Since arcane doesn't gain as much bonus damage as ffb builds is it still CSD no matter what? What if you were a jewel crafter?

vellias · 01/31/09, 8:05 PM

Quick question, is there any mod that tracks non personal debuffs that the user can specify? Sc2 is pretty nice for scorch and winters chill, but just wondering if there's anything similar that can track thunderclap and icy touch, since they work in conjunction with TTW.

Kavan · 01/31/09, 8:21 PM

Originally Posted by Omnia

** I recommend not enabling Ferocious Insiration or Sanctified Retribution (new in 2.1.9), as they increase the 2.1.9 numbers above by varying amounts, instead of the expected 2%. (Boosts Arcane by 0%, FFB by 0.9%, and Fireball by 3%.) Kavan/Astrylian?

Oops, I don't know how this slipped through. In the current live version this buff applies to fire spells only, I'll get a fix in.

galzohar · 01/31/09, 8:33 PM

Why are you using a 3 minute fight though? I don't see anyone killing sarth3d or kt in 3 min, and future fights are most definitely not going to be this short either, considering even naxx, when cleared at first, had fights ~5min. Fights slightly over 2 minutes and generally very short fights will obviously favor arcane, but before the nerfs/fixes I was getting arcane clearly ahead on very long fights as well. Keep in mind I assume a totally optimal raid with 1 mana tide every 5 minutes so it's not like I'm leaving out important mana regen stuff. Then again it doesn't take an extremely short fight to bring arcane to the top, 5 min or less seem to be around where arcane tops, with exact duration obviously depending on more specific details.

If you're looking to improve your DPS on short farm-mode fights might as well go further and look at improving your trash DPS as well, after all if you really don't care about making the hard fights easier then that's what you should be maximizing.