Enhance Shaman: The Collected Works of Theorycraft, Vol I - Page 5

• Nite_Moogle · 06/25/07, 12:55 AM

Originally Posted by slant

I'm not convinced that's actually true. Please correct me if I've missed it, but I haven't seen any math, sims, or tests that showed any reason to have an offhand slower than your mainhand. You just want them both as slow as possible, and while you'll miss out on the extra flurried hit if they aren't the same speed, 2.7 main / 2.6 off is perfectly viable.

I'd tend to agree with this assertion, especially given how well haste procs treat Shamans in terms of itemization efficiency.

Mios · 06/25/07, 1:01 AM

It makes sense to me that a mainhand that is slightly slower than your offhand would be detrimental to your dps. This is because your offhand will always be hitting slightly before your mainhand, which gives a good chance that your offhand will be eating significantly more windfury procs.

How does this effect two weapons with the same speed? If both of your weapons are always hitting at the same time, does the game have an order it chooses when looking for possible windfury procs, or is it just random whether it looks at your mainhand or offhand first with equal weapon speeds? I am saying this because I have two 2.6 speed weapons, and it does seem like I get more offhand procs than mainhand procs... could just be imagining it though I guess.

On a side note, check out this screeny:
http://img187.imageshack.us/img187/5...erhastedl9.gif
Those weapon speeds were originally 2.6. This is with almost every possible haste buff proc'ed: Dragonspine Trophy proc, haste Meta Gem proc, haste potion, bloodlust, flurry, and a mongoose proc. I suppose it could go faster with the crafted mace proc, another mongoose proc, abacus, and maybe some haste gear.

◊ Shalas · 06/25/07, 1:07 AM

Originally Posted by Mios

It makes sense to me that a mainhand that is slightly slower than your offhand would be detrimental to your dps. This is because your offhand will always be hitting slightly before your mainhand, which gives a good chance that your offhand will be eating significantly more windfury procs.

No, that isn't how time works. MH hits don't reset your OH swing timer. Your first few swings the OH will be right before the MH... and then you'll have some swings where they're far apart... and then you'll have some swings where the MH will be right before the OH, and then DOT DOT DOT

• Nite_Moogle · 06/25/07, 1:09 AM

It makes sense to me that a mainhand that is slightly slower than your offhand would be detrimental to your dps. This is because your offhand will always be hitting slightly before your mainhand, which gives a good chance that your offhand will be eating significantly more windfury procs.

No, if you start combat normally your main hand will hit first since your offhand is on a half-speed cooldown. Even if you are primed so both swing at the same time, your main hand will still get the first chance to proc WF. Even at that you're simply alternating swings as normal, so as soon as your off hand "misses" a proc your MH is back in the driver's seat. It isn't like you'll get stuck in a cycle where your offhand will always be swinging first; that actually can happen with same-speed weapons but it's uncommon.

Also, the H in between Jesus and Christ apparently stands for Haste.

slant · 06/25/07, 1:11 AM

Makes sense? Perhaps. Lots of theories made sense; for example, it made sense that you'd want to get 2.0 weapons in both hands to minimize the amount of time waiting for your weapon delay to proc windfury as soon as possible. But it didn't hold up in testing or sims, they clearly proved we should equip slower weapons, so that theory was discarded. We're trying to get away from "makes sense", and figure out how these mechanics actually work.

Negative · 06/25/07, 1:35 AM

Originally Posted by Shalas

There's a 14 DPS difference between them, even ignoring the Cleaver's superior stats and speed. It's possibly the biggest of an upgrade from one item in the history of your character.

I'm not sure why you're trying to decide between two level 70 weapons at level 35, though.

I figure if I decide now I won't be running around not knowing what I'm doing at 70.

Beowolf · 06/25/07, 3:22 AM

Is there any way possible that you could add a toggle feature that can include or exclude certain instances? For instance, my guild obviously doesn't run MC/BWL/AQ40/Naxx anymore, nor are we in Hyjal or the Black Temple. However, when I do a blanket optimization, it includes items from those zones.

Basically, I'd like to build an 'optimized' set for the instances we are farming or bosses we will farm soon, without having to manually search what every boss drops and input numerous different items into your optimizer to see which is 'best'.

Ilmatar · 06/25/07, 3:35 AM

Originally Posted by Negative

I figure if I decide now I won't be running around not knowing what I'm doing at 70.

Which weapon is better is obvious, and the answer to that question is not the one you need to be looking for. You need to figure out what about the class/spec confuses you so that you couldn't determine, without posting on a theorycraft board, which weapon is better.

W.r.t. haste: Correct, by stacking ap you will never hit a statistical valley like you can with haste, however, it's not a "which is better haste or ap" argument. If haste provides a scalable DPS increase (which the theory states, and testing seems to support) than adding a % benifit to your DPS is not something that you can do with a consistent amount of AP. So while a single upgrade may indicate that "oh this bit is better because my dps is higher" adding haste provides a % increase to your dps that will scale with the AP you gain from other gear. Haste potential (static or proc) on an item needs to be valued slightly differently because while you can very easily say, "this pair of boots is better because I have 10 more ap with them," it is not as easy to evaluate, "Well these bracers give me a static haste value, and while I lose a bit of dps now, my next gear upgrade will have a 0.66% modifier to it's addition because of the haste on my bracers." (for example)

Negative · 06/25/07, 3:45 AM

Originally Posted by Ilmatar

Which weapon is better is obvious, and the answer to that question is not the one you need to be looking for. You need to figure out what about the class/spec confuses you so that you couldn't determine, without posting on a theorycraft board, which weapon is better.

W.r.t. haste: Correct, by stacking ap you will never hit a statistical valley like you can with haste, however, it's not a "which is better haste or ap" argument. If haste provides a scalable DPS increase (which the theory states, and testing seems to support) than adding a % benifit to your DPS is not something that you can do with a consistent amount of AP. So while a single upgrade may indicate that "oh this bit is better because my dps is higher" adding haste provides a % increase to your dps that will scale with the AP you gain from other gear. Haste potential (static or proc) on an item needs to be valued slightly differently because while you can very easily say, "this pair of boots is better because I have 10 more ap with them," it is not as easy to evaluate, "Well these bracers give me a static haste value, and while I lose a bit of dps now, my next gear upgrade will have a 0.66% modifier to it's addition because of the haste on my bracers." (for example)

It wasn't so much of a "which is better" question as it was a matter of time investment. I figure a Gladiator weapon will take at least 6 weeks to get (probably just running with a 5-5 group for points), whereas the Hammer can be acquired rather quickly on my server.

Mios · 06/25/07, 5:44 AM

Originally Posted by Shalas

No, that isn't how time works. MH hits don't reset your OH swing timer. Your first few swings the OH will be right before the MH... and then you'll have some swings where they're far apart... and then you'll have some swings where the MH will be right before the OH, and then DOT DOT DOT

Ahh yes, now I realize how stupid I sounded. Thank you.

Tornhoof · 06/25/07, 7:24 AM

Originally Posted by Beowolf

Is there any way possible that you could add a toggle feature that can include or exclude certain instances? For instance, my guild obviously doesn't run MC/BWL/AQ40/Naxx anymore, nor are we in Hyjal or the Black Temple. However, when I do a blanket optimization, it includes items from those zones.

Basically, I'd like to build an 'optimized' set for the instances we are farming or bosses we will farm soon, without having to manually search what every boss drops and input numerous different items into your optimizer to see which is 'best'.

This is partially already possible since blizzard now includes the drop locations in their armory, there is just no ui yet for it.

• Disquette · 06/25/07, 9:06 AM

Ok, at this point I'll ask again (and I wont repeat the question until I do some myself) - who is checking the sim/formulaic dps figures against actual mobs to make sure that Pater/Tornhoof aren't missing anything?

• Disquette · 06/25/07, 11:21 AM

Pater, you might want to add this line into your Windfury damage calcs. I think I did it right for java, but don't know for sure:

		if (hand == "OH") D1 += (bonusAP)*wepSpd*0.5/14

it goes after this block, I think:

		double W =  (maxDmg - minDmg) * Math.random() + minDmg;

		double D1 = W + (currentAP + bonusAP)*wepSpd/14;

		D1 *= elementalWeaponsFactor; // See Disquette's testing

		D1 *= WEAPON_MASTERY_FACTOR;

		if (hand == "OH") D1 *= 0.5; // 50% penalty for offhand attacks

This is adding back in the damage bonus that the most recent patch activated. What is happening is that you do your normal WF damage calc, and then multiply by .5 for the offhand penalty. However, the "bonus damage on windfury" that was in the patchnotes is actually reflected by *not* halving the AP bonus on your offhand weapon.

This is just a quick and dirty way, if I understand your code correctly, to add it back in without disrupting the rest of your code.

Ilmatar · 06/25/07, 3:51 PM

I am very interested in these numbers, but I have not (yet) done testing with them. I have no passive haste to test with and so none of my testing would be deterministic.

It is my theory that haste will have the additional, hidden, benifit of improving consistency of windfury hits (not between MH and OH, but in lessening the time that we spend waiting for WF to proc). The scenario that I think would test that theory is to get several sets of very low dps weapons (any dps value works, speed is the important factor), windfury on both, blasted lands. The sets of weapons should have speeds which corolate to weapon speeds reachable via haste. The weapons should be buffed with Windfury and non-SS/shock combat should be recorded. If possible, Flurry should be de-specced for this test (my character is poor. I have +6 str gems in some stuff, and other stuff still has +hit gems from when +hit was god). The data recorded from the test should be parsed, and presented such that it shows time-> and the state of the windfury cooldown. The total time spent in combat during the 3s WF cooldown (WF procs * 3) should be recorded. The total time spent in combat outside the 3s WF cooldown (combat time - wfcd cimte). And the average time spent waiting for Windfury to proc, outside the cooldown (sequential data parsing needed).

The test should be repeated with the various sets of weapons.

It seems like common sense, to me, that haste should have the additional, hidden, effect of reducing time spent waiting for WF to proc. Like you have said previously, though, testing is key. I'm one of those jerks who likes to theorycraft and hates to test, but honestly, testing in World of Warcraft, for me, is frustrating. I feel like I can't set up deterministic tests; can't test against mobs (easily and solo) that provide accurate simulation data; and with no knowledge of the net code in WoW we have no idea if event order is maintained, or if it is maintained with some sort of update-priority (where some events, of higher priority, will be sent in guaranteed order), so combat log examination is not (necessarily) representative of the order in which events took place on the server.

Pater · 06/25/07, 4:19 PM

Thank you for pointing out that change, Disq. I've made the necessary changes and uploaded the new version.

Lujaar · 06/25/07, 7:57 PM

Originally Posted by Ilmatar

I am very interested in these numbers, but I have not (yet) done testing with them. I have no passive haste to test with and so none of my testing would be deterministic.

The best way to test this would be to tally your WF procs over a given time interval, then spec out of Flurry and do it again. Flurry should be close enough to a passive haste effect that if there's a substantial difference, you'll see it.

• Igniter · 06/25/07, 9:44 PM

Was thinking of taking a condensed version of the weapon enchants /WFT information and posting it on the wow forums for them, assuming no one cares/minds (so many put data and contributed to it, it's hard to ask everyone

)

Pater · 06/25/07, 9:52 PM

Slant's sticky in the official forums doesn't cover it?

• Igniter · 06/25/07, 10:01 PM

From the looks of it no one reads that thread. It would be more effective to have enhancement or WF in the title, to be frank.

slant · 06/25/07, 10:12 PM

Originally Posted by Igniter

From the looks of it no one reads that thread.

Who can tell how many "rockbiter offhand" posts were avoided because of my FAQ? I like to think that I educated a few people.

Enhancement and WF aren't in the title simply because it covers other specs.

◊ Rob · 06/26/07, 1:59 AM

Don't know if you guys are still looking for WWS parses, but here's one for Magtheridon.
http://www.lossendil.com/wws/?report...3729-4286&a=14

tdawg · 06/26/07, 4:11 PM

I can't seem to find this question anywhere, but with patch 2.1 is a 2h enhance shaman build viable? If so does crit chance become more important? Will 2h weapon dmg give me sustainable raid dps, or is DW still the way to go?

Ilmatar · 06/26/07, 4:35 PM

Nothing changed in patch 2.1 that would at all lend viability to 2h Enhancement.

Spike damage would still be a problem, keeping Unleashed Rage up (consistently, and constantly) would still be a concern, and no synergy with Shamanic Rage would mean you'd be out of mana pretty fast.

• Aeolian · 06/26/07, 4:49 PM

I'm sorry to sound like a complete newb if I am but I am trying to understand haste effects a bit more. Only have a single question at the moment. I've looked through most of the posts here concerning itemization and such and came across Pater's model. Why is there a drop in DPS between 44-55% haste?

Pater · 06/26/07, 5:38 PM

First, my results are just the output of my time-step simulation. They're not gospel truth, although I believe they're close enough to reality to base gear decisions on. Testing may confirm or disprove that. (Note on the previous page someone claimed that haste effects reduce the WF cooldown--only one person has made that claim and we have no logs, but it could be true and would require big changes to the model.)

Second, the dps drop happens when your flurried hasted speeds go below 1.50 seconds, putting the 2 swings following WF in cooldown (rather th