The One Stop Elemental Shop - Page 64

innocent · 04/11/08, 6:50 PM

In my gear, this gives about 83 spell damage equivalent when converted from haste to damage equivalent.

Which is basicly 3 more spellpower than Crusade for an increased manacost due to haste,
Depending on how much it synergizes with skycaller it might not be superior.

Daidalos · 04/11/08, 7:07 PM

Originally Posted by Moshne

I agree wholeheartedly, despite that, I think it is easy to overvalue it.

(ROUNDUP(6/Quaghastedcast))*Quaghastedcast) / (45+(1/((1/avgcast)*0.10)))

This gets you an average haste value for the trinket, using your average cast for the proc rate, and the QuagHaste for the proc duration. This does not include any "extra" haste like Skycall, as its impossible to model, and the numbers seems to be showing the other totem as better anyhow.

In my gear, this gives about 83 spell damage equivalent when converted from haste to damage equivalent. I use:

Spell Damage = 1
1 Crit rating = 0.857318064
1 Haste Rating = 1.270297572

Well the numbers on skycall seem to also be conflicting. I"d say haste is a little higher than your DEP values but thats not really the point. What i'm wondering if we can get a synergistic affect of stacking haste procing items so that the relative value of each increases as you stack more. This was why do did so much simming of the skycall relic and I'm begining to wonder if we need a better sim so we can better answer this than going to Dr Boom for 3 hours.

Hmm I guess I need to get alot more free time so I can finish working on my ele sim.

Moshne · 04/11/08, 7:24 PM

My haste/crit/damage values are arrived at by simulating the DPS rotation, then doing 3 identical tests but with each respective stat increased by 1. The difference in the Adjusted DPS and the "normal" DPS is the stat ratio. It obviously varies based on each stat.

I'm curious what methodology others have used to come at their values?

Binkenstein · 04/11/08, 8:34 PM

Originally Posted by innocent

Which is basicly 3 more spellpower than Crusade for an increased manacost due to haste,
Depending on how much it synergizes with skycaller it might not be superior.

I pretty much ignore extra mana cost now that I'm LB spamming.

Skycall uptime is ~60% from what I've seen of raiding recently, so I'm more inclined to use Ancestral atm

The spreadsheet is still being worked on. Atm I think I'll just put static values for haste procs to simplify things (and avoid circular references), re-do the enchant references so I can use that space for individual gem selection, then double check things before a final release.

sheng · 04/12/08, 4:06 PM

hi,Binkenstein.
After dling your new simulator, I have found you missed few items that added in patch 2.4, such as new bg neck with spell haste, new 60 badges spell haste cloak, etc.
Btw, great work as usual, keep it up.

innocent · 04/12/08, 6:16 PM

I pretty much ignore extra mana cost now that I'm LB spamming.

Well I am actually having some mana issue even spaming LB.
Or I would better say that mana is not a non-issue to me since I never put shadowpriests in my group.
We run a classic comp with 1 tank/hunter group, 1 mele group, 1 healer+SP group, 1 mages+SP group and my Elem+4 warlock group. We found it wasnt worth it dps-wise to drop a lock for a SP in group5. SP do not scale amazingly with crit, they are almost hit caped with talents. Both SP and Mages all in all get better use of manatide and improved manapring totems. Destro warlocks are the only ones that can manage to avoid hit cap and benefit more of my draenish totem of wrath.

Binkenstein · 04/12/08, 7:38 PM

Originally Posted by sheng

hi,Binkenstein.
After dling your new simulator, I have found you missed few items that added in patch 2.4, such as new bg neck with spell haste, new 60 badges spell haste cloak, etc.
Btw, great work as usual, keep it up.

Yeah, I know, that's why it's still a Beta...

Originally Posted by innocent

Well I am actually having some mana issue even spaming LB.
Or I would better say that mana is not a non-issue to me since I never put shadowpriests in my group.
We run a classic comp with 1 tank/hunter group, 1 mele group, 1 healer+SP group, 1 mages+SP group and my Elem+4 warlock group. We found it wasnt worth it dps-wise to drop a lock for a SP in group5. SP do not scale amazingly with crit, they are almost hit caped with talents. Both SP and Mages all in all get better use of manatide and improved manapring totems. Destro warlocks are the only ones that can manage to avoid hit cap and benefit more of my draenish totem of wrath.

You're better off having 3xLock/Spriest/You as the locks will get more out of a shadow priest than Mages will

Moshne · 04/12/08, 7:52 PM

Locks tend to get the short end of the stick in the Shadow Priest priority it seems due to archaic beliefs about their lack of mana dependence due to Life Tap. While this is obviously neither a warlock nor a shadow priest thread, there is something to be said for putting the SP with the locks and ele shaman, as you can go into a higher DPS mode, and the locks damage output will go up a reasonably large amount as well. Our traditional makeup gives our higher DPS SPriest to the locks, mages, hunters then lastly the healers, especially post 2.4, healer mana regen is not nearly the same beast it once was.

Chaostheoryx · 04/12/08, 8:51 PM

Someone asked me which was better, a 4:1 rotation or LB spam. I couldn't really give an answer aside from the two popular spreadsheets. I've always assumed that a 4:1 rotation was a considerable DPS gain over LB, mirroring Bink's SS results of adding ~60 DPS with my current gear (no 4pc, using skycall).

However... I tried this test on Dr. Boom, complicated as it was. I did several runs until I had a normal crit rate and imported the combat log entries to a spreadsheet. I used WWS to break out the damage done to Dr. Boom only (Filter, Target, Dr. Boom, Hit. Or Breakdown, Dmg out). I calculated the DPS against Dr. Boom by dividing the total dmg to target by the seconds between first and final cast. Final results from Excel:
4:1 rotation: 1380, 1300, 1407.
LB Spam: 1472, 1528.

That being rather time consuming, I brought a warrior to Blasted Lands east of the Dark Portal. He tanked one of those Servants of Razelikh and I tried about 10 sessions of each rotation. Granted, one would expect this test to be better for 4:1 than the first, as LO/TLC procs and crits may be expended on the bots and uncounted. Results recorded with Recount:
Best 4:1 rotation 1419 (44% crit).
Best LB spam: 1480 at 33% crit (!).

So, anyone else have any real-world data on 4:1 performance relative to LB spam? Is there an explanation for these counter-intuitive results short of the limited sample size?

draconis0101 · 04/12/08, 10:33 PM

I have a quick question bout shaman stats. Being as how this is the elemental shaman. Is it normal for our character screen to not show our talent crit% added? It only shows 20.05% crit but doesn't show Tidal Mastery and Call of thunder. So should we be aiming or 25% crit without that showing so that it equals 35% or including it?

Binkenstein · 04/13/08, 1:26 AM

Coming back to that overload resets the CL cooldown timer point from earlier, has anyone observed this to any extent?
It may be worth sitting on Dr Boom for a while mashing the CL button (with logging active) and then see what the time differences are like.

If it adds ~0.5 seconds on, then this may explain why the math says it's higher, but in practise it's lower. It would also explain why a 4/1 rotation is fine as you'd have to get a 1.625 second cast for it to be a problem again.

Bulger · 04/13/08, 1:42 AM

Depending on the fight really depends what group I support, alot of the time I have 3 mages and a Spriest in my group(4:1 rotation) but if its more of a healer intense fight I dont mind giving up the Spriest and I support the warlock group with a resto in the mage group (more of a LB spam but also situational). I dont feel I ever have mana problems even while not in the Spriest group you just need to take pots early and often (I do have the new alch stone, hax).

You're better off having 3xLock/Spriest/You as the locks will get more out of a shadow priest than Mages will

I dont know if I agree with that, Mage dps depends on being able to drop heavy mana for heavy dmg using a mana efficient rotation kills the dmg output (we run all fire mages, 1 scorch bitch) I really think group placement is situational but reguardless I always end up in the mage group the last 20% for Magecute range.

Nizari · 04/13/08, 4:12 AM

Originally Posted by draconis0101

I have a quick question bout shaman stats. Being as how this is the elemental shaman. Is it normal for our character screen to not show our talent crit% added? It only shows 20.05% crit but doesn't show Tidal Mastery and Call of thunder. So should we be aiming or 25% crit without that showing so that it equals 35% or including it?

If you read the talents a little more carefully you'll notice it only gives the critical strike % to Lightning Bolt and Chain lightning (not shocks and only 5% to heals). In other words you can't really expect it to be shown on the main character screen.

Chaostheoryx · 04/13/08, 4:26 PM

Originally Posted by Binkenstein

Coming back to that overload resets the CL cooldown timer point from earlier, has anyone observed this to any extent?
It may be worth sitting on Dr Boom for a while mashing the CL button (with logging active) and then see what the time differences are like.

If it adds ~0.5 seconds on, then this may explain why the math says it's higher, but in practise it's lower. It would also explain why a 4/1 rotation is fine as you'd have to get a 1.625 second cast for it to be a problem again.

CL #1	4/13 13:24:15.520  SPELL_CAST_START,0x00000000001F6B6E,"Chaostheoryx",0x511,0x0000000000000000,nil,0x80000000,25442,"Chain Lightning",0x8
LB #1	4/13 13:24:16.977  SPELL_CAST_START,0x00000000001F6B6E,"Chaostheoryx",0x511,0x0000000000000000,nil,0x80000000,25449,"Lightning Bolt",0x8
	4/13 13:24:17.385  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213AF8,"Boom Bot",0xa48,25442,"Chain Lightning",0x8,811,8,0,0,0,nil,nil,nil
	4/13 13:24:17.385  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213B37,"Boom Bot",0xa48,25442,"Chain Lightning",0x8,1151,8,0,0,0,nil,nil,nil
	4/13 13:24:17.385  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004F3C207428,"Dr. Boom",0x10a48,25442,"Chain Lightning",0x8,3617,8,0,0,0,1,nil,nil
	4/13 13:24:17.719  SPELL_AURA_APPLIED,0x0000000000000000,nil,0x80000000,0x00000000001F6B6E,"Chaostheoryx",0x511,37658,"Electrical Charge",0x1,BUFF
CL LO ->4/13 13:24:17.761  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213B05,"Boom Bot",0xa48,45302,"Chain Lightning",0x8,856,8,0,0,0,1,nil,nil
CL LO ->4/13 13:24:17.761  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213B37,"Boom Bot",0xa48,45302,"Chain Lightning",0x8,595,8,0,0,0,nil,nil,nil
CL LO ->4/13 13:24:17.761  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213AF8,"Boom Bot",0xa48,45302,"Chain Lightning",0x8,1753,8,0,0,0,1,nil,nil
	4/13 13:24:18.129  SPELL_AURA_REMOVED,0x0000000000000000,nil,0x80000000,0x00000000001F6B6E,"Chaostheoryx",0x511,37658,"Electrical Charge",0x1,BUFF
TLC ->	4/13 13:24:19.247  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004CEC213AF8,"Boom Bot",0xa48,37661,"Lightning Bolt",0x8,736,8,0,0,0,nil,nil,nil
LB #1	4/13 13:24:20.172  SPELL_DAMAGE,0x00000000001F6B6E,"Chaostheoryx",0x511,0xF530004F3C207428,"Dr. Boom",0x10a48,25449,"Lightning Bolt",0x8,1721,8,0,0,0,nil,nil,nil

CL-LB Interval: 0:00:01.457s
LB 1 Time to Cast: 0:00:03.195s

Think that's what you're looking for. I set up AutoHotKey to send CL, wait 1.390s, send LB. Then I just hit that 'macro' as CL came off CD.

I've discovered that my CL rotation is giving me a minimum of 1.62s intervals between CL and the next LB, as opposed to the 1.36s or 1.28s it should be. I dare to say, most everyone is likely losing a lot of time on their CLs. However the client-server cast protocol works (I've read a lot on it, still not entirely sure), spamming LB doesn't cause any problems with the 'mini-GCD' incurred by failed cast requests. Spamming LB every 10ms, I get an average interval of 1.866 (Min of 1.85 @127 haste). Unfortunately, switching from CL to LB is very tricky. I still don't understand where this 'mini-GCD' is coming from. One would think that the client would not send a cast request to the server when it knows full well that it's inside the GCD (which is always true for CL). Anyone have any insight into this?

As it stands, I'm going to try out my new CL macro that just does CL, wait 1.4s, LB.

BTW, I found this _very_ simple tool for displaying your cast interval in real-time: LagInfo | World of Warcraft Addons | Curse

Heleia · 04/13/08, 6:23 PM

Would like to know if anyone has good math on Icon, LC and blessings deck for dps.

Binkenstein · 04/13/08, 6:54 PM

Chaos, not quite what I was looking for.

What I want is a log for CL, CL overload, and the next CL cast. I'm probably going to have to go do the testing myself (stand at Boom, and just cast CL).

Gelanin · 04/14/08, 5:04 AM

How would most of you rate/compare the new Sorcerer Alchemist's stone trinket ?

From a purely DPS perspective i guess it's +63 passive damage isnt as good as some of the alternatives, but it does add quite abit of "usability" with its very nice 40% gains from health/mana pots.

Perhaps especially on fights like the 3'rd boss in MH, where mana is at a premium.

Juice · 04/14/08, 10:41 AM

I imagine you can put a damage equivalent on the extra mana you get from a potion based on your rotation and stats if you're really interested. Then you can compare it to the other available trinkets.

I'm not an alchemist, but I'd only consider using it on Kaz'ragol due to his mana leeching debuff. And for 1 boss who we already defeat without said trinket, I'd be unlikely to devote the creation materials and bag space to said trinket. Unless you're unable to raid with a shadow priest I don't believe this is a good trinket for elemental shaman.

Handyhoof · 04/14/08, 2:23 PM

Chaos, I'm not sure that's being interpreted correctly. The combat log doesn't show when something was necessarily cast, but rather when it hit, meaning delivery time. CL has no delivery time, but LB does. going from CL to LB is going to be cast time of LB + delivery time.

Chaostheoryx · 04/14/08, 7:46 PM

Handyhoof, the new combat log shows two separate entries, SPELL_CAST_START and subsequently SPELL_DAMAGE. I ignore the SPELL_DAMAGE entries as those are affected by travel time (The LB 1 Time to Cast thing in my last post was, in retrospect, entirely irrelevant). I'm pretty sure the SPELL_CAST_START message is one sent by the server to the client when a spell request (button press) has been granted (Server starts the cast after checking range, mana avail, etc.).

Also, I've seen a range of CL->LB intervals from 1.36s to 1.65s. I can keep the interval low by automating the casting entirely (timing w/ AHK). All of my testing has been done from atop the flat rock behind Dr. Boom.

Binkenstein · 04/14/08, 10:51 PM

No, it's not the time between CL and LB, it's the time between CL SPELL_DAMAGE and the next CL SPELL_CAST_START that we want to have a look at. It should be 6 seconds + user reaction time, and what we're looking for is whether two SPELL_DAMAGE entries have a longer period between the first and the next SPELL_DAMAGE entry, or the spacing of the SPELL_CAST_START entries (allowing for cast speed).

Mirranda · 04/15/08, 2:30 AM

So, I'm sitting on what everyone plus myself can consider the #2 and #3 best trinkets in the game for our class.

From everything I've modeled The Skull of Gul'dan is quite secure as the #1 dps trinket, especially once we're far enough that I replace my t6 with the new Sunwell loot (which we're on Felmyst right now so it's not "too" far off). For me it's not a need of spell hit at the moment, but come full best in slot gear, a skull would be needed to avoid gemming some spell hit or going with second best in slot for an item or two.

My question is, is it a large enough upgrade for an elemental shaman to take it over a mage and/or warlock as it's also best in slot for them?

gman · 04/15/08, 2:49 AM

From your armory, you've already exceeded your hit cap. The hit on the skull will just go to waste.

Edited to add: How much hit do you plan to lose?

Mirranda · 04/15/08, 2:50 AM

Originally Posted by gman

From your armory, you've already exceeded your hit cap. The hit on the skull will just go to waste.

Edited to add: How much hit do you plan to lose?

Was referring to being best in slot Sunwell gear. I realize the hit goes to waste as I am now.

Going with keeping 1 piece of old t6 (looks like gloves at the moment), I'd lose ALL of my hit except for gloves. So I'd have 19 hit from gloves, +14 hit from helm enchant.
Total hit would be 33 spell hit rating ~2.5%?

EDIT: Looks like part of the hit from the Skull would still go to waste, maybe drop 1 point in elemental hit talent in favor of mp5 to balance it.

What I'm looking at using:
Helm: Cowl of Guldan
Shoulders: Erupting Epaulets
Chest: Garments of Crashing Shores
Bracers: Skyshatter Bands
Belt: Skyshatter Cord
Legs: Chain Links of the Tumultuous Storm
Boots: Skyshatter Treads
Trinket: The Skull of Gul'dan
Weapon:Sunflare (Dagger, Sunwell): +17 Stamina, +20 Intellect, +23 Haste, +30 Crit, +292 Damage

Binkenstein · 04/15/08, 3:14 AM

Originally Posted by Mirranda

EDIT: Looks like part of the hit from the Skull would still go to waste, maybe drop 1 point in elemental hit talent in favor of mp5 to balance it.

Please tell me you're joking when you say this. Not only does it give you 2% hit per point, but it reduces threat by 3/3/4% which is something we should never ignore. The only hit talents I consider "optional" are the ones in the restoration tree.

Also: 2nd place trinket imo is Hex Shrunken Head