• Binkenstein

Ah, but if they implement that option via replacing Overload with a CL talent, and then do some other tree tweaking to bring LB back up to scratch, problem solved.

Going back to an earlier post about haste, looks like there are some areas to avoid.
Between 9% to 12%, 21% to 24%, 34% to 37%, and of anyone ever gets that far, 46% to 49%
Also, the difference between waiting for the LvB cooldown and just casting an extra LB seems to be about 30 dps, and the gain per 1% haste is ~24.3 dps at 2000 sp, increasing by 0.74 dps for every 100 sp after that.

 

tufy

Lintra, for base castable spells, the general rule is [cast time] / 3.5 = spellpower coefficient. There are very few exceptions to this, two of which just happened to be our lightning bolt and chain lightning. The reason for that is Lightning Mastery and Overload change a few patches back. Now that Blizzard considers that problem solved, they removed the artificial modifier, so we are back to this "normal base".

In essence, a lot of such info is easy to calculate and very logical. However, as Juice said it above, Shamans and the game itself are in a state of constant flux in beta right now, so compiling information beyond basics is a waste of time. Once things stabilize, Think Tank will surely have all that's needed.

Creativity requires the courage to let go of certainties.

 

Loi

I did some testing on beta with my lvl 80 shaman with test dummies. No flasks/food buffs, just self buffs with TOW/WOA totems, and flametongue weapon imbue. Major glyphs I use are Flametongue Weapon, glyph of Flame Shock and Glyph of Lava. No 4pc T7 bonus (only have 3 pieces), so it's not a factor. My spec is 57/14/0. My stats as tested (with totems down and flametongue weapon):
2304 spell dmg
26.77% crit +3% from TOW = 29.77%
11.07% haste + 5% from WOA totem = 16.07%
9.26% hit rating +1% drainei + 3% from talents = 13.26% ( I recorded about 3-4% miss rate, which seems accurate)
Relic I use is Totem of Hex (+182 LB and CL dmg)
trinkets:
Embrace of the Spider
Embrace of the Spider - Item - World of Warcraft
and
Mark of the War Prisoner
Mark of the War Prisoner - Item - World of Warcraft
Mark "on use" used once during each test.

I used "skull" boss dummy for testing. Did about 500K dmg each test. First, I used FS>LavaB>LB>LB>LB>LavaB>FS etc rotation, basically open up with FS, then Lava Burst, and fill in with LB until Lava Burst was up again etc. Refresh FS every time. Using that rotation I got about 2500dps average. If I used lvl 80 dummy with this rotation, my dmg with everything staying the same came up to 2950-3K dps.
Then I dropped FS and Lava Burst from rotation, and used regular CL>LB spam rotation, which gave me roughly 2200 dps. Adding FS to rotation with no Lava Burst added another 100dps or so. So, in my limited testing, our "savior" fire spells don't add that much, at least in solo testing. I'm sure in raid environment numbers will change, but I doubt they will close the 30-40% gap between top dps classes I saw in every 10/25 naxx raid. On Patchwerk for example, before the glyph and 10% increase in dmg for Lava Burst changes, top dps classes (hunters, mages, arms warriors etc) were topping 5k dmg fairly easy. I couldn't break 3.2K. I used full consumables (+120dmg flask, haste food etc), while some of the top dpsers didn't use any when I checked them. Our gear then was pretty much equal - full, or almost full T7.

I apologize if I made any mistakes in the limited testing I've done, it's not meant to be a full scientific testing, just some quick tests I've done while waiting for some raids/heroics to happen. If you have any questions, or requests for some other tests, I will try to help as much as I can. I'm still waiting for a raid to tests our new rotations with new glyphs, but no luck so far - after 3.0 went live, beta is nearly empty :P

 

Agash

Loi,

I think adding CL into your rotation should add about 120-130 DPS and using +damage food instead of +haste would add a bit more. Also, crit is not a big damage boost for us so it might be that changing your Flametongue glyph for an earth shock one and including that in your rotation might squeeze a little more out (though it will make your rotation extremely complicated - trying to balance 4 different CDs).

Even if you did all of this though you might add about 200 to your DPS which obviously leaves you far behind the heavy nukers in your raid and you would have to work harder on maintaining your rotation than any of them.

 

• Binkenstein

It appears that I can't post HTML code to link to a published chart from a googledoc sheet

However, it does give a nice save image function.

I've redone that math I mentioned earlier (wasn't calculating the rotation right, was rounding the sum of (8-GCD)/LB + 8/LB rather than rounding the two portions.



As you can see, there are definately haste percentages we want to avoid, and waiting for the LvB cooldown is fairly pointless (that's the orange line). Atm I just have LvB/FS/LB rotations, in a wait for the cooldown flavour (LB W) and a no-wait version as well (LB NW). Haste is kept in % (for the time being) as it's easier for me to work with.

PS: it looks alot better in the spreadsheet

PPS:
Correct!

 

Torrential

Just so I'm clear by LB W you mean doing nothing for ~1 second rather than cast another Lightning Bolt so you can get Lava Burst in sooner?

Or am I way off?

 

Loi

Finally was able to test new rotations in 10 man Naxx. Group comp was kinda meh - no SP, lock or druids were present. 4 DKs, 1 paly, one mage, one hunter, one enh, one ele and one resto shaman. In addition, the hunter and mage were not up to par to test agains - hunter was in greens doing like 1k dps, and mage was in mix of nice epics and blues, but somehow managed to do 1.5K dps.. dont ask me why (maybe he didnt train spell ranks? no idea). Anyways, the only viable dps besides me was blood/frost DK, with full epic gear. I was #1 most of the time, dropped to #2 a few times during large trash pulls where DKs aoe shined. On Patchwerk, i managed to do 3446 dps over the course of the fight (i had +120 flask and +35dmg food, and one heroism). DK was second with about 3.1-3.2K dps. Before the changes to glyphs and LvB, I only managed to do 3.2K, and that was in 25 man raid, with SPs and warlock present. I did 3.4K without them this time. Too bad those top dpsers from previous runs who were pulling 5k dps were not there. Anyone have an idea by how much my dps would go up if we had SP and lock present with their debuffs? Just curious if we finally getting close to the other classes in dmg.

 

Lucitron

The advantage with using Fire Nova and Magma Totem, is that they build upon what is already there. The drawback is Totem of Wrath. There are several solutions if Blizzard wish to go with this path for Shaman AoE, where possibly the most simple is just to make ToW to an Earth Totem, while the most elegant (and most complex) is to add the ToW effect as a passive effect to fire totems.

The advantage with a new spell is perhaps not so much the single target damage aspect, as it is the simple fact that it would allow Blizzard to create whatever spell they feel we need without being bound by previously set limitations (i.e. things like that totems are stationary and fragile). The big drawback is still as mentioned the demand for new graphics, animation, testing and all other kind of stuff it takes to get it to work.

Thunderstorm was mentioned by Ghostcrawler as a possible route for AoE, which perhaps is hard to understand considering all of its limitations with knockback, cooldown and mana regen. Still, if there is a will, then there is a way. Blizzard could just add something as simple as a damage over time component, if they wanted Thunderstorm to be our AoE. Further, we should also remember that there have been reports about griefing behaviour with Thunderstorm, so it is also quite possible that they will just do an overhaul and tweak it.

Finally, we have Chain Lightning. The advantage with using Chain Lightning would be that once again they build upon what is already there. There is nothing that state that Chain Lightning must hit loads of targets in order to be a viable AoE. The mechanism to allow it hit a 4th target is already there, and what is left is just to reduce its cooldown so that you can spam it. The drawback here is far more likely that of all spells, then changes to this specific spell will affect our single target damage. Out of all possible AoE spells, this solution might cause the most balance nightmares.


Now, to make a summary (and reason for this post), then as it should be apparent by now either AoE option is quite viable and possible. Perhaps the most important is not really what we get, but the fact that we have the opportunity to see Blizzard improve our class. I for one would love the opportunity to use my other fire totems, I would love to see new spells (Earthquake, eh?), I would love if Thunderstorm had a bit more of PvE usability, and I would love to see my Chain Lightning dance among a group of enemies.

At 7:45, offices are empty. Some could get bored, but I stay calm, I know how to adapt. While waiting for them I have time to take a coffee.

 

• Binkenstein

So while the guildies go clear AQ20/40 again, I got bored.

Here's the result:


I'm pretty sure I've got the rotation down, in a LvB/CL/LB priority system and adjusting to be able to go back to the start of the rotation with minimal time between the third LvB cd & use.
Rotation goes FS/LvB/CL/LBx2/CL/LBx2/LvB/CL/LBx2/CL

CL use = more dps, but I'd be impressed if anyone manages to get a full FS/LvB2/CL4/LB6 rotation perfect

[e]I added a second CL rotation only using it after LvB, which is a simpler rotation to handle.

 

tufy

If you'd cut the bolded CL, you would get LvB in faster, giving slightly higher dps. Also, your rotation after the second LvB goes on for 8.5 seconds without haste (provided FS lands after previous CL), meaning it will be somewhere between 8.09 (only wrath of air down) and 7.07 seconds (with 15% haste on gear, which will be the max we'll see on wotlk shaman without Heroism/Bloodlust, imo). This means you should probably throw in another LB at the end (8.73 seconds after LvB, so you can recast LvB now). Just my opinion

Creativity requires the courage to let go of certainties.

 

• Binkenstein

That's the rotation with 1% haste.

At 10% haste the LB counts are 2/3/2/1, 20% 3/3/3/0 and 33% 3/3/3/1 (30% was the same as 20%)

This was the "No Wait" + "Use CL asap" rotation. See the grey line for the one that skips the 2nd CL in each half of the rotation.

 

StErMi

Sorry for my nuub questions, i know LvB, CL, LB. But can you explain me the others?
LB4 = 4xLB?

Thank you very much

 

Tainter

Loi,

The only thing you'll get from SPs is 1 replinish and 3 spell hit. So if you weren't hitcapped your damage would go up ~3%. If you ran out of mana the replinish might add a bit of dps too, but I doubt you did run out of mana.

A talented Lock curse would give you another 9% bonus damage I guess.

Edit:
As a side: Once you end up with multiple abilities that have cooldowns and durations you're probably better off with a priority system rather than a cast sequence. Particularly because Haste can change throughout fights.

If you can't join them?
Beat them.

 

Torrential

Man, that's barely capable of being called a 'rotation.' No slight to you, you're just reporting the facts. To me, seeing that just reinforces the idea of a priority system as the way to go. You seem to have shown that it's ok to postpone Lava Burst a tad if that means getting CLs in.

There's no way to maintain a rotation like that given even the slightest movement and totem dropping.

I think there's an Enhance shaman mod that shows what your currently available (off cooldown) highest damage ability is. I think that's the direction we're going as well.

 

McBeefy

Awesome stuff Bink... I was doing it the manual way calculating haste first then trying to jam in my spell rotations to perfection and then calculating DPS over the rotation... which to say the least was tedious and time consuming.

Doing the FS,LvB,CL (insert filler LB's here) LvB, CL (insert filler LB's here) rotation makes the most sense to me except at obvious haste inflection points... trying to gear and gem around those inflections will be a chore, but thanks for better identifying them.

Having a CL follow a LvB for mana conservation purposes alone and not touching it otherwise should work OK, and help simplify DPS rotation without it becoming an onnerous task every time you get an upgrade.


** edit **

on the 2 CL NW rotation you have plotted do you have the DPS increase per 1 haste (or 100 or however you like looking at it) on average. It's kind of sad that 12% haste and 15% have the same relative dps in that rotation. And you adjust for perfect latency or 100ms?

 

Torrential

Just as a very rough benchmark here is a non-complete set of gear of item level 200 (which appears to be heroic 5-man and 10-man naxx level).

Item Comparison Tool - World of Warcraft

As you can see it's not difficult in the least to get 500+ haste rating. Then don't forget the 5% from ToW and 3% from Moonkin (and something else?) and you're easily over 20%.

Obviously we'll all vary, but if you're like me and haven't been in Beta and didn't really know what amounts of haste to expect, there's a ballpark.

 

McBeefy

Well given 8% haste is pretty much assured in a raid environment. I am mapping our haste vs. spell power.

Spell power looks like a better scaling stat at 80 given my sloppy spreadsheet work, and bink's much more elegant graph shows why I was having so much trouble.

so if I can lock down at say 320-340 or so haste to hit "sweet" spot of ~18% haste and use FS,LvB, CL, LBx3, LvB,CL LBx3 considering about 40-80ms lag... and then attempting to piece together upgrades and gemming around that, that is what I want to do. More haste is worse than less haste, unless you go a lot lower haste or a lot higher haste...

I still don't see how we keep up on DPS as gear gets better with moonkin, but hey I havent spent nearly as much time on that spreadsheet.

 

Jakuniku

The mod is ShockAndAwe by Levva on this site. I think that's the way to go too; I can't imagine trying to do one of those rotations w/ proc-based haste effects (skycall totem, quag's eye) thrown into the equation. I'll take a look at it tonight and see how hard it would be to make a similar mod for elementals, although there are probably people more qualified than me to do it.

 

• Binkenstein

Yeah, it's just easier to type it that way
I think the biggest thing shown from this is that using CL four times is not much better than using it twice, and that LvB is so much stronger than LB/CL that it doesn't matter if you waste a little bit of cooldown time. (~7k average damage vs ~4.3k for CL & LB).

Once I get some RL stuff done today (and lots of dailies) I'll export the googledoc sheet to excel, modify it for spellpower, and put it back up for more shiney graphs. *munches on breakfast*

PS: Jak: some sort of elemental addon would be useful. Especially if you combined cooldown use with a dogtag string to show a warning when you were down X% mana so you can use Thunderstorm effectively in PvE.

 

• Binkenstein

More graphs!
Same basis again, I've just been working on getting Spellpower & Critical Strikes in. These graphs compare the % gain from one row to the previous, eg: damage with 21% crit divided by damage with 20% crit (minus 1 to give a better graph as you can't set the axis start/end point in google docs)

As you can see, the first two look normal, with LB spam gaining the most from Crit and SP (more ammo in the "Elemental Scales Badly" campain?). The haste gain graph is a little odd with the LvB Wait rotation in there, so another copy of that for the NW rotations is included as well. As you can see, the nigh impossible 4CL rotation scales worst with haste normally, while the 2CL drops down from the LB line to the 4CL line just before the major drop, then back up to the standard LB Spam line. It's going to be very interesting to see what haste procs will be available in WotLK and how they can stack.

 

jimmyolsen

Bink, are the spikes displayed in the haste graph continuous curves or is it the classic TI-83 glitch where removable/jump discontinuity is shown continuous?

Example: % haste gain at 19%. Is the value at 18.9% closer to 8.8% as one would expect by the depiction on the graph, or is it nearly 0 as one would expect based on the pattern of previous points?

 

Jakuniku

After looking through wowhead, the haste proc effects I've found so far are just updated versions of Skycall totem and Quag's eye (not sure how to link WoTLK items):

Totem of the Elemental Plane - Your Lightning Bolt spell has a chance to grant 196 haste rating for 10 sec. (15% proc chance)

Embrace of the Spider (trinket) - Equip: Increases spell power by 98. Equip: Your spells have a chance to increase your haste rating by 505 for 10 secs. (10% proc chance, hidden cooldown unknown)

 

• Binkenstein

Those are discrete data points. I can produce some more accurate graphs going down to 0.1% steps rather than 1% steps, because no matter how I do it, it'll just have a series of points rather than a continuous line

 

Agash

I was having a discussion with the other officers in our guild today concerning off spec loot. Without going into detail it headed off in the direction of the value of Hit Rating on gear. I ran some calcs and come up with some suprising figures and I was wondering whether somebody could confirm my calcs or identify where it is I have gone wrong.

Comparing 2 gear sets (all figures are raid buffed);

Set 1 : 50% Crit, 11% hit (capped), 1000 SP
Set 2 : 50% Crit, 0% hit, 1173 SP (same item budget as set 1)

Set 1 Combat Table : Crit 50%; Hit 50%; Miss 0%
Set 2 Combat Table : Crit 50%; Hit 39%; Miss 11%

Set 1 : Average Damage per Hit : 2017
Set 1 : Average Damage per Crit : 4034
Set 1 : Average Damage per Cast : 3026

Set 2 : Average Damage per Hit : 2175
Set 2 : Average Damage per Crit : 4350
Set 2 : Average Damage per Cast : 3023

So, the average Damage per cast is marginally better with the Set 1. However, if we add a little more crit, the tables are turned. With 55% Crit using all other figures the same;

Set 1 : Average Damage per Cast : 3127
Set 2 : Average Damage per Cast : 3241

This seems to indicate that due to the high levels of crit now achievable in raids (largely due to the Winter's Chill/Improved Scorch debuff), it is possible to reach a crit level where spell power becomes the superior stat to Hit.

So - is this right or have I just made a fool of myself?

 

jimmyolsen

Could you just check between 18% and 19% at .1% intervals and post results (a full graph isn't necessary)? I'm curious if the we're looking at a situation where you would want to hold on your haste until you have an additional 70 rating or just 15 rating.

Obviously you can't get a continuous function since the rating system is by integer values only; however, we should be able to see if the spikes are singular outlying points as in a piece wise or an actual trend in the function.