[Shaman] Elemental Shaman DPS Spreadsheet? - Page 2

panny · 06/19/07, 5:48 AM

Originally Posted by Nite_Moogle

There really isn't one-- just spam lightning bolt over and over. At a 1.5 cast time, anything else you do is a global cooldown anyway and takes just as long, and has a lower coefficient. Dropping totems is really the only interruption you should have.

Lightning Bolt is a 2 second cast.

Binkenstein · 06/19/07, 6:27 AM

Originally Posted by Dicho

Honestly, the DPS vs DPM consideration relies intirely on the encounter, and your regen in raids.

Keeping in mind here that my gear is somewhat decent, I always have a shadow priest, and BoW available. I also use stopcasting to shave milliseconds off my LB casts, thus spending slightly more mana.

On the majority of fights, I find myself hard pressed to get rid of my mana. How mana efficient my spell rotation might be seems pointless when I am seemingly never in any danger of going oom.

With that in mind, I would love to find out whether just spamming LB, or CL/LB rotation, or FS/LB rotation, or FS/CL/LB rotation will most increase my dps.
- Assuming again this is in a raid situation with all the appropriate debuffs on the boss, CoE, CoS, Scorch, Misery etc, and that I have my totems and flasks/pots on.
My own rather primitive attempts at calculating this seems to indicate that using FS every 12 secs is more dmg then just spamming LB in the same cast window.

Threat is mostly a non-issue on static fights, even with Stormstrike goodness.

Well, you can see that on there, I just included the mana stuff as an afterthought.
I guess I could include a debuff check in next, but it's pretty safe to say that the only thing that would affect a rotation like this is whether the mob is nature vulnerable, or has +fire damage debuffs. Guess I'll add that into the next build, as I'm kinda curious now :p

This said, I have made some changes. I think I've implemented a sufficient way of accounting for Clearcasts. There's now a second option to use for CL, to say how many CC'd CLs you use (I think this effectively gets around the CC -> CL issue with minimal worry on my part, but allow fairly decent control of how it works in the long run). Assuming your crit % gives that many clearcasts, the remained will be used on LB/FS casts.

So for the debuffs, we want scorch, CoE at least.
Misery is all spell damage iirc, so it would scale everything up by 10%(?) and not be that useful for comparisons (unless we want absolute values).

koushi · 06/19/07, 7:20 AM

CoE doesn't affect Nature or did they finally change it. Misery is 5% iirc.

koushi · 06/19/07, 7:22 AM

nvm me. CoE factoring in for flame shock. Its hot and my brain is tired

Dicho · 06/19/07, 7:25 AM

The best ofc would be if blizzard gives us a Patchwerk-like encounter to test dps on, but that probably isnt likely.

For a comparison study misery is kinda pointless yeah.
The 10% + 15% from CoE and scorch definitely impacts the viability of FS in raids.

If Stormstrike was a flat 20% buff to nature dmg, then there would be no contest. But with only 2 of my 5(or less) bolts getting the SS buff, it doesnt impact the viability of flame shock.

There are other factors to consider when moving away from the pure LB spam ofc. CL can break cc's, and FS requires you to sometimes be unhealthily close to your target.

I also like to keep atleast 2k mana free to do spotheals(i'd consider myself shit if I don't throw heals when it can save someones life), and keep myself alive.

Voltan · 06/19/07, 8:00 AM

I am 99.9% sure the best rotation is LB,LB,LB,CL - Rinse and repeat. I will back this up with some evidence once I have the time (at work at the moment and just about to go on lunch). I myself have tried all possible rotations and found this one to be the best if you have the mana to see it through (its sick mps usage - however I have pretty good gear, 41% crit, and always get a a shadow priest and BoW).

The Rotations I am going to provide maths for are:

LB,LB,LB,CL
LB,LB,LB,CL with FS once every 12 seconds.
LB Spam - Purely for the point of comparison, from experience I've noticed the dmg difference between avg LB, and Avg CL is VERY slight - ~30ish dmg, obviously dependant on +spelldmg, the more you have the closer the gap (?) - Have yet to do any maths for proof.

Watch this space.

Ok, just to edit. Havent had a chance to look through this, still thinking about how to fit them into comfortable rotations without losing .5 of a second every now and then. I'm almost certain now FS is not worth using for a multiple of factors: - 11% less crit, cant proc LO, Short Range, More mana, Losing .5 - 1sec of FS uptime as well with standard spell rotation.

Squrf · 06/19/07, 11:25 AM

There is most definitely a point where LB12 is going to provide higher average damage than CL, simply due to the scaling and talents. With around 700 +damage, it's already a minimal difference for me. There should definitely be a breakpoint where LB spam provides higher DPS than a LB12/CL cycle and it should be roughly where the average LB12 hits 33% harder than CL.

Hamilburg · 06/19/07, 11:47 AM

Originally Posted by Squrf

There is most definitely a point where LB12 is going to provide higher average damage than CL, simply due to the scaling and talents. With around 700 +damage, it's already a minimal difference for me. There should definitely be a breakpoint where LB spam provides higher DPS than a LB12/CL cycle and it should be roughly where the average LB12 hits 33% harder than CL.

Which will never happen, due to the difference in cast times.

LB has a 86% coefficient compared to 71% on the first hit of CL. Dividing by the relative cast times, LB DPS increases by 43% of spell damage compared to 47% for Chain Lightning.

If you take lag into account CL gets hit harder than LB, but you need to be talking about .87 seconds of additional lag on top of casting before LB has a better DPS scaling coefficient than LB, and I dearly hope you are not trying to raid productively with that kind of latency.

Binkenstein · 06/19/07, 6:19 PM

He may have been talking about three-target damage rather than single target.

In that case, yes, it scales better, but it requires three targets to be hit (and often breaking CC).

Voltan, assuming my calcs are correct you should be able to plug that info straight into my spreadsheet

Voltan · 06/20/07, 3:54 AM

Will do, just gotta get my head round the different bits on there and I keep getting interupted... damn work. :P

RK · 06/20/07, 9:25 AM

Originally Posted by Dicho

The best ofc would be if blizzard gives us a Patchwerk-like encounter to test dps on, but that probably isnt likely.

Closest I know of is Maulgar. I (and ranged generally) can plant in the middle of the room and basically never have to stop casting until everything is dead (for maximum efficiency, switch targets with a bolt in the air when the previous guy is down to the point where you know he'll die before your cast finishes). Depending on positioning you might need to take a few steps to get range on the last mob, other than that the only interruption is totem recasts.

Of course, Maulgar is far from ideal for many other class to benchmark on, but it is pretty much ideal for elemental shaman.

Rephaim · 06/20/07, 9:39 AM

Dr. Boom in Netherstorm is perfect for that...

Hamilburg is correct... no amount of +damage will ever make LB better than CL in DPS because CL scales BETTER for DPS with +damage by 4.33%

Dicho · 06/20/07, 10:15 AM

Dr. Boom is a fantastic guy to test your dps on, however he does have some problems.

First of all the boom bots means you have to find interesting topographical features to stand on to use your shocks(unless youre using the pvp gloves).

Secondly, you are forced to use Dmg to target, or CL dmg will be vastly inflated.

Thirdly he is only lvl 68, vastly cutting down on partial resists and resists in general.

If you are testing raid dps, his hp isnt really sufficient either, as a few people whacking away at him will quickly kill him.

The beauty of Patchwerk(apart from his handsome features) was that you could not overaggro him(well, normally anyway), he stood in a fixed position, and there was no movement or anything other then button mashing involved for the dps.

Back to the spell rotation:

Originally Posted by Voltan

Ok, just to edit. Havent had a chance to look through this, still thinking about how to fit them into comfortable rotations without losing .5 of a second every now and then. I'm almost certain now FS is not worth using for a multiple of factors: - 11% less crit, cant proc LO, Short Range, More mana, Losing .5 - 1sec of FS uptime as well with standard spell rotation.

I must admit I did not factor LO proccs in my considerations of FS.
In the long run LB/CL spam is probably the best, albeit boring rotation to use.

Voltan · 06/20/07, 11:46 AM

Ok, finally had a chance to have a look at things and its time for some numbers. I'm using Binkensteins Elemental Rotation spreadsheet. The details I input myself were 1080 spelldmg and 42% crit.

LB Spam over 60 seconds.
30 Casts of LB - 1442 DPS over 60 seconds.

LB/LB/LB/CL rotation over 60 seconds.
22 casts of LB & 10 casts of CL - 1574 dps over 59 seconds.

FS/LB/LB/LB/LB/LB/LB rotation over 60 seconds.
25 casts of LB & 5 casts of FS - 1475 dps over 57.5 seconds (When I added an extra LB, the DPS dropped to 1474 if someone with more time could look at why this happens on the spreadsheet I would be greatful)

Working on an feasible rotation for CL&FS integrated into LB spam, Spreadsheet is kicking out 1607 dps for a rotation containing 18 LBs, 10 CL's and 5 FS's over 58.5 seconds. Just need a bit of time to calculate when its possible to add in the CL&FS's. It does look to be a higher dmg rotation, however it looks like we'd have a hunter-style pain trying to weave these into the rotation (CL having to be cast on CD, and FS as soon as the DoT portion is finished), the may result in the loss of regular LB casts.

Also, the MPS is coming out at 228, which 1140Mp5 resulting in OOM in only 40seconds without Mp5 gear/Raid buffs. hmm.. maybe not possible to keep that sort of mana drain up for long - I'll work out exactly how long it's possible for me with full raid buffs to keep this up later, Although to be perfectly honest, the hassle this will cause in a raid enviroment, may result in more 'dropped' spells due to movement, lag, lack of mana etc etc. for only an extra 33dps. (Hmm, that would be pretty noticeable) - Watch this space.

Just reading back over this post, the numbers are pretty damn high DPS. Are the crits from talents added onto these results?

Anyway... back to work - will look into this as soon as I have more time.

Voltan · 06/20/07, 11:57 AM

Originally Posted by Dicho

Dr. Boom is a fantastic guy to test your dps on, however he does have some problems.

First of all the boom bots means you have to find interesting topographical features to stand on to use your shocks(unless youre using the pvp gloves).

Secondly, you are forced to use Dmg to target, or CL dmg will be vastly inflated.

Thirdly he is only lvl 68, vastly cutting down on partial resists and resists in general.

Yer, DR boom is pretty unrealistic. I cranked out 1.8kdps over 3minutes here. - Lol..

I'd love a boss that spawns little adds as well, things like getting another target whilst DD-ing with CL make me tingly inside (the Totems/pet @ Karathress net me an extra ~50 dps).