Diablo® III

CM/WW DPS Tests and Observations

So I did a whole bunch of tests on MP8 Ghom and tried to remove as many variables out of there as I could. Originally, I did the test to see how much of a dps boost using Fire Walkers w/ Conflag boosted my dps running ShockNadoShards (SNS) but expanded the test to see how much DPS I'd lose if I went to more defensive builds. Figured I'd share my results with those who might be interested in actual field testing.

The Test:
Ran MP8 Ghom (70.6M hp) and timed how long it would take to kill him. At the start of the fight, I'd teleport within melee range of him and hit frost nova. At the moment I hit frost nova, the timer started. I didn't move at all after (held shift the entire fight to stay still) and only focused on maintaining the highest possible DPS with the spec under test while casting Diamond Skin and Frost Nova on cooldown. I did not have any AP issues after initial windup of 5 or 6 twisters. I also did not record these tests with a FRAPS type program, they were all timed by hand so there's going to be some error of I'd say +/- a second in the timing.

There's also few variations in kill times due to movement from Ghom. On some tests, he didn't move at all while on other tests, he'd walk almost a 180 around me. Movement during tests reduces damage output because he probably runs out of range of a few tornadoes. I don't know if this is a variable which can be removed with better gear or not; I'll leave that to someone else who has a higher end set of gear to test with.

Relevant test character stats were done using ShockNadoShards as a baseline. If a spec changes dps, the change and dps are listed.
2.77 APS, 53% Crit, 18 APoC, 119065 Unbuffed DPS w/ Scoundrel out and 3% Crit Buff.
Note, Scoundrel never died and never moved once I did my Frost Nova and started timer. Apparently, 819 LoH is more than enough for him to sit in Ghom's clouds and not lose hp (serious AoE damage reduction on followers now... I want that).
My normal boots are a pair of huge EHP Ice Climbers and also have around 80 more int than my Fire Walkers used for testing.

The Data:
Note, all descriptions are assuming standard SNS with a the change listed.
And disclaimer: there's going to be a few percent of error in these tests due to statistical variation and factors beyond the best of my ability to control (counting time, movement by Ghom, extended button mashing induced fatigue, etc). Numbers given are not meant to be super precise but were intended to be used as general guidelines.

Original Tests

ShockNadoShards (119065 Buffed DPS, 53% Crit) - 1:25, 1:26, 1:28 for an average of 1:26.3 = 817k Sustained DPS (6.86x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!bWg!YYcccc

Pin Point Barrier (124701 Buffed DPS, 58% Crit) - 2:05, 2:05, 2:02 for an average of 2:04 = 579k Sustained DPS (4.87x Baseline DPS, 4.63x Spec DPS)
http://us.battle.net/d3/en/calculator/wizard#VQXYhO!bWg!YYcccZ

Frozen Storm Rune on Frost Armor (119065 Buffed DPS, 53% Crit) - 2:08, 2:08, 2:05 for an average of 2:07 = 556k Sustained DPS (4.67x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhf!bWg!YYcccc

SNS + Firewalkers & Conflag (114968 Buffed DPS, 53% Crit) - 1:24, 1:23, 1:26 for an average of 1:24.3 = 837k Sustained DPS (7.03x original, 7.34x accounting for Fire Walker's loss in int)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!eWg!YYcccc

Prismatic Armor (119065 Buffed DPS, 53% Crit) - 2:12, 2:06, 2:11 for an average of 2:09.7 = 544k Sustained DPS (4.57x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhO!bWg!YYcccc

SNS w/ Blur instead of Evocate (119065 Buffed DPS, 53% Crit) - 1:27, 1:27, 1:26 for an average of 1:26.7 = 814k Sustained DPS (6.84x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!ZWg!YYcccc

SNS w/ Glass Cannon over Evocate and Time Warp Rune on Slow Time instead of Teleport (136924 DPS, 53% Crit) - 1:09, 1:11, 1:08 for an average of 1:09.3 = 1,019k Sustained DPS (8.56x)
http://us.battle.net/d3/en/calculator/wizard#VgXYhT!YWg!YZcccc

SNS w/ Glass Cannon over Evocate and Crystal Shell Rune on Diamond Skin (136924 DPS, 53% Crit) - 1:26, 1:25, 1:25 for an average of 1:25.3 = 828k Sustained DPS (6.96x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!gYW!YYaccc

Additional tests but only tested once because my hands were going to die.

SNS w/ Frozen Mist Rune on Frost Nova (119065 Buffed DPS, 53% Crit) - 1:26 = 821k Sustained DPS (6.90x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!bWg!YYcbcc
Note, tried to cast Frost Nova about once every 7 seconds.

SNS w/ Glass Cannon instead of Evocate (136924 Buffed DPS, 53% Crit) - 1:16 = 929k Sustained DPS (7.80x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!YWg!YYcccc

SNS w/ Cold Snap Rune on Frost Nova (119065 Buffed DPS, 53% Crit) - 1:39 = 713k Sustained DPS (5.99x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!bWg!YYcZcc

SNS w/ Crystal Shell Rune on Diamond Skin (119065 Buffed DPS, 53% Crit) - 1:36 = 735k Sustained DPS (6.18x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!bWg!YYaccc

Frozen Storm Rune on Frost Armor and Blur instead of Cold Blooded (119065 Buffed DPS, 53% Crit) - 2:27 = 480k Sustained DPS (4.03x)
http://us.battle.net/d3/en/calculator/wizard#VQXYhf!bZg!YYcccc

Time Warp Rune on Slow Time instead of Teleport (119065 Buffed DPS, 53% Crit) - 1:15 = 941k Sustained DPS (7.90x)
http://us.battle.net/d3/en/calculator/wizard#VgXYhT!gbW!YZcccc
Note, tried to cast Slow Time about once every 7 seconds.

Stretch Time Rune on Slow Time instead of Teleport (126152 Buffed DPS, 53% Crit, 2.94 APS) -
1:28 = 802k Sustained DPS (6.73x)
http://us.battle.net/d3/en/calculator/wizard#VgXYhT!gbW!Yccccc
Note, tried to cast Slow Time about once every 7 seconds.

Crit Scaling Tests

ShockNadoShards (120052 Buffed DPS, 58.5% Crit, 2.79 APS, w/ Hellfire Proc) 1:19, 1:20, 1:21 for an average of 1:20 = 883k Sustained DPS (7.35x relative to test DPS)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!bWg!YYcccc

SNS + Firewalkers & Conflag (115564 Buffed DPS, 58.5% Crit, 2.79 APS, w/ Hellfire Proc) - 1:14, 1:17, 1:17 for an average of 1:16 = 929k Sustained DPS (8.01x relative to test DPS)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!eWg!YYcccc

SNS w/ Blur instead of Evocate (120052 Buffed DPS, 58.5% Crit, w/ Hellfire Proc) - 1:19, 1:20, 1:20 for an average of 1:19.7 = 885k Sustained DPS (7.38x relative to test DPS)
http://us.battle.net/d3/en/calculator/wizard#VQXYhT!ZWg!YYcccc

SNS w/ Glass Cannon over Evocate and Time Warp Rune on Slow Time instead of Teleport (138061 Buffed DPS, 58.5% Crit, 2.79 APS, w/ Hellfire Proc) - 1:07, 1:07, 1:04 for an average of 1:06 = 1,070k Sustained DPS (8.91x)
http://us.battle.net/d3/en/calculator/wizard#VgXYhT!YWg!YZcccc

The Observations

Shocking Aspect adds a ton of DPS over Prismatic Armor. Definitely worth using if you can as expected.

Pin Point Barrier gives a small boost over Prismatic Armor (~6.5%) as expected.

Frozen Storm Rune really isn't much of a dps boost at all for single target. You're probably better off using Prismatic Armor or going Pinpoint Barrier.

Firewalkers + Conflag definately gives a DPS boost but it might not be as big of one as you'd hope if you have to give up a fair amount of DPS (say from Icewalkers).

Evocate doesn't give much of a tangible benefit for DPS. I'd say its in the error.

Glass Cannon works as advertised (~13.7%).

Cold Blooded does seem to work on chilled effects when comparing "Frozen Storm Rune on Frost Armor" and "Frozen Storm Rune on Frost Armor and Blur instead of Cold Blooded" (~16%).

If you're going to give up one ability for a defensive one going from SNS standard, give up Diamond Shards over Storm Armor (to anything else). Storm Armor is just worth way more DPS and going to Crystal Shell on Diamond Skin is probably a bigger overall boost to survival than any other change you could make for a CM/WW spec.

Frozen Mist does a surprisingly high amount of damage but a chunk of it might be due to not spamming Frost Nova as in the baseline comparison. Casting once every 8 seconds instead of on CD also requires more mental overhead to time. Spamming it eats more CD's which reduces DPS and isn't helpful with this particular rune which leaves a frost patch on the ground for 8 seconds. And it removes your ability to freeze things... just more or less acts like a movement speed down Time Warp with some additional damage.

Stretch Time Rune on Slow Time didn't really help me at all for DPS, likely because I didn't break the next threshold for WW at 3.0 APS and I had to spend a cast to put up Time Warp every 7 or so seconds.

Good use of Time Warp Rune can match Glass Cannon DPS output assuming you can give up Teleport. And its a party buff. Pretty awesome there.

If you want any other specific combinations, post below and I'll squeeze out some time to test them.

Things I'd love someone to test:
SNS DPS scaling w/ Attack Speed... particularly hitting 3.0 APS vs. 2.74 APS and going down on breakpoints if possible while keeping character sheet DPS and crit levels as close to each test as possible. I might be able to do it in a few more upgrades (Enchantress + Stretch Time puts me at 2.99) but can't do it yet. Wondering in particular, how much the multiplier of character sheet DPS changes with higher APS.

Edit: Added a request test and disclaimer.
Edit: Added crit scaling tests.
Edited by holyfork#1768 on 10/31/2012 3:57 PM PDT
Reply Quote
Thanks for the effort. It'll be useful to the community.
Reply Quote
Awesome!! Someone should also do a test of the kiting builds...
Also what about the hybrid WW/Meteor build that Aphrael demoed earlier today? (Meteor/Molten instead of teleport) but maybe that doesn't work with single targets?
Edited by Omar#2813 on 10/29/2012 3:03 AM PDT
Reply Quote
This is fantastic. Thanks a lot! I really appreciate what you did here, it helps :)
Reply Quote
Big ups from Australia. Thanks for the info.
Reply Quote
Stretch Time Rune on Slow Time didn't really help me at all for DPS, likely because I didn't break the next threshold for WW at 3.0 APS and I had to spend a cast to put up Time Warp every 7 or so seconds.


Bubbles use attack turns? Didn't think they did.

Ran MP8 Ghom (70.6M hp) and timed how long it would take to kill him. At the start of the fight, I'd teleport within melee range of him and hit frost nova. At the moment I hit frost nova, the timer started. I didn't move at all after (held shift the entire fight to stay still) and only focused on maintaining the highest possible DPS with the spec under test while casting Diamond Skin and Frost Nova on cooldown. I did not have any AP issues after initial windup of 5 or 6 twisters. I also did not record these tests with a FRAPS type program, they were all timed by hand so there's going to be some error of I'd say +/- a second in the timing.


Prolly a dumb question but we are using Shards correct? Even if evocation didn't cause any effect on EB due to maxed out server cast rate, it should have a full 15% effect upon shards cast rate. I'd except that to have at least a ~2% effect upon your test just from that alone. Your seeing a 0.3% effect. Margin and error as is that's massively disconcerting.
Reply Quote
So far I think Iceclimbers+GC outdamage FW+Confla, I just used GC in MP 10 and the speed is the same as when I used Confla in MP 9. ICs boost EHP massively though, so it can even offset GC, and 15% > 10% obv. Also I think GC allows me to Tele around more freely because even if I'm not in melee range, mobs still take 15% more damage no matter what. Any thought?

Also pure DPS from IC is far higher than FW, but maybe the ticks can compensate a bit...
Edited by novice#1750 on 10/29/2012 6:00 AM PDT
Reply Quote
Thanks for the test info
Reply Quote
Prolly a dumb question but we are using Shards correct? Even if evocation didn't cause any effect on EB due to maxed out server cast rate, it should have a full 15% effect upon shards cast rate. I'd except that to have at least a ~2% effect upon your test just from that alone. Your seeing a 0.3% effect. Margin and error as is that's massively disconcerting.


I'm disconcerted about this, too, as I feel that Evocation is necessary for people, like myself, who are skimming 2.5 AS and so we should see the biggest benefit from it in our DPS.

holyfork, could you test the below build the same way?

Thanks,
Mitsuragi#1806

http://us.battle.net/d3/en/calculator/wizard#VTYXhg!YWg!YccccZ
Reply Quote
Mad props... this info is awesome.
Reply Quote
Posts: 13,644
View profile
hard to take these multipliers as universal with only 18 apoc on single target. surely u miss attack turns especially with conflag running. Like I know I did on SB with my conglag firewalker test cuz I lost evo. I was running like a 9x multiplier on 3 elites but only 7.5 on SB. But it's nice work man cuz it gives us baselines.
Reply Quote
Posts: 7,035
View profile
If there is no nearby enemy, can the shocking aspect effect hit the same enemy, e.g. Ghom? Or is the bonus damage just from the storm armor lightning strikes? Whenever I run SNS, I can't tell what's shocking what, it's just lightning everywhere!
Edited by ChangBooster#1510 on 10/29/2012 7:35 AM PDT
Reply Quote
Can anyone with 3.0 ias try running a test with meteor + ww variant please?

Is it safe to assume sns cms does the most dmg amoung all classes?
Reply Quote
10/29/2012 05:18 AMPosted by Shandlar
Stretch Time Rune on Slow Time didn't really help me at all for DPS, likely because I didn't break the next threshold for WW at 3.0 APS and I had to spend a cast to put up Time Warp every 7 or so seconds.


Bubbles use attack turns? Didn't think they did.

You're probably right about it. I did stop mashing WW for a split second to make sure it came out. I forgot it was off the cast honestly.

10/29/2012 05:18 AMPosted by Shandlar
Prolly a dumb question but we are using Shards correct? Even if evocation didn't cause any effect on EB due to maxed out server cast rate, it should have a full 15% effect upon shards cast rate. I'd except that to have at least a ~2% effect upon your test just from that alone. Your seeing a 0.3% effect. Margin and error as is that's massively disconcerting.

I'm not sure exactly what caused this effect. The way I do SNS is to bind 1/2 to Diamond Skin and Frost Nova and hold 1/2 down the entire time. 3 is EB while my left mouse is WW and I constantly mash those. From what I can tell, this casts Frost Nova and Shards on CD automatically without having to mash those two buttons. I'm not sure if this particular method of executing SNS causes the delay or not but I think its more consistent (and saves my hands) then mashing everything. And I think its pretty clear from this description that I don't use keyboard/mouse macros to do the casting so there might be some variation from just fatigue.

And yes, I was using Shards were on the spec under test. I just referred to the skill instead of the rune because one of the tests used the Crystal Shell rune instead of the Diamond Shard rune to test the DPS loss.

I don't think I'll be able to really resolve a 2% margin of error without a huge number of tests however and if you want to get that precise, you're probably going to need to record every attempt (w/ Fraps or something). These were done more so to get general trends than being super accurate.

That said, I should probably put a disclaimer up there that says there's probably at least a few percent margin of error on all numbers, especially the ones based on one test.

EDIT: there's also a test which compares shards to Crystal Shell (which of course, does no dps).
Crystal Shell- 735k Sustained DPS (6.18x)
Shards - 817k Sustained DPS (6.86x)

Evocate was in the spec for both of these but just say for discussion's sake, it wasn't. The damage that shards does when comparing the two specs is 82k DPS. If you do 15% more of that in an ideal situation, that means Shards does 94k DPS or when added to the case with no Shards, 829k DPS. If you compare 829k DPS back to the Shards damage, 819k damage, that's around 1% damage. Now, if you go back to how long it took to kill Ghom and assume I was off by even 1 second in my timing, for a kill time of 1:25.3 or 1:27.3, you end up with a DPS of 827k or 809k respectively. Of course, this doesn't account for Storm Armor procs from the increased number of shards and the additional CM procs from more shards but trying to include that into the model probably complicates it more than its worth when you're just looking for a quick analysis/sanity check.

Evocate is likely going to always be in the noise given the method I used for testing. If you really want to try and resolve 1-2% differences, you need to be super precise and accurate on everything from start to finish (and you can't have Ghom running around like I did haha).

10/29/2012 07:27 AMPosted by Aimless
hard to take these multipliers as universal with only 18 apoc on single target. surely u miss attack turns especially with conflag running. Like I know I did on SB with my conglag firewalker test cuz I lost evo. I was running like a 9x multiplier on 3 elites but only 7.5 on SB. But it's nice work man cuz it gives us baselines.

I think this is more due to Shards cooldown and maybe Bone Chill uptime assuming the 15% wears immediately whenever the mob is unfrozen again. Single target freezing is not a true permafreeze at my gear levels, there's some stuttering as I mentioned and even movement. A better comparison might be to find one of those cooldown shrines and try Ghom again to see if that makes an impact.

Whether or not that actually makes up that much of a difference, I don't know. But I will note that I agree with you in that regard; hitting more mobs in a super concentrated area is just way more DPS. I have a feeling there's some non-linear increase in DPS when going from hitting 1 mob to hitting 3 mobs to hitting 10 mobs, all in the same space.

Edit: Added some more notes regarding Shards for Shandler.
Edited by holyfork#1768 on 10/29/2012 12:04 PM PDT
Reply Quote
This was on the to-do list.

I'll measure my DPs output with my avaliable gear and update later tonight.

3.0 APS 150k dps here.
Reply Quote
holyfork, could you test the below build the same way?

Thanks,
Mitsuragi#1806

http://us.battle.net/d3/en/calculator/wizard#VTYXhg!YWg!YccccZ

Odd... "Your post contains restricted HTML code" from quoting the calculator. The heck...?

Anyways 1:09, 1:11, 1:08 which is an average of 1:09.3 seconds = 1,019k DPS

And, I'll add it to the first post as well.
Reply Quote
I Tried with /without evocation with my little crit with follower(46%) and I got only about 3-5% more dps. Maybe wth a macro every 25ms you can get a better percentage but with finger, the difference is pretty small.
Reply Quote
I tried crystal shell and glass canon who give more dps and more survivablity than shard+evocation.
Edited by einsteinz#1738 on 10/29/2012 11:45 AM PDT
Reply Quote
10/29/2012 11:29 AMPosted by einsteinz
I tried prismatic armor and glass canon who give more dps and more survivablity than shard+evocation.

Wouldn't it be better to run Crystal Shell + Shocking Aspect then?
Reply Quote

Please report any Code of Conduct violations, including:

Threats of violence. We take these seriously and will alert the proper authorities.

Posts containing personal information about other players. This includes physical addresses, e-mail addresses, phone numbers, and inappropriate photos and/or videos.

Harassing or discriminatory language. This will not be tolerated.

Forums Code of Conduct

Report Post # written by

Reason
Explain (256 characters max)
Submit Cancel

Reported!

[Close]