Diablo® III

[SCIENCE!] Disintegrate and APS

IMPORTANT: These results have been called into serious question due to flaws in my methodology (but have raised additional questions about why CM is proccing way more than it should). Trust no-one pending further investigation.

-----

Some dude:

http://us.battle.net/d3/en/forum/topic/7349425050

...just wanted to know if APS influenced how often channeled spells like Disintegrate proc CM. He was told that it did not, because Disintegrate ticks twice per second no matter what your APS is, and that stacking ASI was just a good way to drain your AP really fast. I was pretty sure that wasn't right, so I checked it out myself. Turns out I was totally wrong.

-----

Methods

I popped Archon, cancelled it, and then spammed Disintegrate on a single MP10 zombie until Archon's cooldown was over. I measured how many seconds of constant channeling it took to refresh Archon. The cooldown-relevant factors were:
  • Archon: No Rune, the cooldown I was resetting.
  • Critical Mass, the actual proc I was triggering.
  • 22.5% CC, to proc CM.

For Disintegrate, I used both the Chaos Nexus and Convergence runes, chosen because they have the same proc coefficient ( http://us.battle.net/d3/en/forum/topic/7004036888 ), so any difference between them is purely due to the crits from CN's secondary beams.

I also used two different attack speed setups, one with an ASI dagger (total character APS = 1.67) and one with a 2H mace (total character APS = 0.98).

I used Energy Armor: Energy Tap, and Diamond Skin: Prism to keep my AP from running dry on the fast tests. This should have had no effect on cooldowns whatsoever.

Procs/second were calculated by taking 120 (the total time Archon should take to reset its cooldown without CM) minus T (the actual time it took to reset Archon's cooldown), and dividing the difference by T, i.e.:

Procs/sec = (120 - T) / T

Comparison of the rate at which secondary beams are fired by Chaos Nexus for slow vs. fast weapons was eyeballed by counting the number of beams fired over a 5s window in a frame-by-frame viewing of the video.

-----

Results
  • Chaos Nexus, fast weapon: 88s to refresh Archon. 0.364 procs/sec. 18 secondary beams fired between 42s and 47s in the video.
    Video: http://youtu.be/67MDetFyVN0
    .
  • Chaos Nexus, slow weapon: 86s to refresh Archon. 0.395 procs/sec. 11 secondary beams fired between 14s and 29s in the video.
    Video: http://youtu.be/a205pR-qEEU
    .
  • Convergence, fast weapon: 88s to refresh Archon. 0.364 procs/sec.
    Video: http://youtu.be/7CkFqaLu4Xc
    .
  • Convergence, slow weapon: 87s to refresh Archon. 0.379 procs/sec.
    Video: http://youtu.be/cbu1VVWXvWc

-----

Conclusions
  • Attack speed appears to have no effect on the rate at which Disintegrate procs CM.
  • Attack speed does appear to cause CN to proc secondary beams more quickly.
  • ...but it doesn't show up in cooldown reduction from CM procs; CN shows no advantage over Convergence, and fast CN shows no advantage over slow CN.
  • The fact that these beams don't seem to speed up the CM procs may indicate that they have a very low or zero proc coefficient, or that my testing just wasn't rigorous enough to notice the difference (if I really wanted to do this right, I'd average over a whole bunch of trials, but I'm too lazy).
  • The encoding of my videos was crappy enough to make CN beam counting a little difficult, but seriously, try it yourself in-game with a fast vs. slow weapon setup; it's night and day. Fast weapons definitely proc secondary beams faster.

tl;dr Using fast weapons with Disintegrate sucks. It doesn't help you proc stuff more quickly, just burns your AP really fast. Go get yourself a Skorn.
Edited by AS00#1295 on 12/7/2012 3:35 PM PST
Reply Quote
So you didn't believe what you were told to be the truth and went off and did a bunch of experiments to, lo and behold, arrive at the same truth?
Reply Quote
12/07/2012 01:54 PMPosted by JawnV
So you didn't believe what you were told to be the truth and went off and did a bunch of experiments to, lo and behold, arrive at the same truth?

Yep, pretty much. Thought I remembered somebody else saying that channeled spells had a hidden tic rate, based on APS, and that the 2/sec numbers were summaries that added up the tics since the last number popped up. I think that's all true, except for the "based on APS" part; channeled skills just have a native tic rate that doesn't care about APS.
Reply Quote
I appreciate the effort and am sorry for the snark in my first reply. I used Chaos Nexus back in the day and never made the connection to IAS that you noted here. Good work :)
Reply Quote
12/07/2012 02:00 PMPosted by JawnV
I appreciate the effort and am sorry for the snark in my first reply. I used Chaos Nexus back in the day and never made the connection to IAS that you noted here. Good work :)

Snark = entirely justified ;) .
Reply Quote
Some dude:


that's me :)

This is great work. Thanks for the great info!
Reply Quote
Is archon beam tick rate works the same as disintegrate?
Reply Quote
12/07/2012 02:23 PMPosted by Aurex
Is archon beam tick rate works the same as disintegrate?

Well, the Archon abilities have their own proc coefficients (0.10 for the beam, according to ZzEzZ's list), but as far as how they tic, I'd guess that all channeled abilities work the same. I could be wrong, though; I only tested Disintegrate. You wanna check it out?
Reply Quote
Something you may want to look into is how Disintegrate's AP cost is determined when using slow 2handers and adding cost reduction.

I did a bunch of looking into this back before I had a trium, and the results were jsut plain odd.

http://www.diablofans.com/topic/69583-disintegrates-costodd-behavior/

I did many more tests than what are outlined there, but as your attack speed dips below 1aps along with a -5 cost reduction you can spam disintegrate for infinity time...even if your expect expenditure should be more than your passive regen. Also these tests were done back when the spell still has a listed cost of 20 AP.
Reply Quote
There's probably a better way to see if IAS changes procs using LoH with various attack speeds. If the proc rate changes with attack speed, the LoH returns will increase with attack speed, but you'll want to use a very wide variety of attack speeds, I recommend going from 1 to 2.51+. Using LoH instead of Archon CD negates the RNG of crits. Also, you didn't mention what your crit % was, which plays a huge role in the RNG of the experiment.

If I'm not mistaken, some people in europe did just that and the results are posted here: http://eu.battle.net/d3/en/forum/topic/5208511084

They concluded it has a tic rate of about 3*APS tics per second.
Edited by Loroese#1415 on 12/7/2012 2:55 PM PST
Reply Quote
Something you may want to look into is how Disintegrate's AP cost is determined when using slow 2handers and adding cost reduction.

I did a bunch of looking into this back before I had a trium, and the results were jsut plain odd.

http://www.diablofans.com/topic/69583-disintegrates-costodd-behavior/

I did many more tests than what are outlined there, but as your attack speed dips below 1aps along with a -5 cost reduction you can spam disintegrate for infinity time...even if your expect expenditure should be more than your passive regen. Also these tests were done back when the spell still has a listed cost of 20 AP.

Huh, that is indeed totally weird. Testing just now, I find that even with Disintegrate's AP cost at 10 and no additional regen, I hit perma-channel well above 1 APS (where it should be), and at my normal APS (1.31), I can channel about 25% longer than I should be able to (assuming your 6 AP min still holds). Weird.

There's probably a better way to see if IAS changes procs using LoH with various attack speeds. If the proc rate changes with attack speed, the LoH returns will increase with attack speed, but you'll want to use a very wide variety of attack speeds, I recommend going from 1 to 2.51+. Using LoH instead of Archon CD negates the RNG of crits. Also, you didn't mention what your crit % was, which plays a huge role in the RNG of the experiment.

If I'm not mistaken, some people in europe did just that and the results are posted here: http://us.battle.net/d3/en/forum/topic/7350725148

They concluded it has a tic rate of about 3*APS tics per second.

You relinked to this thread; can you link to the real one? I'd like to read it.

*edit* And my CC was 22.5% for this test (took off my source and dismissed Scoundrel to make the test cleaner).
Edited by AS00#1295 on 12/7/2012 2:52 PM PST
Reply Quote
What build are people using with Disintegrate? I'm obviously CM now, but even when I was using Dis/Archon I switched the non-Archon portions over to CMlike skills to get back into Archon faster.
Reply Quote
12/07/2012 02:49 PMPosted by JawnV
What build are people using with Disintegrate? I'm obviously CM now, but even when I was using Dis/Archon I switched the non-Archon portions over to CMlike skills to get back into Archon faster.

I like Dis/Archon.

*shrug*

Besides that, there are some passable Disintegrate-based builds for Ubers if you don't have CM gear. Though given how cheap an Uber-capable CM set is, they're probably not worth bothering with.
Reply Quote
Something you may want to look into is how Disintegrate's AP cost is determined when using slow 2handers and adding cost reduction.

I did a bunch of looking into this back before I had a trium, and the results were jsut plain odd.

http://www.diablofans.com/topic/69583-disintegrates-costodd-behavior/

I did many more tests than what are outlined there, but as your attack speed dips below 1aps along with a -5 cost reduction you can spam disintegrate for infinity time...even if your expect expenditure should be more than your passive regen. Also these tests were done back when the spell still has a listed cost of 20 AP.

Huh, that is indeed totally weird. Testing just now, I find that even with Disintegrate's AP cost at 10 and no additional regen, I hit perma-channel well above 1 APS (where it should be), and at my normal APS (1.31), I can channel about 25% longer than I should be able to (assuming your 6 AP min still holds). Weird.

There's probably a better way to see if IAS changes procs using LoH with various attack speeds. If the proc rate changes with attack speed, the LoH returns will increase with attack speed, but you'll want to use a very wide variety of attack speeds, I recommend going from 1 to 2.51+. Using LoH instead of Archon CD negates the RNG of crits. Also, you didn't mention what your crit % was, which plays a huge role in the RNG of the experiment.

If I'm not mistaken, some people in europe did just that and the results are posted here: http://us.battle.net/d3/en/forum/topic/7350725148

They concluded it has a tic rate of about 3*APS tics per second.

You relinked to this thread; can you link to the real one? I'd like to read it.

*edit* And my CC was 22.5% for this test (took off my source and dismissed Scoundrel to make the test cleaner).


Ok, relinked. No idea why it linked to this thread. Also, 22% crit is VERY LOW to be testing something that relies on critting, so no wonder you didn't see much of a difference. I highly recommend using the LoH approach since it just requires having 1 piece of LoH gear (which can be a storm crow) and some IAS gear. You should notice a difference in the LoH returns at 2.5 APS vs 1.0 APS.
Reply Quote
That's interesting. If that's true, bascially it means CM procs are NOT the same as LoH procs.
For LoH, you can visually see the difference between slow and fast weapons.

So if CM is not tied to the LoH attack speed scaling, this raises several new questions. What is CM tied to? What about APoC? What about Shocking Aspect?
Reply Quote
That's interesting. If that's true, bascially it means CM procs are NOT the same as LoH procs.
For LoH, you can visually see the difference between slow and fast weapons.

So if CM is not tied to the LoH attack speed scaling, this raises several new questions. What is CM tied to? What about APoC? What about Shocking Aspect?


I know shocking aspect has a very odd interaction with WW that is not as simple as the LoH breakpoints but I believe CM and APoC are still tied to the LoH returns. This believe comes from the simulator I made being fairly accurate for most skills besides SA. It seems to work very well for WW, which I expect to exhibit similar behavior to Disintegrate.

Also, I have a lot of doubts about the results in the OP because of the very low crit% used. If Disintegrate tics 3*APS per second, at 1 APS that's 270 tics over 90s, which is about 61 crits, which is about 7 CM procs. At 1.68 APS that's 454 tics over 90s for about 11CM procs. If only one run was done, I'm not surprised that there's no noticable difference betwen the two cases. If several runs were performed, there might be more of a difference, but I think it would need to be tested with a high crit (50%+) to reduce the varience.
Edited by Loroese#1415 on 12/7/2012 3:08 PM PST
Reply Quote
12/07/2012 02:58 PMPosted by Loroese
Ok, relinked. No idea why it linked to this thread. Also, 22% crit is VERY LOW to be testing something that relies on critting, so no wonder you didn't see much of a difference. I highly recommend using the LoH approach since it just requires having 1 piece of LoH gear (which can be a storm crow) and some IAS gear. You should notice a difference in the LoH returns at 2.5 APS vs 1.0 APS.

Hmmm...given the total refresh times I was seeing, I was critting often enough to be generating 30-35 procs over the course of the Archon timer. That seems like it should be a large enough sample that a difference of about 70% should have shown up very reliably. But I can't think of the appropriate statistical test to demonstrate that off the top of my head...

...I'll see what I can do with the LoH method.

*edit* If the post above this one is correct, then something else fishy is going on, which is that I'm getting WAY more procs than I should be. For Archon to reset in 90s, you have to get 30 procs over those 90s.
Edited by AS00#1295 on 12/7/2012 3:10 PM PST
Reply Quote
12/07/2012 03:09 PMPosted by AS00
Ok, relinked. No idea why it linked to this thread. Also, 22% crit is VERY LOW to be testing something that relies on critting, so no wonder you didn't see much of a difference. I highly recommend using the LoH approach since it just requires having 1 piece of LoH gear (which can be a storm crow) and some IAS gear. You should notice a difference in the LoH returns at 2.5 APS vs 1.0 APS.

Hmmm...given the total refresh times I was seeing, I was critting often enough to be generating 30-35 procs over the course of the Archon timer. That seems like it should be a large enough sample that a difference of about 70% should have shown up very reliably. But I can't think of the appropriate statistical test to demonstrate that off the top of my head...

...I'll see what I can do with the LoH method.

*edit* If the post above this one is correct, then something else fishy is going on, which is that I'm getting WAY more procs than I should be. For Archon to reset in 90s, you have to get 30 procs over those 90s.


Are you using Evocation Passive?
Reply Quote
12/07/2012 03:13 PMPosted by Loroese

Hmmm...given the total refresh times I was seeing, I was critting often enough to be generating 30-35 procs over the course of the Archon timer. That seems like it should be a large enough sample that a difference of about 70% should have shown up very reliably. But I can't think of the appropriate statistical test to demonstrate that off the top of my head...

...I'll see what I can do with the LoH method.

*edit* If the post above this one is correct, then something else fishy is going on, which is that I'm getting WAY more procs than I should be. For Archon to reset in 90s, you have to get 30 procs over those 90s.


Are you using Evocation Passive?

Nope. Blur, Astral Presence, and CM.

Just tested via LoH; LoH procs are definitely proportional to APS. Also, CN beams definitely don't have a proc coefficient of 0; they are capable of proccing both LoH and CM.

Gonna try to get a better test of CM procs set up with higher CC and a larger APS difference.
Reply Quote
CN beams should also be at a proc coef of 1/9.

There's definitely something strange going on. I can understand why the difference between the two setups is not very high, but as you said, the total value is more than three times of what it should be.

PS: For low attack speed, search for grey 2handers. They can have -aps. Don't know who sells this stuff, but the EU AH always has like 3 pages of them. I was able to get a nice 0.63 APS weapon :)
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)

Reported!

[Close]