Diablo® III

Plague Bats: Thoughts and Techniques

PLAGUE BATS RESEARCH

Overview:
-> initial (base cost) response on frame 8
-> initial damage frame and mana response occur on frame 10
-> one stack deals aps * 1.925% (577.5%/300) weapon damage each frame per stack
-> stacking (if you keep channeling) occurs at a floor(30/aps) frequency
-> the maximum number of stacks is 5; as soon as you achieve 5 stacks, the 6th one will overwrite the 1st one
-> one stack max duration is (300 frames / aps) or (5.0 sec / aps)
-> first stack costs 220.5 mana (initial cost) on frame 8 + channeling tooltip mana cost / 2 on the initial damage frame (frame 10)
-> each subsequent stack costs channeling tooltip mana cost / 2 on its initial damage frame


Special thanks to Gotcha and Wachati for pointing out some mistakes in the skill mechanics description

Note about the Firebats damage being increased in patch 1.0.8. The Firebats buff modifier is 77/36 = 2.138888 repeating.

This means base Firebats was changed from 180% to exactly 385% weapon damage (aps * 6.41667% per frame). Plague Bats was changed from 270% to 270 * 77/36 = 577.5% weapon damage (aps * 1.925% per frame per stack) and Cloud of Bats was changed from 234% to 234 * 77/36 = 500.5% (250.25-500.5% damage per tick).
-----
supporting research data

Plague bats damage sequence:
100-100 base weapon damage; 1.00 aps, 2400 int; 31% crit chance; +136% crit damage
tooltip damage = 100 * 25 * 5.78 = 14450
expected damage per frame = tooltip / (hidden base duration * 60) = 14450 / 300 = 48.167
crit damage per frame = 48.167 * 2.36 = 113.67412

Note that I believe the tooltip damage is a rounded value and the actual damage is slightly lower because the buff multiplier in patch 1.0.8 is a weird value (something like 2.138-2.139). I don´t understand how the devs came to the exact buff value and it´s not something that I can answer with a test due to damage being applied frame by frame and the game displays only whole numbers on health bars.

In the following sequence, I began the frame number counting from the initial damage frame (f1) which occured on frame 10 after the skill was activated.

monster health<->frame number<->damage per frame
230,880 f27
230,832 f28 -48
230,784 f29 -48
230,736 f30 -48
230,639 f31 -97 (second stack initial frame)
229,773 f40 -866/9 = 96.22
228,811 f50 -962/10 = 96.2
227,944 f59 -867/9 = 96.33
227,848 f60 -96
227,704 f61 -144 (third stack initial frame)
227,559 f62 -145
227,415 f63 -144
227,271 f64 -144
224,961 f80 -2310/16 = 144.375
223,806 f88 -1155/8 = 144.375
223,662 f89 -144
223,517 f90 -145
223,324 f91 -193 (fourth stack initial frame)
221,592 f100 -1732/9 = 192.44
219,667 f110 -1925/10 = 192.5
218,127 f118 -1540/8 = 192.5
x f119 game skipped frame
217,742 f120 -385/2 = 192.5
217,550 f121 -192
217,436 f122 -114 (crit frame)
217,130 f123 -306 (5 stacks; fifth is a crit)
216,824 f124 -306
216,518 f125 -306
214,682 f130
211,621 f140
209,478 f147
209,172 f148 -306
208,867 f149 -305
208,561 f150 -306
208,188 f151 -373 (6th stack crit and 1st stack non-crit ended)
207,817 f152 -371
207,455 f153 -362
207,074 f154 -381
206,702 f155 -372
206,331 f156 -371
204,844 f160 -1487/4 = 371.75
199,271 f175 -5573/15 = 371.53
198,900 f176 -371
198,528 f177 -372
198,157 f178 -371
197,785 f179 -372
197,414 f180 -371
196,977 f181 -437 (7th stack crit and 2nd stack non-crit ended)
196,540 f182 -437
188,675 f200 -7865/18 = 436.94
185,179 f208 -3496/8 = 437
184,742 f209 -437
184,305 f210 -437
x f211
183,431 f212 -874/2 = 437 (8th stack non-crit and 3rd stack non-crit ended)
182,994 f213 -437
182,557 f214 -437
etc.

Conclusions (at 1.00 aps):
-> initial (base cost) response on frame 8
-> initial damage frame and channeling mana response on frame 10
-> if you continue channeling the next stack´s initial damage frame and mana response occur 30 frames after the previous one
-> one stack deals roughly 1/300 weapon damage each frame
-> one stack duration is 300 frames (5.0 sec)

I´ll check what happens with more aps but the skill mechanics seems pretty straightforward if you take the time and analyze what´s happening frame by frame.

Will update this post with higher aps results/conclusions.

Here´s the update:
100-100 base weapon damage; 1.50 aps; 2400 int; 21.5% crit chance; +80% crit damage
expected damage per frame = 48.167 (same as before)
expected crit damage per frame = 113.67412 (same as before)

250,737
250,665 f1 -72 (hmm this is non-crit * aps)
250,592 f2 -73
249,293 f20 -1299
249,149 f21 -144 (2nd stack non-crit)
249,005 f23 -144
248,860 f24 -145
247,994 f29 -866
247,850 f30 -144
247,705 f31 -145
247,561 f32 -144
246,550 f39 -1011
246,406 f40 -144
246,189 f41 -217 (3rd stack non-crit)
242,291 f59 -3898
242,074 f60 -217
241,786 f61 -288 (4th stack non-crit)
235,939 f80
235,578 f81 -361 (5th stack non-crit)
235,217 f82 -361
229,442 f98
229,081 f99
229,817 f100 (lol this happens sometimes)
229,398 f101 (6th stack crit 1st stack non-crit ended around here)
227,824 f102
227,406 f103 -418
226,987 f104 -419

Well, it seems the stacking frequency follows the floor(30/aps) formula and its duration is adjusted to 5x that result.

APS has the following effects:

1. increases the stacking frequency = floor(30/aps)
2. linearly increases damage per frame per stack = aps * (tooltip/300)
Edited by Nubtro#2147 on 7/21/2013 10:07 AM PDT
Reply Quote
Posts: 2,184
View profile
07/20/2013 05:35 AMPosted by Nubtro
Plague Bats damage doesn´t scale with APS, which only affects the stacking frequency. But due to each stack´s duration being adjusted to 5*(floor(30/aps)), you won´t deal more damage per second with more aps.while channeling the skill.


So AS doesn't increase the start-up stacking either?

If true, then AS would do exactly zip for PB.
Reply Quote
Plague Bats damage doesn´t scale with APS, which only affects the stacking frequency. But due to each stack´s duration being adjusted to 5*(floor(30/aps)), you won´t deal more damage per second with more aps.while channeling the skill.


So AS doesn't increase the start-up stacking either?

If true, then AS would do exactly zip for PB.


I will not believe this. I notice a difference when I drop to different Attack speed breakpoints (to get the initial ramp up) yet I do not use PB as a main skill only as a ramp up for the fast DOT before switching to bears. With that said I can also get my DPS and CHD, etc. close (with a skorn slow attack speed and a 1h much faster attack speed) and there is a huge difference in the startup time to ramp up PB before I switch to bears with a slow skorn Attack speed vs a much faster breakpoint 1h setup. Thats why I do not run a skorn in general in my build due to the difference.
Edited by Gotcha#1167 on 7/20/2013 7:34 AM PDT
Reply Quote
Posts: 2,184
View profile
07/20/2013 07:30 AMPosted by Gotcha
I will not believe this. I notice a difference when I drop to different Attack speed


I think the same way. Let's see if Nubtro answers.
Reply Quote
I´ve deleted the last two sentences in my post. Let´s recap the whole thing, because APS has 3 effects.

1. increases the stacking frequency: floor(30/aps)
2. linearly increases damage per frame: aps * (tooltip/300)
3. decreases stack duration: 5 * floor(30/aps)

It´s kind of difficult to make the conclusion considering all 3 aspect lol my brain´s not working properly.

Some help would be nice here haha.

At 1.00 aps, you deal 1/300 of tooltip damage each frame for 150 frames per stack. The fifth stack begins on frame 121, at which point you deal 5 stacks * 1/300 = 1/60 tooltip damage each frame = tooltip damage per second.

At 1.50 aps, you deal 1.5 * 1/300 of tooltip damage each frame for 125 frames per stack. The fifth stack begins on frame 101, at which point you deal 5 stacks * 1.5 * 1/300 = 7.5/300 tooltip damage each frame = 1.5 tooltip damage.

I guess the damage does scale with aps if I didn´t make any mistakes ;)
Edited by Nubtro#2147 on 7/20/2013 8:22 AM PDT
Reply Quote
hello everyone... good info on PB... after reading everyones posts and checking variations of builds people are running, I wanted to share my build. I have no problem running MP10 solo. I can hit 200k but I rather have more life for a bit less damage. I started running this build on MP10 with 155k damage

I found that Toads compliments PB well. It allows to attack over barriers, its cheap on mana when you go OOM with Bats and has good proc rate so you can still leech enough with it to stay alive. The pets help create the barrier between elites long enough to reposition or to move around on difficult long elite battles.
The two passives (Medicine/Jungle) also help tremendously and I recommend them especially if you are not hurting for damage or mana. You can switch out Jungle for Grave if you need mana or want faster CD on skills(which u don't really need for this build).

I also switch the dogs or Garg for Jinx or even voodoo for damage output or in parties. The two passives like I said make all the difference in MP10 with this build especially if you are doing less than 150k damage as it will give you a total of 35% damage reduction when casting bats on mobs or elites. Back that with at least 4k armor and at least 50k plus life and you will find it fairly easy to run MP10. I am at 70k life(was at 80k) and while it was a bit overkill, it allows me to run different skills(voodoo) for more damage since having more life not only lets you stay alive and keeps your pets alive on long elite fights.... more life also increases pet life pool without the need the main pet passive fierce loyalty.

last note... I tried COB build and dont like it... it is aweful against certain elites(knockback) and locust swarm sux as a back up skill. COB is great if you are doing massive damage(over 240k) you just want to kill mobs in open areas... but when it comes to certain elites or especially elites without a lot of mobs to get mana back your practically hosed. My 2 cents and experience...
Edited by AztroTrain#1796 on 7/20/2013 12:49 PM PDT
Reply Quote
Posts: 2,184
View profile
At 1.50 aps, you deal 1.5 * 1/300 of tooltip damage each frame for 125 frames per stack. The fifth stack begins on frame 101, at which point you deal 5 stacks * 1.5 * 1/300 = 7.5/300 tooltip damage each frame = 1.5 tooltip damage.

I guess the damage does scale with aps if I didn´t make any mistakes ;)


If you are correct in your conclusion this time, if the damage stacks in the formula AS X base damage, then increasing AS should linearly increase damage output.

This would agree with what some of us are seeing.
Reply Quote
Posts: 2,184
View profile
07/20/2013 08:42 AMPosted by AztroTrain
hello everyone... good info on POB... after reading everyones posts and checking variations of builds people are running, I wanted to share my build


First the skill is just PB = plague bats.

As far as your build goes, you have way too much AS to be using PB all the time. While you have a ton of mit from dogs, BM, and JF, your damage output is poor becasue you can't maintain PB.

I would think you would be much more effective subbing in BR, or SA for GF, even then you would still be running OoM on PB, but if that's your play style...
Reply Quote
07/20/2013 09:33 AMPosted by skywalkerfx
hello everyone... good info on POB... after reading everyones posts and checking variations of builds people are running, I wanted to share my build


First the skill is just PB = plague bats.

As far as your build goes, you have way too much AS to be using PB all the time. While you have a ton of mit from dogs, BM, and JF, your damage output is poor becasue you can't maintain PB.

I would think you would be much more effective subbing in BR, or SA for GF, even then you would still be running OoM on PB, but if that's your play style...


BR why? staying alive is not a problem... and dropping GF only would add to the mana problem. I rarely go Oom... unless i have to recast it 3 times due to knockback elites... Because PB is the only skill that I am using other than when i need to recast SH or SW.

I will give SA a try... rather lose some mitigation from from BM or JF. I didn't know 205k damage was considered poor... not a fan of skorn but like you say everyone's playstyle is different.
Reply Quote
Some help would be nice here haha.

At 1.00 aps, you deal 1/300 of tooltip damage each frame for 150 frames per stack. The fifth stack begins on frame 121, at which point you deal 5 stacks * 1/300 = 1/60 tooltip damage each frame = tooltip damage per second.



Nubtro something still does not make sense. Can you test how a sinlge stack works and see if this matches my findings?

What I found is actually this
At 1.00 aps, you deal 1/50 of tooltip damage each 6 frames for a total of 300 frames per stack.

Here are my test results for a single stack - burst quickly on a single target

Setup
=====
weapon damage = 369
aps 1.0
Intel damage boost - 11.39 + 1 = 12.39
cc = 5% (min)
cd = 50% (min)

Expected damage = 369*12.39*5.78 = 26425

Damage from a single stack = 26400
Damage from a single stack is calculated as 50 ticks of 6 Frames each, each tick does 528 damage.

Single Stack Damage Numbers (every 30 Frames)
Damage / # of Ticks * 528 /
2640 = 5 * 528
2640 = 5 * 528
3168 = 6 * 528
2640 = 5 * 528
2640 = 5 * 528
3168 = 6 * 528
2640 = 5 * 528
3168 = 6 * 528
2640 = 5 * 528
1056 = 2 * 528

Total "ticks" = 50, each one lasting about 6 frames.

To sum -
1) A single stack takes 300 Frames and not 150 Frames.
2) A single stack does the full tooltip damage.

Now, how does that match your findings -
Looking at your numbers I think there is a limit of 5 on how many stacks can stack :)
Essentially during "ramp up" the effective dps is twice the tooltip and once you reach the limit of 5 the steady-state damage is the tooltip damage.

Note:
This also supports Gotcha's findings, during ramps up the effective dps is very high (twice the tooltip) and after max damage is reached it is better to switch to ZB.
Edited by Wachati#1647 on 7/20/2013 1:10 PM PDT
Reply Quote
Thanks for reminding me I forgot to test the most important thing, to double check one stack duration.

Yes it seems one cast is 300 frames long. After looking at the whole sequence of just one stack again it´s exactly as you´re saying. One stack takes 300 frames and deals 1/300 of tooltip damage each frame. The skill can stack up to 5 times and my previous tests suggest that the first stack is overwritten by the 6th.

Updating the post. Thanks again.

EDIT: Added the following information.

Overview:
-> initial (base cost) response on frame 8
-> initial damage frame and mana response occur on frame 10
-> one stack deals (aps * 1/300) tooltip damage each frame
-> one stack damage duration is 300 frames (5.0 seconds)
-> stacking (if you keep channeling) occurs at a floor(30/aps) frequency
-> the maximum number of stacks is 5; as soon as you achieve 5 stacks, the 6th one will overwrite the first one
-> first stack costs 220.5 mana (initial cost) + channeling tooltip mana cost / 2 on the initial damage frame
-> each subsequent stack costs channeling tooltip mana cost / 2 on its initial damage frame


Special thanks to Gotcha and Wachati for pointing out some mistakes in the skill mechanics description


Note that I´m not sure about the initial cost which the tooltip says is 220 but I always lose 221 mana. The cost before 1.0.8 was 98 IIRC and 220.5 would make sense (*2.25). It´s kind of an assumption but I have no clue how to test its exact value.
Edited by Nubtro#2147 on 7/20/2013 2:10 PM PDT
Reply Quote
Posts: 2,184
View profile
So in conclusion AS ramps up the 5 stacks to max damage faster, but once you hit 5 stacks, damage is constant.

This is what most people assumed was true.
Reply Quote
Overview:
-> one stack damage duration is 300 frames (5.0 seconds)


One correction Nubtro, I think this line should say:

-> one stack damage duration is 300/aps frames (5.0/aps seconds)
Reply Quote
So in conclusion AS ramps up the 5 stacks to max damage faster, but once you hit 5 stacks, damage is constant.

This is what most people assumed was true.


Yes, but most importantly, the damage does scale with AS.

AS effects the following:
1) DoT duration of a single stack is reduced (5.0/aps)
2) DoT stack more quickly (one stack every 0.5/aps seconds), max is still 5
3) Overall sustained damage scales with AS (tooltip*aps)
Edited by Wachati#1647 on 7/20/2013 4:47 PM PDT
Reply Quote
I think a very interesting conclusion from Nubtro's testing is that during the ramp-up period on a single target, PB effective DPS is double the tooltip damage (Higher than CoB!)

578%*(aps/0.5) = 1156%*aps

Edit: ramp-up period duration is (2.5/aps) seconds.
Edited by Wachati#1647 on 7/20/2013 4:33 PM PDT
Reply Quote
You all are the math guys.

I just know that PB for the ramp up folowed by bears is wooooooooooooooooooot!

Its great to have you all cranking out the numbers.

Long live us WD'rs no matter what skills you use.

:P
Reply Quote
Overview:
-> one stack damage duration is 300 frames (5.0 seconds)

One correction Nubtro, I think this line should say:
-> one stack damage duration is 300/aps frames (5.0/aps seconds)


You´re right but maybe I should also add a note that one stack duration while channeling won´t be higher than 5 * floor(30/aps) frames because after that the 6th one overwrites the first one which never gets its full duration etc. This applies for as long as you channel and overwrite stacks. The moment you stop channeling, the base max duration is applied.

For example at 2 aps the duration of one stack should be 2.5 seconds but it only takes 75 frames to get 5 stacks: f1(#1)->f16(#2)->f31(#3)->f46(#4)->f61(#5)->f76(#6 overwrites #1).
Reply Quote
I know what you are saying Nubtro but I think that explanation will be just as confusing. Also, what you saying is only valid if you sit there and attack a single target, but most PB users "spray" the bats across multiple targets, if a single target only gets every other stack then it will get the full duration.
Edited by Wachati#1647 on 7/20/2013 5:46 PM PDT
Reply Quote
Posts: 2,184
View profile
I think you guys have to bake this down some more.

If I have a 1.0 APS weapon, and a 2.0 APS weapon with the exact same DPS:

1. The 2.0 APS weapon hits max damage faster.

2. The 2.0 APS weapon after hitting max damage, makes stacks quicker so it delivers more damage over time.

Are these the basic conclusions?
Reply Quote
Posts: 4,303
View profile
I think you guys have to bake this down some more.

If I have a 1.0 APS weapon, and a 2.0 APS weapon with the exact same DPS:

1. The 2.0 APS weapon hits max damage faster.

2. The 2.0 APS weapon after hitting max damage, makes stacks quicker so it delivers more damage over time.

Are these the basic conclusions?


In addition to that, I would want to know if 2 APS is exactly twice the damage in the same time interval. If not, then your sheet dps vs effective dps is going to get skewed the more attack speed you have.
Edited by GunnersDream#1109 on 7/20/2013 6:16 PM PDT
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]