The Gearbox Software Forums Go to Gearboxity.comGo to Gearbox Software's Web site

Go Back   The Gearbox Software Forums > Gearbox Games > Borderlands > GearChat

Reply
 
Thread Tools Display Modes
Old May 1st, 2010, 05:20 PM   #1
Scottes
Elemental Scientist
 
Scottes's Avatar
 
Join Date: Jan 2010
Posts: 4,534
Default Some Tests Of Elemental Repeaters

This is the second in a series of tests of elemental weapons, this time with repeaters. Once again ampersand made a set of repeaters for me - thanks!

The first part concerned revolvers, and can be found here:
Some Tests Of Elemental Revolvers


The Repeaters

I had originally planned on testing 4 repeaters from x1 to x4, and all four Effect Chance levels - Base, High Elemental Effect Chance, Higher, and Very High. But I am finding that the testing is quite time-consuming, and the nuances aren't as important as the extremes. Besides, the information from two weapons is close to overwhelming...


First up are these two, an x1 and an x4, both with base chance to proc.



The x1 has 32 Pool_Tech and the x4 has 68 Pool_Tech.


Next up are a couple of x4 repeaters with varying chances to proc:



Both of these have a Pool_Tech of 60.

It's interesting to note that the first x4, the one with a base chance of elemental effect, has a larger Pool_Tech than the repeaters with more chance to proc..


Proc Cost - Initial Surprise

I then fired a few clips of the base chance x4 to check the proc cost. I recorded this in Fraps and replayed the movie frame-by-frame using VirtualDub. The results seem a little weird to me...

The gun didn't proc for a few rounds, but when it did the tech pool went from 68.000 (full) to 64.5679, showing a proc cost of 3.4321. I was expecting a cost of a whole number, not a fractional one. I automatically assumed that the FPS had dipped for a bit, and the pool had regenerated a fair amount between captured frames. So I checked the next few frames, noting the Pool_Tech size and the frame time stamp:

68.0000 - 0.733s - Start
64.5679 - 0.767s - Proc
64.6764 - 0.800s
64.8837 - 0.833s
64.9778 - 0.867s

From the time stamps we can see that Fraps is recording at 30 frames per second, just like it should. Checking the Fraps FPS display, I am getting 62 frames per second, so Fraps is capturing every other frame, approximately.

Considering the second and last frames from the above list, the tech pool regenerated 0.4099 in 0.1 seconds, so it's regenerating damn close to the expected 4.0000 per second.

Conclusion: Proc costs aren't always whole numbers - at least for repeaters.


Proc Cost - Continued

I then checked the costs of the next ten procs (rounded to tenths):
3.8 - 3.9 - 3.9 - 5.1 - 4.9 - 4.9 - 4.9 - 3.8 - 4.9 - 4.9 = Average 4.5 points per proc

I tried again with the x1, recording the costs of 10 procs:
3.3 - 3.9 - 3.9 - 3.4 - 3.9 - 3.9 - 3.9 - 3.9 - 3.9 - 3.9 = Average 3.8 points per proc


Proc Likelihood - Full Auto
The first two repeaters have the base chance to proc - that is, they do not have a High Elemental Effect Chance, or Higher, or Very High. The next two have increased chances to proc.

I fired these non-stop, just holding the trigger down counting the procs. The guns initially started with their Pool_Tech at maximum.

For the x1 Base chance to proc 10 times, it took 31 bullets over 9.5 seconds*.
For the x4 Base chance to proc 10 times, it took 19 bullets 4.0 seconds.
For the x4 High chance to proc 10 times, it took 25 bullets 5.9 seconds.
For the x4 Very High chance to proc 10 times, it took 14 bullets over 3.0 seconds.

* The x1 fired more than 1 clip, so this includes a reload time of 0.55 seconds.


OK, pause for a second here for something very interesting

With the base chance x4 above, and it's Pool_Tech of 68, it took 19 bullets to proc 10 times.
With the High chance x4 just now, and it's Pool_Tech of 60, it took 25 bullets to proc 10 times.

Personally, I do not think that the difference in Pool_Tech is the reason for the discrepancy. I think the discrepancy is caused by a small test sample of 10 procs.

Chances are that 100 procs would be more consistent, but this is probably not possible since the tech pool will drain below a proc'able value before 100 procs are achieved.

Also, I am not going to spend many more hours recording and analyzing a 15 GB Fraps movie. Sorry.


Proc Likelihood - Max Pool_Tech

Next I tested them by firing a single bullet at a time, waiting for the tech pool to reach maximum before I fired.

For the x1 Base chance to proc 20 times, it took 62 bullets.
For the x4 Base chance to proc 20 times, it took 30 bullets.
For the x4 High chance to proc 20 times, it took 28 bullets.
For the x4 Very High chance to proc 20 times, it took 41 bullets.

41???? WTF? The High chance repeater took only 28. So I ran that one again:

For the x4 Very High chance to proc 20 times, it took 35 bullets. STILL more than the High chance. I don't know why, but I'm going to blame it on Borderlands randomness and/or BL's tendencies towards ambiguous description lines.


It is interesting to note that the last two tests showed a margin of error of approximately 20%. That's not very good. Don't risk your life on these tests...


Proc Size

I went through the movie of the base chance x4 repeater and checked the sizes of all the procs. I saw two different sizes, which I will call x1 and x2 procs.



For reference, here are the proc sizes from my revolver test.



As you can see, the repeater x2 proc is considerably smaller than the revolver's x2 proc, but I have no choice but to call them "repeater x1 and x2 procs."


A couple of interesting things to note is that the x4 repeater produced many x1 procs, but they were far fewer than the number of x2 procs. Also, these repeater x1 procs did not drain the tech pool, which is also how they worked on the revolvers.


Proc Size - Firehawk

In the thread Elements, Multipliers, and YOU, instant karma says that the Firehawk procs only at x6. This thread also mentions a report from Majikal_ninja saying that Hellfires have a long DoT effect. I was curious is this also might be true for Firehawks. I wanted to test both by firing a bunch of bullets, waiting between procs to make sure the tech pool was at maximum.

For reference, the Firehawk I used is this one:



It has a Pool_Tech size of 92, and took 27 shots to proc 20 times when ensuring the tech pool was full before firing.

This is one example of a Firehawk's proc:



All of the procs were the same size, and they all come close to matching the size of the revolver's x2 proc. The Firehawk's proc is considerably larger than a typical repeater's x2 proc. Is it an x6 proc? I'm not sure, but I won't argue that it isn't x6.


Considering the theory that a Firehawk may cause more DoT, I checked the movie to see how long the Firehawk flames lingered compared to the typical repeater's flames.

I ran the movie frame-by-frame, and noted the timestamp of the first frame that showed flames on the target. I advanced, and noted the timestamp of the last frame that showed flames on the target.

x4 Repeater flame duration: 0.8 seconds
Firehawk flame duration: 10.3 seconds.

So yeah, I'd say that the Firehawk does more DoT than a typical incendiary repeater. Though a better test would be to actually check the damage done, but I haven't gotten to those tests yet.



Conclusion & Summary

Eh, what? Yeah, this is even more interesting than the revolver test, probably because I'm beginning to accumulate more knowledge and do better testing. At least I think so. But I don't consider these tests to be definitive in any way. We are still stuck with Borderlands' inherent randomness, and the propensity to use description lines that mean something slightly different than the actual effect. Definitive tests would require much, much more time and effort.

But these tests are informative, and interesting.
__________________
DPS Visualizer - A Damage Calculator with Charts and Graphs (Dead)
My Borderlands 1 Test Threads
Random loot is random, with a lot of modifiers, many of which are unknown to us.

Last edited by Scottes; May 1st, 2010 at 05:33 PM.
Scottes is offline   Reply With Quote
Old May 1st, 2010, 05:26 PM   #2
duncanfogg
Cool Hwip
 
duncanfogg's Avatar
 
Join Date: Feb 2010
Location: United States
Posts: 724
Default Re: Some Tests Of Elemental Repeaters

Good stuff Scottes. I'll have to go through your results later and use it to clarify or confirm what the game files say.
duncanfogg is offline   Reply With Quote
Old May 1st, 2010, 05:28 PM   #3
Scottes
Elemental Scientist
 
Scottes's Avatar
 
Join Date: Jan 2010
Posts: 4,534
Default Re: Some Tests Of Elemental Repeaters

Please do! There are some findings that go against things you've said, and/or things I've seen in your Gear Calculator.
__________________
DPS Visualizer - A Damage Calculator with Charts and Graphs (Dead)
My Borderlands 1 Test Threads
Random loot is random, with a lot of modifiers, many of which are unknown to us.
Scottes is offline   Reply With Quote
Old May 1st, 2010, 08:12 PM   #4
ampersand
 
ampersand's Avatar
 
Join Date: Feb 2010
Posts: 226
Default Re: Some Tests Of Elemental Repeaters

Quote:
Originally Posted by Scottes View Post
Next up are a couple of x4 repeaters with varying chances to proc:



Both of these have a Pool_Tech of 60.
I was really looking forward to the results of these, because the tech pools are the same, and the only significant difference is the high/very high chance.

For a while now I've assumed the high/very high chance made no difference, but I always hoped I was wrong.

Quote:
Originally Posted by Scottes View Post
For the x4 High chance to proc 10 times, it took 25 bullets 5.9 seconds.
For the x4 Very High chance to proc 10 times, it took 14 bullets over 3.0 seconds.
So, this made me very happy. That's a pretty huge difference.

Quote:
Originally Posted by Scottes View Post
For the x4 High chance to proc 20 times, it took 28 bullets.
For the x4 Very High chance to proc 20 times, it took 41 bullets.

41???? WTF? The High chance repeater took only 28. So I ran that one again:

For the x4 Very High chance to proc 20 times, it took 35 bullets. STILL more than the High chance.
Then, I read this and my brain caught on fire.

Quote:
Originally Posted by Scottes View Post
I don't know why, but I'm going to blame it on Borderlands randomness and/or BL's tendencies towards ambiguous description lines.
Indeed.


This has nothing to do with the test, but I was shooting Athena in the face with a volcano the other day (as you do...). The x3 procs lingered for a long time (didn't time it, but about 7-8 seconds), but the x4 procs (****ing huge) disappeared in less than a second. What does this mean? Absolutely nothing. I just never noticed it before.
__________________
))<>(( Forever.
ampersand is offline   Reply With Quote
Old May 1st, 2010, 08:35 PM   #5
Scottes
Elemental Scientist
 
Scottes's Avatar
 
Join Date: Jan 2010
Posts: 4,534
Default Re: Some Tests Of Elemental Repeaters

Quote:
Originally Posted by ampersand View Post
This has nothing to do with the test, but I was shooting Athena in the face with a volcano the other day (as you do...). The x3 procs lingered for a long time (didn't time it, but about 7-8 seconds), but the x4 procs (****ing huge) disappeared in less than a second. What does this mean? Absolutely nothing. I just never noticed it before.
I might have to check that out. That is weird.
__________________
DPS Visualizer - A Damage Calculator with Charts and Graphs (Dead)
My Borderlands 1 Test Threads
Random loot is random, with a lot of modifiers, many of which are unknown to us.
Scottes is offline   Reply With Quote
Old May 1st, 2010, 08:39 PM   #6
Maxxfury
Proud Troll Enabler >;-p
 
Maxxfury's Avatar
 
Join Date: Jan 2010
Posts: 2,697
Default Re: Some Tests Of Elemental Repeaters

The best point i derived from this is something i already knew.........

firehawks are the tits!!!

__________________
<Not wasting 5 points in trespass with your vengeance? priceless!!>
From Hunter to Ninja! in one new installment!
Dont call a mag, a clip! FFS!
Maxxfury is offline   Reply With Quote
Old May 1st, 2010, 08:56 PM   #7
FeenicksPhyre
Champion of Metal!
 
FeenicksPhyre's Avatar
 
Join Date: Mar 2010
Posts: 931
Default Re: Some Tests Of Elemental Repeaters

Everything i know is a bloody LIE!!!!!!!


Quote:
Originally Posted by Maxxfury View Post
The best point i derived from this is something i already knew.........

firehawks are the tits!!!

Cept this... this is always true....
__________________
*working on new sig*
FeenicksPhyre is offline   Reply With Quote
Old May 2nd, 2010, 12:32 AM   #8
duncanfogg
Cool Hwip
 
duncanfogg's Avatar
 
Join Date: Feb 2010
Location: United States
Posts: 724
Default Re: Some Tests Of Elemental Repeaters

I think you are considering the wrong definition of the Maliwan "high chance" type effects. Their benefits are not in the probability variables as you are testing, but in the increase in tech level, and thereby tech pool. I suppose you could consider it as increased total proc power.

I've been using your numbers to match to my Elemental Simulator spreadsheet. So far, I've done several full-auto trials but I don't think I can conclude much since the variability is large for even the 40 or so independent trials I did, I'll probably have to write a script to get it to converge. I'll try single-shot tomorrow.
duncanfogg is offline   Reply With Quote
Old May 2nd, 2010, 06:23 AM   #9
Scottes
Elemental Scientist
 
Scottes's Avatar
 
Join Date: Jan 2010
Posts: 4,534
Default Re: Some Tests Of Elemental Repeaters

I had originally thought the same thing until I ran into a couple of Volcanos with the same size tech pool yet wildly different "effect chance" descriptions. That's what made me think about probability.

But after these tests, I am leaning back, and feel that the higher chances just indicate a larger tech pool. But it's still not conclusive. I think that someone skilled with weapon creation could dig deeper into that portion.
__________________
DPS Visualizer - A Damage Calculator with Charts and Graphs (Dead)
My Borderlands 1 Test Threads
Random loot is random, with a lot of modifiers, many of which are unknown to us.
Scottes is offline   Reply With Quote
Old May 4th, 2010, 03:46 PM   #10
duncanfogg
Cool Hwip
 
duncanfogg's Avatar
 
Join Date: Feb 2010
Location: United States
Posts: 724
Default Re: Some Tests Of Elemental Repeaters

I just wrote a quick script that simulates elemental procing encompassing most conceivable variables (regen, pool size, probability, proc cost, reload, fire rate).

My first few tests I've set it to terminate test runs whenever 10 procs occur on full-auto, to mirror your test above. I let it run for 30000 runs and got the average results.

Here they are for your TK5 C:


AVERAGES
---------------
Tech X1: 3.59
Tech X2: 2.10
Tech X3: 2.49
Tech X4: 1.82
Cumulative: 10.00
Elapsed Time: 3.91
Shot Count: 18.39
Depleted Count: 0.00


It shows the average number of procs at each proc level comprising the 10 procs. It also shows averages for elapsed time for 10 procs, the number of shots it took, and the number of times the tech pool was depleted to a non-procable level during firing.

I think it agrees pretty well with your full-auto tests on X4 repeaters.


Scottes, I'm curious if you've done any tests on elemental damage. I think if we can get some numbers for damage, I can combine that with my simulator to calculate elemental DPS.

Last edited by duncanfogg; May 4th, 2010 at 03:49 PM.
duncanfogg is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -5. The time now is 11:09 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.