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 October 17th, 2010, 08:09 PM   #1
Scottes
Elemental Scientist
 
Scottes's Avatar
 
Join Date: Jan 2010
Posts: 4,533
Default Testing Scavenge Extra Items

Testing Scavenge Extra Items

As most of you know, I tested Find Rare Items against a large number of Lance Chests., and wrote up the results in the "+2 FRI" thread listed in my sig. Short version: FRI didn't do anything detectable.

Some time later, a new member named Anarch came along, and started talking about statistic and probabilities and all sorts of math stuff. We had a great discussion, and the two of went off and tested it again. Short, more accurate version: Find Rare Items does not influence the rarity levels of weapons obtained from high-level Lance Chests.

This bugged the hell out of me. I wanted to know how it worked, and what it affected, and how to best use a COM with +FRI.


Work, Work, Work

So about 5 or 6 weeks ago I started working on my test map, gearing it up to test FRI against enemy drops. Heck, if it didn't work against chests, maybe it will affect the quality of loot dropped by enemies.

So I started working on my test map and making runs collecting loot. Anarch wrote a great program to dump a character's backpack to a data file. And I started learning C# in order to write a better program to analyze loot.

All in all, we've put in a *ton* of hours over the last 5 or 6 weeks.


Configuring My Test Map

Obviously, I *really* want to know how FRI works. So I configured my cages to spawn all different sorts of enemies found in Knoxx, but not the Badasses or Bosses. I made 25 spawn cages, and had each cage spawn 200 enemies. That’s 5,000 enemies, for the math-impaired. I killed them all and collected the loot.

I did this with my Mordecai with a low-end Ranger COM, and once again with a +2 Find Rare Items COM. I picked up all the loot - over 12,000 pieces of loot. And I did some quick analysis, and found little difference. Both runs yielded almost the same exact number of items and the average rarity was almost identical. Yeah, the numbers where *that* close.


WTF?

A few of the enemies were poor choices for such a test. I killed 200 Alpha Skags and they dropped about a dozen pieces of loot. Drifters, on the other hand, are veritable loot fountains. It was a serious pain to pick up all the loot before it disappeared. (Even after configuring the loot to stick around longer.)

I re-configured my map, and set it to spawn only Bandits and Crimson Lance. Once again, the differences between two runs were insignificant.


Poking Through The Editor

So I started poking through the editor, looking at enemy loot pools, looking for enemies that tended to drop decent amounts of good loot. I concentrated on Crimson Lance, particularly Knoxx-based Lance, since they seemed to drop good stuff. I figured that two runs with a greater amount of better loot would be better to compare.

I found a bunch of good candidates, so I re-configured my map changed to have all Crimson Lance spawn, just the ones with good loot pools. Another 10,000 killed, another 12,000 pieces of gear collected. And the same results. Without the Ranger COM, Mordecai collected 6,119 items. The weapons had an average rarity of 6.7 (low-level Green). With the FRI COM he got 6,143 items - only 24 more. And the average rarity was lower - only 6.6.


WTF??? I had killed 30,000 enemies, and FRI was doing nothing for the drops.


Won A Prize

I went back into the editor, and stumbled - yes stumbled - onto something called ScavengerSkillPool. Digging in, it seems to improve the odds of loot drops, particularly for Weapons and "A_Pool_Gear" - which is Shields, COMs, and Grenades.

Interesting!

And then it hit me: The FRI COM also had +3 Scavenge Extra Items. Why wasn't I getting more loot?? With SEI active I should be getting Extra Items - more loot. But the enemies were dropping the same number of items every time! (Well, damned close to the same numbers...)

So I started tracking down where ScavengerSkillPool is used, that is, what other things in the game contain a reference to it. Here's the list:



If you peruse that list, you'll see that there's no mention of either Bandits or Crimson Lance. As a matter of fact, none of the 30,000 enemies I had killed are on that list.

This is proof of sorts. I just killed 30,000 enemies that are not on that list, and I didn't see any change. This is a very strong indication that Scavenge Extra Items only works on the things on the list.


But it's not enough proof. To prove it, I'd have to kill some things that are on the list and see a difference between the drops. And I'd also have to kill more things that are not on the list, and find no difference between the drops.


Killing Some On The List

So I re-configured my test map once again, and set it to spawn 500 Marleys, since I remembered that this mangy mutt dumped a good amount of loot. I ran three times - Ranger COM, SEI COM, and FRI COM which also has SEI. Both of the runs with SEI produced about 10% more items. Perhaps I was on to something? So I tried again, a few more enemies from the list.

I killed 500 Marleys, 1,000 Badass Guardians, 1,000 Badass Spiderants, and 1,000 Badass Scythids. I did this three times. Another 10,500 enemies bit the dust in the name of science.
Code:
      Run & Task |500 Marley                             |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |    997  1,104  ( +10%)  1,110  ( +11%)|


      Run & Task |1000 PrincipleBadassGuardian           |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |  1,252  1,449  ( +15%)  1,431  ( +14%)|


      Run & Task |1000 BadassShockSpiderants             |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |    514    685  ( +33%)    696  ( +35%)|


      Run & Task |1000 Scythid                           |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |     68    117  ( +72%)    114  ( +67%)|
Every one of these enemies is on the above list, and every one of them consistently dropped significantly more loot for both SEI runs.


Killing Some That Are NOT On The List

So what if I tried a few enemies that are *not* on that list? They should drop equal amounts of loot for all runs, right? I line up 1,000 Badass Desperados, 500 Boneheads, and 500 Hanzes. Again, three times each. And I grabbed the results from two runs against 5,000 Lance, just the Ranger and FRI COMs. I never did an SEI run against the lance, but the FRI COM has SEI, so it's a valid test.

Code:
      Run & Task |5000 AllLanceMix                       |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |  6,119                  6,143  (  +0%)|


      Run & Task |500 Bonehead                           |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |    636    646  (  +1%)    643  (  +1%)|


      Run & Task |1000 BanditBadassElite                 |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |  1,102  1,097  (  -0%)  1,100  (  -0%)|


      Run & Task |500 Hanz                               |
       Character |Mordecai                               |
             COM |   None    SEI             FRI         |
=================|=======================================|
        All Loot |    161    153  (  -4%)    116  ( -27%)|
There is, consistently, an insignificant difference in the amount of loot. Random loot is random.

Well, Hanz's FRI run produced lower results, but he didn't really drop much to begin with, so it's difficult to detect the difference.

In this game, due to the randomness, one generally must really consider large amounts of loot.. The numbers on all these runs isn't large enough for Average Rarity tests, like those from my previous FRI tests. But they do show sizable differences, and consistency.


Summary

IMHO, the results above prove when Scavenge Extra Items works, and where it doesn't.

Well, I'll never be absolutely certain until I test every item on that ScavengerSkillPool list. But indications are extremely strong with this one. The above results show consistent significant differences with things on the list, and consistent insignificant difference with things that are not on the list. The results match the theories. That's science.


It seems extremely likely that Scavenge Extra Items only works on the enemies and items listed in this pic:



That covers many, many enemies - just about every single one except humans. So don't bother using SEI against Bandits or Crimson Lance.


Further investigation of the list shows that it should work against Crawmerax and his green minions. It should also work against skag piles (called "ScrapPile" on the list).
CORRECTION: Crawmerax is not on that list. His minions are, but not him. So it seems that SEI does not work on Crawmerax. Sorry.

It will do nothing for machines or chests - but that makes sense. How do you get extra items from a machine? And chests have set item counts.


Coming Soon

My investigations through the editor, looking at loot pool info, have given me some leads on Find Rare Items. I will be working on that next. I have already done over 20 runs, but I have yet to analyze anything. I expect to need another 10 or 15 runs before I can prove anything around FRI. Hopefully I will have some concrete info by next weekend.


Further Reading
Here are the full reports from all the SEI runs mentioned above. There's tons of info, showing individual counts for most everything - weapon types, grenades types, COMs for each characters... Tons.

Please note that some of the numbers are far too small to be anything other than random. For instance, you will find that the Lance dropped 0 Pearls on the Ranger COM run, but dropped 2 on the FRI run. That's just randomness. Note that the FRI run also produced 6 fewer Oranges. With results that low, it's just random.

Also, I can't measure the Rarity of COMs or Grenades. For COMs, I calculate the number, like 660 Scavenger or 550 Mercenary. I then calculate the Average Number, so it's kinda like calculating the Average Rarity for weapons.

http://www.itsanadventure.com/postim...orks_Proof.txt
http://www.itsanadventure.com/postim...Work_Proof.txt
__________________
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; October 25th, 2010 at 08:10 AM.
Scottes is offline   Reply With Quote
Old October 17th, 2010, 08:22 PM   #2
Smeggit
A What Now?!
 
Smeggit's Avatar
 
Join Date: Mar 2010
Location: New Zealand
Posts: 873
Default Re: Testing Scavenge Extra Items

Awesome investigation dude. I was beginning to think SEI only gave extra money/ammo/health drops but I was just killing the wrong things.
Smeggit is offline   Reply With Quote
Old October 17th, 2010, 08:39 PM   #3
Co2 Elite
Okami > Zelda
 
Co2 Elite's Avatar
 
Join Date: Aug 2010
Location: Raleigh, North Carolina
Posts: 2,790
Default Re: Testing Scavenge Extra Items

So much insight from the extremely wise. Kind of odd how it only works on specific enemies but oh well. It makes sense that the scavange extra items would only be on enemies because they actually drop loot (as opposed to chests spawning more specific amounts) but I don't see why Lance and human enemies wouldn't be able to drop extra.
__________________
Misery's sig is full of lies._
Co2 Elite is offline   Reply With Quote
Old October 17th, 2010, 09:19 PM   #4
Newonemax
The path less traveled.
 
Join Date: Mar 2010
Posts: 170
Default Re: Testing Scavenge Extra Items

SEI has always worked fine and results in a much better chance of finding good loot off of monsters

If you have ever modded +50 SEI then you will remember the fountain of loot that ensued
__________________
The Original Borderlands Samurai-Ninja. Master of Melee
Advocate of Repeaters,Pounders,Thumpers and now Gambles or otherwise refereed to as the stuff everyone hates.
Newonemax is offline   Reply With Quote
Old October 17th, 2010, 09:39 PM   #5
REEEPR
Lil late 2 the game
 
REEEPR's Avatar
 
Join Date: Jan 2010
Location: Outside, looking in
Posts: 2,825
Send a message via Skype™ to REEEPR
Default Re: Testing Scavenge Extra Items

Scottes, you give hope to the hopeless. Thank you. You get the "Zed peace prize" it's more prestigious than that cheap Nobel thing lol.
REEEPR is offline   Reply With Quote
Old October 17th, 2010, 10:09 PM   #6
Szyslak
Cantankerous Masher
 
Szyslak's Avatar
 
Join Date: Sep 2010
Location: NY
Posts: 1,810
Default Re: Testing Scavenge Extra Items

That's some great work Scottes. Thanks to you and Anarch for putting in the work to help us understand the characteristics of those COMs. I'm looking forward to any light you can shed on the FRI situation. It sure would be nice to have more information about it one way or the other.

I really admire the thoroughness, ingenuity, and dedication of your testing here. Amazing stuff.
__________________
(PC) GameSpy: Szyslak22 GMT-5 134 Pearls (breakdown)

Awesome Level Thread
Szyslak is offline   Reply With Quote
Old October 17th, 2010, 10:35 PM   #7
Nublarian
First Access Club Member
 
Join Date: Oct 2010
Location: Missouri
Posts: 197
Default Re: Testing Scavenge Extra Items

your the man
Nublarian is offline   Reply With Quote
Old October 17th, 2010, 11:17 PM   #8
ense7en
Party
 
ense7en's Avatar
 
Join Date: May 2010
Location: Edmonton, Alberta
Posts: 1,718
Default Re: Testing Scavenge Extra Items

Glorious work!

Always a wealth of info to be found in your investigations, thanx.
__________________
Xfire
Steam
Gamespy = ense7en
"Best" chests in Armory (highest Awesome Level)
ense7en is offline   Reply With Quote
Old October 18th, 2010, 12:55 AM   #9
Ice Car
life sucks, drop out
 
Ice Car's Avatar
 
Join Date: Dec 2009
Location: Hawaii
Posts: 15,074
Default Re: Testing Scavenge Extra Items

About damn time someone did this.

And nice work Scottes. I only read a few lines and have a general idea. So SEI only works on Skag Piles and everything other than humans? Craw works? Damn. Well I'm never going to be able to solo him as a hunter so I guess that's out of the question...
__________________
Tumblr」「Anime List」 「Gifs
Ice Car is offline   Reply With Quote
Old October 18th, 2010, 03:53 AM   #10
Vashts1985
e-sama bin laden
 
Join Date: Nov 2009
Posts: 948
Default Re: Testing Scavenge Extra Items

i take it there is no illusive list for FRI to "stumble" upon?

be nice to finally put that one to rest right?
Vashts1985 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 08:00 AM.


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