Forum:Help me find out if the +2 rare item finder works on chests!!!!!

At this point in the experiment we do not have enough data to see valid results.

Experiment: Does wearing a +2 Team Find Rare Items Mod effect the items which spawn in a chest?
ok so this is how we can compile statics and help up find out if they really work on chests. to do this we all have to follow the same rules that way all our numbers will work together instead of work against each other.

RULES MOVED TO TOP OF PAGE by Doctorgray

also if someone knows how to set up a chart that would be nice, i do not.

so i got 4 runs i'll put up

WITH +2 RARE finder CLASS MOD ON
 * runs                   25   -   25   -   25   -   25
 * average item per run  14.52 - 13.48  - 14.56  - 13.92
 * pearlescent              1  -   0    -   0    -    0
 * orange                   2  -   5    -   1    -    8
 * yellow                   5  -   8    -   4    -    7
 * purple                  64  -  51    -  49    -   30
 * blue                    77  -  69    -  87    -   86
 * green                  122  - 121    - 125    -  120
 * white                   89  -  83    -  96    -   91
 * eridian                  3  -   2    -   2    -    6
 * total                  363  - 323    - 364    -  348

ideas, thoughts and constructive criticism are all welcome. i hope others out there are fed up that they will not tell us what exactly what it affects and want to try and find out.

Wylde bil 06:26, December 12, 2009 (UTC)

- As per the older thread Forum:Rare_Item_Effect_Mods, i did a bit of the same thing --Raisins 06:30, December 12, 2009 (UTC)

Newhaven 6 chest runs. Seven times each with and without +2 Find Rare Items.

I really hope it works because I've done all my looting there w/ a +2 rare item mod. I've noticed that when I don't wear it, I usually see white/green items in chests and when I do I see more blue and purple. Then again, it could just be my imagination. I'll do some runs tomorrow and post them.

Anyway, I did an analysis of the stats above and so far there's not enough data to tell whats up. What I can tell you is this: You're likely to get about 2.3 items per chest regardless and orange items are going to come up about or less than .05% of the time. This means that you're likely to see an orange item every 20 chests or so. Now this doesn't mean you can't go 100 runs w/out an orange item but it means that you're likely to average out to about that. I'm guessing the mod has an effect, but I don't think its anything remarkable. It's probably just a difference of .01 % chance per chest for orange and a bit more for less rare items.

Suggestions: Instead of asking people to do this in sets of 25, ask them to do this in multiples of 5 that way people can still contribute even if they don't have the time to spend an hour in New Haven.

I do not recommend doing the 6th chest (near the bandit encampment) in this run. A) It takes more time than its worth B) Its too easy to forget not to kill enemies thereby creating a risk for corrupt data. C) It's much easier to compute data w/ multiples of 5 or 10 (5 chests, 5 runs, etc)

--Doctorgray 11:21, December 12, 2009 (UTC)


 * i fear i must respectfully disagree with my esteemed colleague. i am unconvinced the bonus does not apply to the entire area occupied by the player vice played as an instant (MtG ref) i.e. when each item spawns.  i am interested in this thread as my theory is the mod provides an area of effect/affect and to be fair my proof would require raiding _all_ boxes _and_ killing every spawn-able enemy in said area (new haven).  this is tedious i know but i should point out that i rarely get [0] yellow+, but from  your data neither do you.  Dr. Clayton Forrestor 14:37, December 12, 2009 (UTC)

doctorgray that makes complete sense it just frustrates me that my data is no good for that now cause the numbers are skewed for the 6 chest and that was 4 hours of work, umm for the 6th chest if you go the back side to the chest you never fight bandits and if you hug the trash pile the bugs don't even come out as you see i have extensive knowledge on this because i did it for 4 hours lol if we shorten it ot the 5 chest in the town then do you think we have to throw my data out? also side note i have found weapons i use out of that chest though i will admit i had to go through tons of trash guns to get to them. Wylde bil 13:57, December 12, 2009 (UTC)

GOD DAMN there are alot of doctors in this room LOL, makes me feel sooo soo soo small anyway dr. clayton one step at a time i was thinkin the same thing but just wanted to do the chest first, also doing enemies require documentation of what type of enemy. i mean superbad drop better stuff then normal bandits, and the random generator they have doing the guns is similar if not the same as the one that does enemies sooo we could go through once and get no superbads then turn around and hit 6 or 7 and that will skew our numbers so that would have to be done separately, look at all the work we have to do cause gearbox cant give us the simple answer of what +2 rare finders effect. Wylde bil 15:47, December 12, 2009 (UTC)


 * it is my sneaking suspicion mr bil, that you have hit upon the very reason _why_ gearbox has not and may never reveal the exact process of +x find rare items. being the fact that gamers such as ourselves will exhaust hours of gametime and compose titanic treatises such as this just to uncover a theory as what it does.  gametime = >product = >$,£, etc.  Dr. Clayton Forrestor 16:59, December 12, 2009 (UTC)

To both Dr's points, either it works on chests or it doesn't. Testing any set of chests (or set of enemies) should show the effect if it exists, you shouldn't need to test the whole area. Wylde bil's data is fine, probably. Paranoid musings: Loading-in is not the same a zoning in. I know this because the secret Weapon Vending Machine never spawns if you start the game in New Haven, but has a chance to whenever you travel to New Haven from anywhere else. Maybe this effects item generation. --Raisins 17:21, December 12, 2009 (UTC)
 * my dear raisins you make an excellent point however comma your reasoning is contradictory. im not practicing proctology here, just saying, this mod either affects the area or it doesnt.  i had not put together that the vendor never opens when spawning excuse me loading in.  and i express no refutation whatsoever of Mr. Hiccock's data.  i was just throwing my sandal into the cogs for a bit of unearned glory on the OP's coat-tails. i shall shut up now as the nurse is coming for my weekly sponge bath and im not allowed to be "playing-on-the-internet" excepting outdoor periods.  Dr. Clayton Forrestor

In response to Wilde Bill:

We shouldn't have to throw out your data as long as you didn't kill any enemies. Every thing will work out fine, I was just making the point that not only is it quicker but its easy to compute. Anyway, if some more people take the time to write down their results we should be able to figure this out quite quickly. It'd be nice if we could get 10,000 chests worth of data. This means we'd need 2,000 runs I.E. 200 people to do 10 runs each. (5 w/ 5 w/out) Please please don't take this to mean that you should do it yourself. I'm sure some people will contribute over time as this issue isn't going away. I'll try to get in around 50 runs today since I've got some time on my hands, I'll post the results here and the computations at pastebin.com --Doctorgray 17:31, December 12, 2009 (UTC)

really you think, cause i figure once they got our money they could really careless how long we play, they already got our money, look how long it took to fix core issues on all plat forms, then take in the account DLC are coming every month, i figure for 2 reason they know they got a good structure but lack things to make it re-playable other then us obsessive gun hunters and to make more money, telling us won't keep either of those facts from people playing the game but at the most change their strategies while playing them, now i only play 2 games borderlands and NBA 2K10 both are major titles for 2K and both came out with serious serious core issues that have taken months to patch had made me contemplate trading both in but because the structures of thee games were so sound i was basically sitting there and taking it and askin for more til they fixed the issues, both have made me restart because of these issues too, i guess i went on this rant for one simple thing telling me wouldn't change how long i play the game sooooooooooooo GEARBOX just tell me what it does and doesn't effect already cause the only thing i'm doing more is typing on this message board because of you not telling us. o and raisins i get that secret machine fine from loading in to the zone it happens about 1 out of 4 times for me and im glade i don't have to do it alone cause it wouldn't get done lol Wylde bil 17:34, December 12, 2009 (UTC)

125 chests - New Haven - +2 Mod

--Doctorgray 21:06, December 12, 2009 (UTC)

Hey everyone. I ended up doing 10 New Haven runs of the 5 chests, both with and without the rare item finder mod. I also sorted out the weapons I found by material grade (e.g. "ZZ" is uncommon and "XX" rare for Jakobs weapons) to see if the mod affected material grade rather than color rarity. Anyway these are my findings, feel free to consolidate the values into the groupings you want if you don't want them so separated.

Notes: I also placed low quality weapons into the Common Category (C). Also, I'm not 100% sure if Torgue and Atlas weapons display their material grade all the time, so the numbers may be skewed towards common for every Torgue and Atlas weapon I found. The Misc. category contains shields/grenade mods/class mods/eridian weapons.

100 chests - New Haven

--DrKirth 22:53, December 12, 2009 (UTC)

my results are percent of total (not percent per chest):   Dr. Clayton Forrestor 23:37, December 12, 2009 (UTC) 7 runs each. 6 chests.

*= notably two of these were ultra-white (cyclops and invader pistol).

ok wow this changed real fast anyway i have 100 more runs of 6 chest so a total of 600 more chest with the mod on, since i see we have more of that the next set i do will be without i also notice we got another doc too lol

WITH +2 RARE finder CLASS MOD ON
 * runs                   25   -   25   -   25   -   25
 * average item per run  13.68 - 12.28  - 13.64  - 14.36
 * pearlescent              0  -   1    -   0    -    0
 * orange                   5  -   4    -   6    -    6
 * yellow                   3  -   1    -   8    -    5
 * purple                  42  -  34    -  41    -   43
 * blue                    77  -  71    -  58    -   72
 * green                  128  - 115    - 133    -  149
 * white                   86  -  74    -  91    -   79
 * eridian                  1  -   7    -   4    -    5
 * total                  342  - 307    - 341    -  359

i can't wait for the next dlc so i can stop have 12 other character on another Xbox 360 id full just to hold all my gear Wylde bil 07:37, December 13, 2009 (UTC)

ooo also someone brought up about the extra machine in new haven if you want i have stats on how often it comes out to play and i have stats on what was in both of them, though i was only counting the pearlescent, orange, yellow and purple if anyone wants to know, o yeah i have started to notice on the runs i can start to guesstimate what is in the last 2 chest depending on the first 4 so far that i have been doing it I've been right about 75% of the time, i wanna see if anyone else notice some things. if you get more then whites then you get an orange, yellow or a couple of purples, but more greens or blues mean at best a yellow or purple. it might be random but each run seems to have a balance, i've had a run were i got 4 yellow / oranges but mostly everything else was white, i also seem to get more oranges when i open revolver or repeater cases it makes me wonder if that is cause of what i have in my hand or due to my proficiency level. anyway i'm glade we could work together to crack this mystery. Wylde bil 07:59, December 13, 2009 (UTC)

--Doctorgray 10:43, December 13, 2009 (UTC)

Not to mess up the flow here or just interject without contribution (though I am playing on xbox). Do you think it would be wise to do a small sample with 4 players using +2 rare item find mods? This SHOULD have a more pronounced effect on the results. Something just seems plain wrong about the current results... as there is more than enough data collected for looting the chests while the mod is on and nearly enough to get a rough guess of the results for looting with the mod off. (sorry if my formatting is off, i'm new to wikis) --sX-- 05:55, December 14, 2009 (UTC)

At this point in the experiment we do not have enough data to see valid results. I predict that round 2000 chests each we'll start to see what the percentages are going to end up like. To get a valid data set I'd like to see 5000 chests each. This will obviously take time and no one is in a hurry here.

Your suggestion to have a small sample might seem valid but it's completely wrong. For example, you can flip a coin 5 times and have it come up tails each time. According to your method this would mean that coins always come up tails. However, if you flip a coin 1000 times the chances of it coming up tails 100% of the time is minutely small. This is why larger data sets are better... Odd runs may occur, but over time your data is going to come to its center. --Doctorgray 07:03, December 14, 2009 (UTC)
 * "Completely wrong" is somewhat arbitrary. The data is valid, your confidence is what increases with the amount of data. 95% level confidence in our proportions is approximately B=sqrt(1/N). Even at populations 359 and 1459 items, that's 5.2% and 2.6% (N.B. this is P/item not P/C. We don't have data on individual chests so we should switch to P/item anyway.). I agree it's not yet conclusive, but it's close to enough to start worrying. --Raisins 08:36, December 14, 2009 (UTC)


 * What i'm saying is that either the +2 rare item is going to have a profound effect on loot, or it is going to have a very small effect on loot, requiring a very large sample size to come to a definitive conclusion. Right now from the data that has been collected, something seems very off, with the +2 rare item find yielding more whites.  +8 rare item find (from 4 players in a game each with a mod) SHOULD yield much more noticeable effects requiring a smaller sample size to come to a conclusion.  at least i'm assuming that the point of this whole experiment is PRIMARILY to determine if +2 rare item find effects chests and secondarily how much it effects them.  --sX-- 18:09, December 14, 2009 (UTC)

yes i agree this is and will be a slow and arduous task and expecting a quick answer or to be able to see anything in incomplete data is very presumptuous; also it is not limited to the 4 people doing it now, it is open to anyone that can follow the guide lines on the top of the page. Wylde bil 13:54, December 14, 2009 (UTC)

I'll start running this tonight. Pdboddy 17:11, December 14, 2009 (UTC)

@--sX-- I don't know how many times I can explain this, but that wouldn't help at all. It would still be a smaller data set and you would still have odd data because you don't have enough. The fact that the data looks weird at the moment does not mean that the experiment is bad or not working, it just means that we're not done yet. The way we're doing it now is the best IMO, for a couple of reasons.

A) Its a pain to open lots of chests w/out killing w/ multiple people in game. B) We don't know the how +2 rare stacks with multiple people in game. C) No matter how wierd you think the data looks it's working just fine. D) Less Data = Less Reliable Result... It doesn't matter if the effect is profound, the more data you have the more reliable the result will be, and collecting a mass of data your way would be extremely difficult. Not to mention uneccesary.

@Raisins I know that the data seems odd but in reality we have absolutely no idea what effect the item will have on gameplay/finding of items so maybe it's not odd. Maybe the +2 rare negatively effects chests and we're on the right track. None of us know and none of us will know unless we get a decent sample size (~2000 chests each).


 * "@unsigned poster" you are just making excuses to go about things the hard way. If over 1000 chests isn't enough data to start showing a trend, then the mod either doesn't effect chests or it's going to take a unreasonable amount of time for this information to come together.  i could see you requesting people at this point to only open chests without the mod on, but i can't see wasting time going to or over a sample size of over 2000 chests for each without knowing the mods effect chests in the first place.  considering a way to get more dramatic results, thus PROVING that these class mods effect chests seems more important to me than having people waste god knows how many hours providing 1200 or more chests.  "if we equip +2 rare item class mods, it should provide better loot from chests."  that is the question you're trying to answer here, not what how often you'll get better loot.  if four players in a game with these mods on don't produce enough of a change over a small sample, i don't believe this mod would be practical in the first place and if you want to waste time finding out just how little impact it has, be my guest.  as for you blatantly shooting this idea down by saying "we don't know how these stack",  it's pretty safe to assume if any of the other team class mod effects stack, that this also stacks.


 * it would have been much more efficient to attempt to prove the mods work in the first place before aiming for a very large sample size. yes, smaller sample sizes will provide a larger swing in data.  yes, they are less accurate. reread the title.  "help me find out if +2 rare item finder works on chests."  four people could've had that much knocked out in a couple of hours.  instead, now you're stuck collecting data from multiple people in their spare time, hoping the information provided is reliable, and marching on into an infinite amount of data to collect to make sure it's accurate enough.


 * additionally, there are more things to consider here, such as whether what the game considers a better weapon has anything to do with the color of the weapon text. it is possible that the game might rate certain whites higher than greens or blues and that would be skewing data... though i am jumping the gun because i'd like to see more data on un-modded chests.  --sX-- 20:16, December 14, 2009 (UTC)
 * Say you get 1000 items in multiplayer, 500 with mods and 500 without. That would give results with plus or minus 4.4% confidence (19 times out of 20). Assuming the effect is four times as strong with +8 that might be enough to see it. But 1000 items is about 75 runs. That's a serious time commitment for four people. This is also assuming that multiplayer doesn't mess with the mod-relative rarity. --Raisins 22:10, December 14, 2009 (UTC)


 * do you know? i have been pondering that myself.  does this mod focus on loot/swag?  my detonating bitch flew out of a scythid and the quality of weapons used against me does seem to be higher with this mod.  also i have noticed enemies performing a draw/switch animation on approach - this makes no sense as they have only one weapon each on spawn, unless the mod gives it to them!  still i would _think_ the mod would apply to items _found_.  keeping in mind that rare could mean >green(blue+).  should we perhaps look at single chests like above scooter's and the deathbox in north Rust Commons West?  not in this study of course.  <+>Dr. Clayton Forrestor 19:41, December 14, 2009 (UTC)
 * i was thinking the same thing earlier about doing it on a single chest. considering we have no idea how chests work in the first place.  it is possible that gearbox could've designed the game to allow certain chests to lean toward a specific type of weapon or rarity.  the only problem with this is that it would take a long, long, long time.  i think if this approach was done (in another test), that it would probably be best to still loot the 5 chests in new haven, but to record the results of each chest individually.  another potential issue that i noted above is that the game may treat the actual "value" of loot different than our own perceived value (text color).  it may be possible that the resale value of the item might be a better indicator.  perhaps a new test that could address several of these questions would be a good idea (after this one is done).  --sX-- 20:29, December 14, 2009 (UTC)
 * actually, it just occurred to me that if player level effects drops in any way, this data is flawed if everyone hasn't been looting these chests at 50. (observe chest drops early game vs item drops late game)  --sX-- 20:30, December 14, 2009 (UTC)
 * If these results go bad, removing things like level as systemic bias is an idea. Also I agree there are other things we could be measuring. As i understand it, every part has a rarity, which sum to a rarity total for the gun. The colors are ranges for that score. It would be best if we could somehow use that. --Raisins 22:10, December 14, 2009 (UTC)

ok were are the stats i posted i just posted 600 chest no mod and now its gone,

WITHOUT MOD
 * runs                   25   -   25   -   25   -   25
 * average item per run  13.48 - 13.88  - 12.76  - 13.76
 * pearlescent              0  -   0    -   0    -    0
 * orange                   4  -   7    -   5    -    9
 * yellow                   2  -   5    -   3    -    2
 * purple                  45  -  45    -  53    -   40
 * blue                    78  -  75    -  60    -   82
 * green                  128  - 129    - 115    -   99
 * white                   72  -  83    -  80    -  104
 * eridian                  8  -   3    -   3    -    8
 * total                  337  - 347    - 319    -  344

here it is again Wylde bil 22:14, December 14, 2009 (UTC)

Raisins -- If you're so convinced that your method will work better and faster, go do it instead of trying to convince us that our method isn't working.


 * Bil -- Thanks for the data.


 * Sx -- There are problems on both sides of the isle. On this side you don't know how Xbox/PC effects drops, whether level effects drops, whether cash effects drops, etc. On the Multi-Player (4 people w/ mods side) you don't know most of the same things (besides platform) plus you have to wonder how the mod effects stack; However, if you collect data from only one source, its a reliable way to do this. (It doesn't matter if its 4 people or 1 person, 1 source whose party/person doesn't change their $$, level, inventory and area throughout the entire experiment would yield the best results) Either way, the more data the better. However, this method should at least give us a good idea of what happens, if it doesn't, oh well. Also, trying to do it the other way could be a bit problematic. Most people won't be willing to not pickup a perfect {insert favorite item here}.

To those who think this is a waste of time... People aren't being forced to work on this. They don't have to submit findings. I imagine that most of the people submitting more than a small amount of data enjoy doing things like this. Just because we spent X hours collecting data doesn't mean we wasted our time. Not only are we collecting data for this experiment, we're also collecting many orange items we previously didn't have. For those of us who don't enjoy dupes/hacks/etc that's a good thing. --Doctorgray 23:18, December 14, 2009 (UTC)


 * your problem is that you're overly defensive about all of this. and yes... it is a waste of time if you have to open another 4000 chests because the experiment wasn't thought through and you wind up realizing you have to throw all of the old data out..  --sX-- 23:44, December 14, 2009 (UTC)
 * I never said anything was a waste of time. If this works then we don't need to do anything else. But if it doesn't work we will need to improve the next experiment. If --sX-- has the people and the time, his multiplayer idea could be useful as an independent experiment. As for this one, with Bil's latest we are well back to sanity, the data is fine now. So far we've proved the mod's effect on chests is ~2% or smaller:
 * RareFindItem 3339 2158.PNG--Raisins 00:08, December 15, 2009 (UTC)
 * I'm working on getting a siren to 50 right now, and i have a friend who has a 50 siren, so if anyone has a copy of borderlands for the xbox, let me know, and we can hook up later sometime and knock out some chest runs. i'll make a forum post when i'm ready for it.  --sX-- 09:01, December 15, 2009 (UTC)

yeah i really have nothing better to do all day got hurt at work and am waiting for my 2nd surgery, anyway boy do i have more stuff now i have 16 characters filled with stuff not to mention what i have on my 3 characters. now things i have noticed no clue if its luck or the class mod, but the items i have found with it on are far better, some of the purples i found with the mod have been ridicules. also i have notice that people that play multi-player seem to see higher damage guns, maybe that's cause they face worst enemies cause the highest normal damage gun i found was a sniper with only 901 damage, i've seen on here where people have volcanos with more damage then that. anyway, i'll still do runs, so we'll get to 2,000 each. Wylde bil 23:37, December 14, 2009 (UTC)

nice graph raisins! i've also been lookin into its effects on machines have those stats too. Wylde bil 00:31, December 15, 2009 (UTC)

I'm not attempting to be overly defensive, I'm saying even if this experiment fails it wasnt a waste of time and/or useless or anything else. It'll still serve a purpose. While you might be correct and we'll have to do this again, don't tell us that this experiment is ruined. We're having fun.

Anyway, I forget who mentioned it, but I think the idea that "colors" matter less than the rarity of the weapons themselves could be valid. I was in a game today where we both had +2 rare and I found my first DVL Volcano... Almost every other I've found has been GGN or VRR. So maybe the mod doesn't effect what colors come up but how rare the actual item under that color can be. In other words, you guys are probably right and we probably will have to do another experiment if we really want to know the effects of the mod. Even so, let us finish this in peace please :).

--Doctorgray 02:46, December 15, 2009 (UTC)


 * i'm just critical. but i, myself, would be fearful of having everyone contributing feeling burned out and discouraged if the data had to be scrapped and restarted after 4000 chests opened.  i also have fun figuring things out and chewing on puzzles, especially the mechanics of video games.  WoW had me doing lots of math when i played it a couple of years ago.


 * at any rate, the new total results for chest runs is troubling at the least... even if you account for margin of error, the numbers aren't adding up right. with the widest swing of error, the amount of purples you get WITH the mod on is 4% less than with it off.  the white drops are also similarly strange, and the weirdest thing to me is that the other colors are about even.  unexpected.


 * also, i understood what you were trying to get at when you were talking about the margin of error vs sample size. this wikipedia article on margin of error is pretty helpful if anyone else needs clarification ---> http://en.wikipedia.org/wiki/Margin_of_error --sX--
 * Are you using data from the spreadsheet error where it's adding no-mod purples wrong? (it said 375 39.10% 17.38% when it was really 313 32.64% 14.50%) Look at my chart, that's using correct data. Purples with the mod on could still be up to 2% better within error. --Raisins 10:29, December 15, 2009 (UTC)

Okay, 5 runs (25 chests total) with the mod on.

Please note: I have not modified the table at the beginning of the article. It's a little late, and I don't want to make a mess of things, I'll do it in the morning, or, someone else can if they wish. As an aside, the two orange weapons I got in this set of runs were from the same chest, a Maliwan Pestilent Defiler and an Atlas Troll. Two permanent additions to my backpack! :D Pdboddy 04:12, December 15, 2009 (UTC) - Before I forget, can someone with a Google account correct the error in the spreadsheet? On the No Mod tab, the Purple-ALL RUNS total in cell B13 is "=SUM(B2:J3)". It should be "=SUM(B3:J3)". --Raisins 09:51, December 15, 2009 (UTC)