Forum:Knoxx's Gold - It looks like it does increase rarity (maybe)

I decided to see if I could work out if using Knoxx's Gold to kill Knoxx made any any difference to the rarity of items found in the Armory. I didn't expect it would, but it may just be possible that it does!

First a few disclaimers - I'm NOT a mathematician, let alone a statistician. I only did 10 runs for each condition, and that's probably too small a sample size for reliability, but each run took me several hours to do, as I had to record and spreadsheet every item found by level, rarity and cost (cost turned out to be meaningless, but I didn't know that when I started). I recorded every item, not just weapons, but unfortunately I didn't think to to code whether each item was a weapon, class mod, grenade mod, or shield. This has made analysis harder than it should have been.

Also, I missed 3 chests, which I didn't know were there until halfway through the experiment. I don't think this invalidates the results, as I checked them (but didn't record them) in subsequent runs, and they were not unusual in contents.

My method:
In the Super-Marcus Sweep mission, I did 10 runs where I transitioned into the Armory with Knoxx's Gold equipped and in hand. I then made sure the first shot that hit Knoxx was from the KG, then switched to other weapons for most of the fight. I made sure that the final killing shot was also with KG. I kept KG equipped and in hand while I did the glitch and ransacked the Armory. I don't know if any of these conditions are necessary, but I wanted to give it as good a chance as possible of getting a result.

In the other 10 runs I got rid of KG entirely - it wasn't even in my inventory, as I wasn't sure whether just having it in the inventory might make a difference. I just killed him, and then went to do the ransack of the Armory.

Just in case there might be an effect from the order in which I did the runs, I did them as 4 sets of 5 runs: first I did 5 runs with KG, then 10 runs without KG, then 5 more runs with KG.

So - on to the results:
In the runs without KG I found 3072 items in total - here's a table by level of items found, and another of items by rarity. In the runs with KG I found 3161 items in total - again, here's a table by level of items found, and another of items by rarity.

As you can see there is a noticeable increase of Orange items when KG was used to kill Knoxx: 3.8% of all items as opposed to 2.8%. In addition, I found 4 Cyan items as opposed to zero when not using KG. Whether these results are statistically significant I have no idea, but it's interesting, isn't it?

Or to put it another way: without KG I found an average of 8.6 Orange items per run; with KG the average was 12 Orange items per run. Here a graph of rarity of items under the two conditions by percentage:

I'll let you draw your own conclusions, but it looks to me like the "extra" Orange and Cyan items come at the expense of Green and Purple items. I'm happy to have less Green items - not so happy to have less Purple ones LOL.

I've also graphed level of items by percentage:

I've no idea whether this is meaningful - it's probably just sample error - but with KG you seem to get slightly less high level items and more level 58's. If this is a real result it's not good news. On the other hand, you appear to get less level 49's (low quality Eridian weapons), which is a good thing.

Since I had coded each item by level I was able to do a little filtering to see if the differences remained when I removed level 48 and 49 items (the low-level chests) and level 55 items (which are all class mods: no weapons, grenade mods, or shields). The short answer is: yes!

Here's a table of filtered rarity without KG:

And another table of filtered rarity with KG:

And a graph of the percentage under the two conditions:‎

Again, draw your own conclusions.

For completeness: average item price without KG was slightly higher at $ 1,047,777 Vs $1,035,992 with KG ( I only recorded cost to the nearest $10,000 - I'm OCD, but I'm not that OCD). Average level was essentially identical: 55.40 for both conditions.

For the spreadsheet I coded white rarity as "1", green as "2", blue as "3", purple as "4", orange as "5", and cyan as "6", which let me work out average rarity for all items found. Without KG average rarity was 2.73; with KG it was 2.76. I didn't try to split orange into light and dark, as this seems to be a continuum rather than a either/or thing, and anyway, the sample size was too small.

If anyone wants to do more runs to add to the data, that would be great. Don't bother recording cost though - it's pointless - just record each item by level and rarity; if you want to code it by item type, that's up to you, but it will make it a much slower process. A second monitor for your spreadsheet helps.

Make sure that you follow the exact methodology regarding killing Knoxx that I used, or else the results won't be comparable. Also, all my runs were done during Super-Marcus Sweep; make sure that your runs are too, as the level and rarity of weapons changes in the the other Armory missions.

The three chests I didn't record are the ones that you get to by going across across the drawbridge, and down the elevator, (they are under the platform). You probably shouldn't record these ones either if we want the data to be consistent. You can safely access the chests behind the octagonal room: don't go through the octagonal room (or you'll set off the Steal Loot task); instead access them via the drawbridge and lift from the top floor and come back the same way. Don't forget that there are two red chests on top of the big containers on the top floor.

If we have any statisticians on the Wikia who could do some analysis of my data to determine if the differences I've found are statistically significant I can supply the raw data as an Excel spreadsheet.

I'm off now to get my life back. Outbackyak 07:45, June 29, 2010 (UTC)