Drop stats database

Color coding thanks to TracerX

Will keep updating like a nerd with OCD.

It’s almost all my own (except about 150 or so drops from TracerX), the wiki is too sparse and messed up to be worth the effort to convert one drop at a time. Plus this way I can maintain constant Luck scores for the drops instead of having to use ranges. Although I do have columns for that which were intended for other people to use…

If there are any forums people want more accurate info on let me know and I’ll set my cheated characters to grinding as soon as I get reasonable confirmation of all the different files of the last few forums (Obscure Phobia, clbum warfare and 1 or 2 other Episode 2 Forums) or if you REALLY care, you can send me 2BP and I will run 69 visits on that forum with that cheated character.

mterek Posted:

I have about 40 drops at 31 luck (lucky!) for World of Whorecraft that I could give you, should you accept it. I don’t have results for maybe 2-3 threads for various reasons, but 99% of the data is solid.

Missing a thread here and there happens every now and then, but I’ll accept any data you want to give. Just TM me and I’ll toss it into the appropriate column for you. Same goes for anyone with a reasonable amount of drops. (less than a whole Forums worth AT LEAST is pretty much statistically invalid, may as well pick and choose which drops you include at that point which from my point of view may have happened, otherwise why have only part of the forum?)

By the way, I need all the files that drop, not individual files otherwise I can’t get proper drop stats on everything and the pieces that are dropped are at random, Luck doesn’t affect it and neither does Forum. At least not from what I’ve seen so far or from what Banhammer has said.

Dirt_E_ATM Posted:

This could be quite useful.

I will try and start tracking my drops better. Do you have a preferred format for us to provide stats?

Not really, just send me your Luck at the time, and the file names followed by how many pieces of that file. Then send it when you feel like it. I only intend on updating all the numbers once a month anyways so it’s not as though I need it all ASAP.

Something like

Homicide Girls – 53 Luck

Buyer’s Guide to Expired Ground Pork.dtbs – 2

hollywood starlet fine upstanding member of society slip collage.xpeg – 4

Journal – what I plan to do differently the second time I have sex.sdoc – 1

would be perfect.

Knowing there’s someone doing something organised like this is making me want to start actually tracking my drops better; might have some extra data for low-luck runs of HoFF just as soon as I figure out an easy way to keep count of what pieces I get.

Interesting… going on current data, 10 luck gets a higher rate of Floyds than 1 luck. When I did a quick and dirty test of the same comparison (a handful of visits at each) I did better with 1 luck. There’s something to test.

man-man Posted:

Interesting… going on current data, 10 luck gets a higher rate of Floyds than 1 luck. When I did a quick and dirty test of the same comparison (a handful of visits at each) I did better with 1 luck. There’s something to test.

I’ve seen a lot of disconnect between 1 and 13 Luck which leads me to believe that whatever equation is used probably causes a curve in the drops so that 1 Luck is equivalent to something like 20 Luck. Either that or Luck has 0 effect until you hit a certain amount and what I’m seeing are just random fluctuations between equivalent amounts. Personally I think it’s the former but only ET knows the truth Log in to see images!

mterek Posted:

We could run controlled experiments using those values. I know I was planning on working out Defense and Upkeep’s critical values so I could calculate the optimal Re-Re point allocation, but with multiple people sharing data it becomes an immensely easy task.

Unfortunately I don’t think it’ll work with anyone who gets starter packs…so it would have to be a high level Re-Re or Permanoob to get the most significant data.

The fastest and probably most accurate way of doing that would be many runs on How Stuff Breaks and just keep track of the Andy files. Since they have such a high percentage, differences would be most noticeable and you could easily get statistically valid numbers with 12 runs since there are so many, not to mention it’s a lower difficulty and can be pwnd faster.

Do this for Luck from 1-20 and see if there is a curve if possible see the entire curve, but I’m going to focus on drop stats for everything else before I work something like this. At least until my low level placeholders get to a high enough level to use their visits for this.

The only thing this will NOT determine is if different clbumes get different drops or if any outliers where specific clbumes have gotten 2-3x the amount of a single file are just that, random outliers.

Dunatis Posted:

The fastest and probably most accurate way of doing that would be many runs on How Stuff Breaks and just keep track of the Andy files. Since they have such a high percentage, differences would be most noticeable and you could easily get statistically valid numbers with 12 runs since there are so many, not to mention it’s a lower difficulty and can be pwnd faster.

Do this for Luck from 1-20 and see if there is a curve if possible see the entire curve, but I’m going to focus on drop stats for everything else before I work something like this. At least until my low level placeholders get to a high enough level to use their visits for this.

Good suggestion.

man-man Posted:

How many runs would you recommend for gathering reasonably reliable stats?

Something in the ~1,000 range would be a high enough to be statistically significant, and 5,000 would be ideal. For those numbers, we’d need a fleet of Hackers with identical Luck pwning the same forum over and over.

How many Spiders can you equip? Even with two Spiders, with 10 Hackers on a single run in a 13 thread forum, that’s 260 drops right there. Repeat 4 times and we’re significant at 1040.