Skyrim talk:Hawk Feathers
Silver Hand chanceEdit
First there are two different type of Silver Hands. Those like LvlSilverHandAmbush1 will use LootSilverHandRandom which includes LootSilverHandIngredients. But those generated by CRSilverHandNameChangerQuest will carry one instance of LootSilverHandIngredients directly.
Now the chance for LootSilverHandIngredients to generate hawk feathers is simply 4/11*25% ~= 9% (75% chance none, 4 entries in 11 item list).
LootSilverHandRandom is more difficult, it has a 25% chance none. After passing that, we have 3 times a 9% chance each at getting one. But taken together the chances are not independent. If we already got one in the first "try" we apparently passed the 25 %chance none, and the chance for a second is now 9%, not 75%*9% like it was for the first. Not sure how to write that nicely, but still precisely - use average number of feathers received instead (around 0.22)?. Anyway, no idea how the original 19% was derived, but the fact one has to write a paragraph to justify such numbers is a bit troubling as far as "accurate and verifiable" goes. --Alfwyn (talk) 19:51, 12 February 2014 (GMT)
- Empirically testing the list LootSilverHandRandom like described at Help:Testing Leveled Lists, I arrive at 2028 to 2123 hawk feathers for 10k counts of different runs, that's nearer to an average of 0.21 . --Alfwyn (talk) 20:09, 12 February 2014 (GMT)
- I think we need to decide exactly how this all wants writing out, and then doing the maths, test like the above, or a combination of both to get correct figures. Which would be a mammoth job, for this, I think that documenting the overall number (21%) over multiple item calls is a better option and less confusing, than the technically more accurate and correct but ultimately useless split up version. I can't think of an instance where documenting the separate values would be better as anyone truly interested in those figures can look it up. I'm still not sure exactly how this translates into layman's terms though, but that's just my (probably confusing) opinion. --Kiz(email - talk) 21:36, 12 February 2014 (GMT)
- Some loot tables may have a reasonable format, would be "1 to 3 hawk feathers (average 0.2)" then. I think "(average 20%)", may be too confusing. It just doesn't match with the other data given as probability. And if we are only interested in the average, the answer will be easier: 3*4/11*25%*75% ~= 0.2045, it doesn't matter that probabilities are not independent then. --Alfwyn (talk) 23:20, 12 February 2014 (GMT)
- I think we need to decide exactly how this all wants writing out, and then doing the maths, test like the above, or a combination of both to get correct figures. Which would be a mammoth job, for this, I think that documenting the overall number (21%) over multiple item calls is a better option and less confusing, than the technically more accurate and correct but ultimately useless split up version. I can't think of an instance where documenting the separate values would be better as anyone truly interested in those figures can look it up. I'm still not sure exactly how this translates into layman's terms though, but that's just my (probably confusing) opinion. --Kiz(email - talk) 21:36, 12 February 2014 (GMT)