User:MindlessGames

From A KoL Wiki
Revision as of 15:50, 21 October 2009 by imported>MindlessGames (Frosty's Mug: oops, didn't catch that on preview)

My main is RoyalTonberry

Tuxedo Shirt

I am keeping this here for the time being, until I have enough data to present on the tuxedo shirt page.

As far as spading this goes, the base assumption that it adds 1-3 randomly needs to be evaluated first, then the others can happen.

If it's random 1-3, then we expect the following to happen.

martinis

  • 5s -> 50% -> 6, 7, 8 as 1/3 of 50%
  • 6s -> 50% -> 7, 8, 9 as 1/3 of 50%

after shirt:

  • 6s -> 1/6
  • 7s -> 1/3
  • 8s -> 1/3
  • 9s -> 1/6


and with blender

  • 5-6.6
  • 5s -> 20% -> 6,7,8 as 1/3 of 20%
  • 6s -> 50% -> 7,8,9 as 1/3 of 50%
  • 7s -> 30% -> 8,9,10 as 1/3 of 30%

after shirt:

  • 6s -> 6.66%
  • 7s -> 23.33%
  • 8s -> 33.33%
  • 9s -> 26.66%
  • 10s -> 10%


Ode

if Ode acts before Tux, we expect:

  • 7-8.0
  • 7s -> 50% -> 8, 9, 10 as 1/3 of 50%
  • 8s -> 50% -> 9, 10, 11 as 1/3 of 50%

after shirt:

  • 8s -> 1/6
  • 9s -> 1/3
  • 10s -> 1/3
  • 11s -> 1/6

if Ode acts after Tux, we expect:

  • 5s -> 50% -> 6, 7, 8 as 1/3 of 50%
  • 6s -> 50% -> 7, 8, 9 as 1/3 of 50%

Ode maps 6->8, 7->9, 8->11, and 9->12

after shirt:

  • 8s -> 1/6
  • 9s -> 1/3
  • 11s -> 1/3
  • 12s -> 1/6

We have observed 10s with Ode, Tux, and Blender.

Note that Blender will not affect how Ode adds adventures. Observing any 10s at all with Ode active rules out the possibility of Tux acting before Ode.
So Ode acts first!

Frosty's Mug

if Frosty acts before Tux, we expect:

  • 6s -> 50% -> 7, 8, 9 as 1/3 of 50%
  • 7s -> 50% -> 8, 9, 10 as 1/3 of 50%

then apply shirt

  • 7s -> 1/6
  • 8s -> 1/3
  • 9s -> 1/3
  • 10s -> 1/6

if Frosty acts after Tux, recall the distribution from only Tux as a modifier. Now apply Frosty's mug and get:

  • 7s -> 1/6
  • 9s -> 1/3
  • 10s -> 1/3
  • 11s -> 1/6

Disgorging and Pickpocketing

for n items all with drop rate p:
integrate over x: p*(1-p*x)^(n-1)
gives: -((1-p*x)^n) / n
evaluate 0 to 1 gives: -((1-p)^n) / n + 1/n
simplifying: 1/n * (1 - (1 - p)^n)

Castle data

it looks like you can expect:

  • +0% combats --> (75+0)/((75+0) + (25-0)*2/3) = 81.82% combats
  • +5% combats --> (75+5)/((75+5) + (25-5)*2/3) = 85.71% combats
  • +10% combats --> (75+10)/((75+10) + (25-10)*2/3) = 89.47% combats
  • +15% combats --> (75+15)/((75+15) + (25-15)*2/3) = 93.10% combats
  • +20% combats --> (75+20)/((75+20) + (25-20)*2/3) = 96.61% combats
  • +25% combats --> (75+25)/((75+25) + (25-25)*2/3) = 100% combats