fane7's game - Page 5

vulture · October 8, 2003, 09:27

So you won 30 battles, and got 6 GLs. Somewhere in that you built the Heroic epic.

With no HE, you have a 1/16 chance of getting a leader per elite win. On average you expect 30/16 = 1.875 in 30 battles. The probability of getting exactly 6 is 0.93%.

With the HE, you haved a 1/12 chance, giving you an expected 30/12 = 2.5 leaders. Probability of getting 6 is 2.8%.

Very roughly speaking then, you have between a 1 and 3% chance of getting 6 leaders, depending on when you built the HE. That's not that implausible as a one off case. If it was happening regularly that would be a different matter.

sabrewolf · October 8, 2003, 09:46

well, you're right

in the following several turns with just as many elite victories i only got one more leader...

i just was surprised to get so many at once and it reminded me of fane7's game where he produced tens with just a few warriors.

just a short mathematical question: how did you calculate these figures? what kind of deviation did you use (sorry, i don't know the correct english expressions). binomial, poisson, normal?
or did you calculate all "by hand"?

Mad Bomber · October 8, 2003, 19:57

Hmmm.... So Fanes reloads does he? Now who would have thought that possible ?

vmxa1 · October 8, 2003, 20:16

Not to beat up on Fane7, but he jumped over to CFC to post his earlyt leader on their gotm23 and never came back to doc anything or finish the game.

sabrewolf · October 9, 2003, 08:20

actually i quite liked him, he always brought a smile on my face when i think of how he was trying to ignore the "are you reloading"-yes/no-questions

i also miss coracle... poly hasn't been as much fun since culture flipping is an option

vulture · October 9, 2003, 11:59

Quote:

Originally posted by sabrewolf
well, you're right

in the following several turns with just as many elite victories i only got one more leader...

i just was surprised to get so many at once and it reminded me of fane7's game where he produced tens with just a few warriors.

just a short mathematical question: how did you calculate these figures? what kind of deviation did you use (sorry, i don't know the correct english expressions). binomial, poisson, normal?
or did you calculate all "by hand"?

Used the Poisson distribution. Strictly speaking, it should be the binomial, but the Poisson is easy to do quickly, and in this case is very close (the Poisson distributiom formally allows for more than 30 leaders from 30 battles, but since the probability of getting even remotely close to 30 is so miniscule, the difference between this and reality is very small).

The normal (gaussian) distribution would be wrong for two reasons - it allows results below zero (we got -2 leaders

), and it is a continuous distribution (unlike binomial or Poisson, which deal with discrete outcomes). Oh, and it also allows more leaders than elite victories...

vulture · October 9, 2003, 12:13

Out of curiosity, I went back and re-calculated with the binomial distribution (which should be exact) for the no HE case, and got 0.752% chance for 6 leaders instead of 0.925% chance. It's a bit of a bigger difference than I expected, but it's still easily close enough for most purposes.

vmxa1 · October 9, 2003, 13:14

Quote:

Originally posted by sabrewolf

i also miss coracle... poly hasn't been as much fun since culture flipping is an option

Not sure how anyone could miss his constant repeating of the same old thing.