Feb 9th American Predicition League - Time to pay the Butcher!!

[68soul]

kamil said:
@Oyvoyvoyv: That's not how you count average accuracy. You counted accuracy of "average prediction". I guess you don't think that average accuracy for Culdcept saga was 99% and you have to get 99% or more accuracy to win on this bet, do you? Also there are no situation when average accuracy is 0 (unless every one person will get 0% accuracy).

To count average accuracy you have to count accuracy for every 140 users, add them up and divide by 140. Saying that what you counted still may be useful to predict results but it's not accurate way to do so.

Are you really sure that's the way it's calculated? Because then, it's really ok, half of the gamers will be below the average, the second half above...

I thought the number used for our winnings was the accuracy of the average prediction, which is WAY different, and which is the number we find on the League page when the game is closed and finished...

And as you may remember, i've had quite a discussion with ioi about that unfamous 96% accuracy for PS3 in the first japanase prediction... with such a number, it's nearly impossible to win, of course, so that was the source of my confusion and my worries... maybe i've worried a little bit too much then... but there's no clear FAQ yet, so...