You are currently looking at Flamebate, our community forums. Players can discuss the game here, strategize, and role play as their characters.
You need to be logged in to post and to see the uncensored versions of these forums.
Viewing a Post
|
BirdofPrey Posted:
Techically, it is possible. Here’s a simple plan: 1) create n(n=2**k) RNG that are used in ruby with such seeds: seed of i-th RNG = currend time BITAND BITNOT (n-1) [clear last k bits] + i 2) get m(m->oo) samples 3) Run all generated RNG parallely. If samples can not be generated from some RNG —- **** it. 4) in the end only one generator will be.
but there’s a ****ing flaw —- FW is mutltiplayer game. So if we get in blackjack p-th random number, next time blackjack asks randon number, it can get not (p+1)th random number, it can get (p+t)th random number, because other players had ‘stolen’ (t-1) random numbers from RNG. Log in to see images!
ADDENDUM: to clarify, here’s a little example.
First we create 4 RNG, and they produced these sequences(every number is modulo 52): 1) 1,52,3,45,5,5 2) 28,1,37,19,3,8 3) 51,1,19,44,32,4 4) 13,42,1,19,34,32
and we have collected sample from blackack: 1,19,32. According to this we can conclude that sample is more likely to be generated with RNG#3 or RNG#4 than with RNG#1.
PS. and there is a ****ing good question – if we have card, with what random number(s) it is bumociated? Zre edited this message on 04/19/2008 10:58AM |
||||||
Posted On: 04/19/2008 10:48AM | View Zre's Profile | # |