[RC5] Statistics of Key Distribution

Phil Gregory pgreg430 at neors.cat.cc.md.us
Mon Mar 30 23:27:40 EST 1998


> > The experiment (aka contest) isn't to crack
> >the code. Simple statistics answers that one (if placed on a graph, the
> >cracking of the code would form a bell curve. Sometimes it would solved right
> >away. Sometimes it wouldn't be solved until all the keys were checked. Most of
> >the time the key is near the half-way point in the pool of all the available
> >keys, 
> Ah, no, if that were the case we should start checking at the middle of the "random or linear" pool. Not from the start of the pool (not that 'start of randomly distributed key space' 
> means much) 
> Each key is just as likely to be the one. If we ran this project 72 quadtrillion times (for DES or rc5-56), we would approxiamtely expect each key to appear once.
> Therefore the distribution will be flat.

Yes, but if we change where we start, then the middle moves, too.

Seriously, the original poster is confusing two different levels.  If we
solved a whole bunch of these contests and graphed where in the keyspace
the "winning" key was, then it would form a bell curve--statistically, the
key tends to be found about halfway through.  Statistics, however, are of
no use in what is essentially a one-shot contest.  The key is equally
likely to fall anywhere within the keyspace.  The only thing that we can
say is that, as we check more keys, the change of finding the right key
becomes greater and greater.

Windows 95 --- The only OS with the Year 2000 bug built into its name
                                 Phil Gregory
                              phil_g at geocities.com
            (Finger pgreg430 at neors.cat.cc.md.us for my public key.)

Version: 2.6.2

To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest

More information about the rc5 mailing list