[RC5] Gaining A Statistical Advantage

Andreas Kunz andreas.kunz at ira.uka.de
Thu Dec 3 18:50:41 EST 1998

Hi everybody!

I think there are several BIG problems about our statistical thoughts
and DES-II-3 in common:

1.) perhaps most important: if this *stays* a theoretical discussion
until mid-January, it's nice, but fruitless. If the d.net people don't
want to use strategies like beginning in the middle of the keyspace or
anything, then we won't do it.
It's only about 40 days now...

2.) EFF - do they care about us? Are they interested in maximizing the
chance to have the key broken, no matter, by whom? So is there any
chance at all, to "divide up" the keyspace officially?

3.) EFF (2) - Will we know anything about their plans? Do we know how
fast their machine will be? Do they have a new cracker or did they
upgrade the old one? Will they crack at all? Do we know for sure where
they will start cracking? 
We will have increased our computing power by factor two or three. If
we'd be  *faster* then EFF this time and start at the opposite end of
the keyspace, WE would decrease our chance to be the first to crack
the key (but still increase the chance of having the key broken at

4.) time - 42 hours are a short time. Bruce Ford predicted a 42.0%
chance for d.net if we could start at once with the full computing
power and 16,9% if we started like in DES-II-2.
I think this shows very clearly the importance of a strategy to get
the clients cracking as soon as possible! I'd like to repeat that it
was better to have new clients available a week *before* DES starts
than a week after, if there will be any ;-)
But see 1.) - only the people on top of d.net can influence this.

Further comments:
robert zwaska wrote:
> However, I would submit that we should
> coordinate our attempts so that we maximize overlap. That way the group
> with the faster system stands a better chance of winning. After all,
> this is a competition.

Well, that way the group with the faster system *will* win the contest
for sure because the other team will always try blocks which have been
tried by the faster team some minutes, hours, whatever, before.
The one question is: which team will be faster?
The other is: "win the contest" stands for "find the key before the
other team", but when will this be? After 20 hours? After 60? You
remember the part with the Dollars?
With minimum overlap (e.g. starting at opposite ends of the keyspace)
the chances of each team exactly correspond to the speed. If EFF is
twice as fast as d.net, they will do the first 66% in the same time we
do the last 33%.
They have 66% chances to have the key first, we have 33%. This is a
2:1 advantage for them - the relation of our computing power. Isn't
this a fair competition, too?
The chances of having that key cracked right in time are the SUM of
both of our seperate chances. Not only the greater of both.
(I've said most of this already... perhaps not clearly enough)


The best accelerator available for a Windows98 computer 
is one that causes it to go at 9.8 m/s^2 
-------Andreas Kunz-----andreas.kunz at ira.uka.de--------

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