[mnet-devel] analytical solution to the FEC reliability problem

[Kyle Hasselbacher]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, Feb 24, 2003 at 03:14:35PM -0500, Zooko wrote:
>
>(Thanks to my combinatorics consultant, Amber, who explained this to me once 
>again.)
>
>Suppose there are M shares, and you need to recover any K of them in order to 
>win.

I made graphs.  Find them here:

http://www.toehold.com/~kyle/mn/

There are graphs for these values of M and K:

 M   K
128 64
 64 32
 32 16
 24 12
  8  4

The X axis of each graph is P(S), the probability of recovering one share,
but the LABELS are wrong because I'm too lazy to fix them.  The labels go
from 1 to 1000, while the actual probabilities used go from .001 to 1.

The Y axis is W, the probability of winning (getting enough shares to
recover).

I should probably put this explanation on the web page also.
- -- 
Kyle Hasselbacher | I thought my brain was the most wonderful organ in my body.
kyle at toehold.com  | Then I realized who was telling me this. -- Emo Phillips
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+WuA210sofiqUxIQRAu/fAJ0RPDFO0B1P1dHUk24z3TPsUE3M9QCgkbA0
RzoIM9eb222e+BgopSOoDAE=
=3bGr
-----END PGP SIGNATURE-----


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
mnet-devel mailing list
mnet-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mnet-devel