Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!husc6!linus!philabs!pwa-b!mmintl!franka
From: franka@mmintl.UUCP
Newsgroups: sci.math,sci.crypt
Subject: Re: dice
Message-ID: <2108@mmintl.UUCP>
Date: Tue, 14-Apr-87 22:14:07 EST
Article-I.D.: mmintl.2108
Posted: Tue Apr 14 22:14:07 1987
Date-Received: Sat, 18-Apr-87 03:30:33 EST
References: <5712@reed.UUCP> <4397@utcsri.UUCP> <438@faline.UUCP> <2881@sunybcs.UUCP>
Reply-To: franka@mmintl.UUCP (Frank Adams)
Organization: Multimate International, E. Hartford, CT
Lines: 23
Xref: utgpu sci.math:898 sci.crypt:316

In article <2881@sunybcs.UUCP> colonel@sunybcs.UUCP writes:
>> Not being into D&D, I use a pair of ordinary 6-sided dice.  Roll one die
>> from each hand, being careful to keep them separate. If a die comes up 5
>> or 6, roll it again.
>
>This is pretty slow, considering how many 5s and 6s you'll throw.  Why
>not take 6 * red + white, subtract 7, and convert to 5 bits?  That way
>you don't need to throw again unless you throw 32-35, and 13 good throws
>suffice to generate your 64-bit key.

When generating bit patterns from 6 sided dice, I count 1 to 4 as bits 00
through 11, 5 is 0, and 6 is 1.  This gives 5/3 bits per roll, vs 4/3 from
the original method.  The Colonel's method is somewhat better, giving 20/9
bits per die; applying the same trick to the values 32-35 improves this to
7/3.

Really, though, if you do this to any extent, it is worthwhile to go out and
buy some eight-sided dice.  They are available at most any game or hobby
store these days.

Frank Adams                           ihnp4!philabs!pwa-b!mmintl!franka
Ashton-Tate          52 Oakland Ave North         E. Hartford, CT 06108