🚀 go-pugleaf

In <5dsm20$s4s@ausnews.austin.ibm.com>, babka@austin.ibm.com (Jim Babka) writes:
>Actually, I have a theory about this.  I'm not "in the know" about this - I
>don't know the people who put TD-GAMMON together - but I have noticed the
>rather blatant appearance of cheating (program has three unprotected pieces,
>only way I *won't* hit it at least once is if I roll a 1 2, and guess what I
>roll).  My theory is that this is a game that was trained via massive
>reiteration of playing against itself.  During the course of that learning,
>it also "learned" the pattern to the pseudo-random numbers used to generate
>the die rolls.  I understand that they probably change the seed from game to
>game, but depending upon the quality of the random-number generator, there can
>still be a repeating pattern, and that pattern might be built into the neural
>net (or whatever other algorithm is being used) that drives the computer play.
>
>In article <33014410.20749346@NewsServer.cc.monash.edu.au>,
>David Zuidema <David.Zuidema@sci.monash.edu.au> wrote:

I think you may be on to something there.  I have noticed that the computer's luck
with the die seems to be directlly related to the skill setting.  Possibly the higher
skill levels used more of the learning you describe.

Possibly the skill levels relate to ones ability to prevail over an increasingly
dis-honest opponent?   


>>On 11 Feb 97 18:44:51 -0500, dantonio@process.com (David D'Antonio)
>>wrote:
>>
>>>In article <01bc182f$5ab0d880$132ee926@dlane>, DLane@ContactPt.com says...
>>>>
>>>>Is a version that will work with Warp 3 available?
>>>>
>>>>Anybody know?
>>>
>>>The version that one could get for fee from IBM's ftp server works fine
>>>under Warp 3 FP17. Actually, that's no quite true as the colour of the
>>>pieces got wierd with the ATI MACH64 drivers I was using.
>>>
>>>Oh and there was alot of net speculation that the damn thing cheats! I'm
>>>not a backgammon slouch but it could trash me consistently. And it always
>>>seemed to get the die rolls it needed to cover an otherwise vulnerable
>>>position...
>>
>>I'm glad that someone else thinks the die rolls were too good to be
>>true ! I thought I was going paranoid. :-)
>>
>>________________________________________
>>David Zuidema                
>>http://www.physics.monash.edu.au/~davidz
>>
>
>
>-- 
>Jim Babka         babka@austin.ibm.com (Internet), BABKA at AUSTIN (VM)
>AIX Kernel Service and Support                     USIB6UKS at IBMMAIL
>What my employer thinks about what I say is best left unprinted...
>  If we could just get everyone to close their eyes and visualize


//=======================
// Chuck Cook
// cookc@clark.net
// www.clark.net/pub/cookc 
//=======================