Actually, I think the distribution of scores (as a function of game time) demonstrates what we already know with respect to man versus machine: Machines are computationally much faster, but man has more efficient algorithms. What *is* interesting is that you can get a quantitative idea of the algorithmic efficiency of the human opponent by looking at the time constraint where they perform equally well, especially since we know how fast the computer is and what algorithms it uses. I find this to be much more interesting than the Kasparov match because we have time data of differing lengths.
Or perhaps this demonstrates that the human algorithms are better suited for the human brain than the software algorithms are suited for the hardware they run on.
At 06:58 PM 7/23/98 -0700, Michelle Jones wrote:
>the news media do not seem to have noticed another milestone
>in machines immitating human intelligence has been passed with
>the world's second highest ranked chess player (vishy anand) falling
>in a match to a computer. you recall world champ kasparov lost to
>the ibm mainframe deep blue last summer, but the kicker is this:
>the mighty anand lost to a 450 mhz pc, running commercially
>available software, rebel ten.
>of the 8 game match, 4 games were blitz (game in 5 minutes)
>result: silicon 3, carbon 1.
>2 games were semiblitz (game in 15 minutes). result: silicon 1, draw 1.
>2 games were tournament speed (40 moves in 2 hours). result :carbon 1,
>draw 1. final score silicon 4, carbon 2, drawn 2. congrats silicon!