Re: XieXie Master -- a recommendation

On Feb 29, 3:55=A0am, mark <mschr...@[EMAIL PROTECTED]
> wrote:
> On Feb 27, 8:31 pm, mikeh...@[EMAIL PROTECTED]
 wrote:
>
> So you say in checkers andxiangqiwe have reached the limit of playing
abil=
ity.

of course that's not what i've said.  reread my previous posts.

> Because no one can ever play better than tinsey and Xu Yin-chuan because
t=
insey and Xu Yin-chuan would never lose a match.

chinook plays better than tinsley, but only but a marginal amount.  NC
plays better than Xu because NC never blunders.  but again, the
difference is very small.

>  They can always force a draw. That sounds unimaginative.

whether it "sounds unimaginative" or not is irrelevant.  all that
matters is if it's true or false.

> Would you say the same for international draughts and world champion
Alexe=
i Chizhov? He would also never lose to a match to a computer

draughts (i.e. 10x10) is far more complex than checkers.  furthermore,
far less research has been done into the game than its 8x8
counterpart.  however, the top draughts programs are easily on par w/
the world's top 10, but probably not on par w/ the world's #1.

> Any other games?

u mean games where humans are stronger than programs?  in that case,
shogi, go, and arimaa are the immediate ones that come to mind.

> I have not heard of anyone doing this because its suicide. If there are
no=
 GMs playing computers in checkers then why? Is it because they would
always=
 lose? Maybe I will ask this question in 1 of the checkers forums.

get back to us.  actually i can save u the time.  a google search will
tell u that:

1. checkers is dying (at the professional level).  this is why they've
implemented the 3-move random opening selection.
2. the decisive game percentage in checkers for top programs is under
5%.  that means u see a win/loss in less than 1 out of 20 games.
3. GM's dont play computers because there's no money in it.  it's as
simple as that.  of course they'll spar w/ 'em for practice, but
that's another story.

> Schaeffer will strongly solve checkers. Schaeffer says on the chinook
page=
 "We will continue our computations to eventually solve all the 3-move
openi=
ngs". Seehttp://www.cs.ualberta.ca/~chinook/news/

sorry, but "strong solved" is not equivalent to "we will solve all 3-
move openings."

> When that happens will you still say tinsey would still draw a match
with =
chinook?

if the match is short enough, yes.  if long enough such that tinsley
may blunder on an off-day, then no.  but like i said, the draw ratio
will easily exceed 95%.  easily.

> Now this is a blunder only at the perfect level of play. To see this
blund=
er causes a loss, you have look ahead more than 80 moves. Only a computer
ca=
n do that. No one could use it to their advantage so it's
> not a blunder to our human eyes.

tinsley can already look 60 moves ahead on a regular basis, so 80
should be that much of a prob for him, :-)  furthermore, these deep
tactical shots only occur when u blunder in the opening.  what
evidence do u have tinsley would do such a thing?