Reply
 
LinkBack Thread Tools Display Modes
  #1   Report Post  
Old January 28th 04, 02:20 PM
henri Arsenault
 
Posts: n/a
Default Pandolfini and rating

In this week's column on the Chesscafe web site, Bruce Pandolfini
replies to a question about rating by saying that US ratings are about
equivalent to FIDE raitings.

For over 30 years, I have been reading that US ratings are about 100
points lower than FIDE ratings. Has something changed? Is Pandolfini
right? And if US ratings are lower, why is that? I remember back in the
old days the US Federation used to hold tournaments where one could
raise one's rating (I forgot what they were called). Could that be the
reason? Or could it be the relatively lower number of strnger players
compared to FIDE?

And BTW, how about other countries? Are Canadian ratings the same as the
US ones, or more like the FIDE ones?

maybe this should go into a separate thread, butIn the same column,
Pandolfini repeats the oft-stated idea that chess computers are not very
useful at evaluating one's games because they are tactically-oriented
and don't take into account positional considerations.

If this is true, how come no world champion player can beat them in a
match over the past five years? If the computer finds the best concrete
move, isn't that by definition also the best positional move? Especially
for 1600-rated players!? Isn't the best move the best move period?

I don't doubt that computers think differently than humans. But I doubt
that any computer can beat Kasparov by neglecting at least implicitely
one of the most important factors of chess. Aren't positional
considerations and deep analaysis two sides of the same coin, both
leading to the best move or at least to the same evaluation?

Henri
  #2   Report Post  
Old January 28th 04, 05:27 PM
David Richerby
 
Posts: n/a
Default Pandolfini and rating

henri Arsenault wrote:
In this week's column on the Chesscafe web site, Bruce Pandolfini
replies to a question about rating by saying that US ratings are about
equivalent to FIDE raitings.

For over 30 years, I have been reading that US ratings are about 100
points lower than FIDE ratings. Has something changed? Is Pandolfini
right? And if US ratings are lower, why is that?


You could turn that question on its head and ask `Why would you expect
FIDE and USCF ratings to be the same?' They're computed using (I
believe) different systems and they rate different populations.


And BTW, how about other countries? Are Canadian ratings the same as the
US ones, or more like the FIDE ones?


The British Chess Federation has a completely different rating scheme in
which a beginner is rated about 50 and a GM about 250.


Pandolfini repeats the oft-stated idea that chess computers are not very
useful at evaluating one's games because they are tactically-oriented
and don't take into account positional considerations.

If this is true, how come no world champion player can beat them in a
match over the past five years? If the computer finds the best concrete
move, isn't that by definition also the best positional move? Especially
for 1600-rated players!? Isn't the best move the best move period?


The computer plays the move that maximizes its evaluation function.
There's no reason to assume that this would be the best move according to
any (subjective) human criteria and computers are susceptible to the
so-called `horizon effect': failing to see a move because it's too far in
the future. The human can see that, eventually, that move will need to be
played.


Dave.

--
David Richerby Addictive Solar-Powered Book (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a romantic novel but it
doesn't work in the dark and you can
never put it down!
  #3   Report Post  
Old January 29th 04, 12:03 AM
RPM1
 
Posts: n/a
Default Pandolfini and rating


"David Richerby" wrote ...
henri Arsenault wrote:


Pandolfini repeats the oft-stated idea that chess computers are not very
useful at evaluating one's games because they are tactically-oriented
and don't take into account positional considerations.

If this is true, how come no world champion player can beat them in a
match over the past five years? If the computer finds the best concrete
move, isn't that by definition also the best positional move? Especially
for 1600-rated players!? Isn't the best move the best move period?


The computer plays the move that maximizes its evaluation function.
There's no reason to assume that this would be the best move according to
any (subjective) human criteria and computers are susceptible to the
so-called `horizon effect': failing to see a move because it's too far in
the future. The human can see that, eventually, that move will need to be
played.


Modern day chess programs do suffer from the horizon effect but they
have pushed the horizon pretty far. In order to take advantage of this
you have to be a pretty strong human player. For 1600-rated players I
wouldn't worry about whether computers are tactically oriented or
not. After all tactics are part of the game. There are programs that are
more positionally oriented than others. Crafty and Comet are two examples
of more positional programs. Little Goliath would probably be considered
more tactical.

Patrick



  #4   Report Post  
Old January 29th 04, 12:21 AM
Steve Grant
 
Posts: n/a
Default Pandolfini and rating

"David Richerby" wrote in message
...

computers are susceptible to the
so-called `horizon effect': failing to see a move because it's too far in
the future.


That's not quite how I'd define the horizon effect, although it's close.

Occasionally a program sees that its opponent will gain a decisive advantage
within its current search depth. At that point it will try to delay the
moment of truth by any means possible. This often takes the form of spite
checks and other common "hangnail" tactics. Once it no longer sees the
danger (because it has pushed the decisive enemy move out past its own depth
of search), it mistakenly thinks that its position is all right. It has
"successfully" moved the disaster out past its own horizon.


  #5   Report Post  
Old January 29th 04, 10:08 AM
David Richerby
 
Posts: n/a
Default Pandolfini and rating

Steve Grant wrote:
"David Richerby" wrote:
computers are susceptible to the so-called `horizon effect': failing to
see a move because it's too far in the future.


That's not quite how I'd define the horizon effect, although it's close.

Occasionally a program sees that its opponent will gain a decisive
advantage within its current search depth. At that point it will try to
delay the moment of truth by any means possible. [...]


Yes, that's a much more accurate definition; I lost a little in trying to
be brief.


Dave.

--
David Richerby Lead Tool (TM): it's like a handy
www.chiark.greenend.org.uk/~davidr/ household tool that weighs a ton!


  #6   Report Post  
Old January 30th 04, 04:38 AM
skoonj
 
Posts: n/a
Default Pandolfini and rating


"henri Arsenault" wrote in message
...
In this week's column on the Chesscafe web site, Bruce Pandolfini
replies to a question about rating by saying that US ratings are about
equivalent to FIDE raitings.

For over 30 years, I have been reading that US ratings are about 100
points lower than FIDE ratings. Has something changed? Is Pandolfini
right? And if US ratings are lower, why is that? I remember back in the
old days the US Federation used to hold tournaments where one could
raise one's rating (I forgot what they were called). Could that be the
reason? Or could it be the relatively lower number of strnger players
compared to FIDE?

And BTW, how about other countries? Are Canadian ratings the same as the
US ones, or more like the FIDE ones?

maybe this should go into a separate thread, butIn the same column,
Pandolfini repeats the oft-stated idea that chess computers are not very
useful at evaluating one's games because they are tactically-oriented
and don't take into account positional considerations.

If this is true, how come no world champion player can beat them in a
match over the past five years? If the computer finds the best concrete
move, isn't that by definition also the best positional move? Especially
for 1600-rated players!? Isn't the best move the best move period?

I don't doubt that computers think differently than humans. But I doubt
that any computer can beat Kasparov by neglecting at least implicitely
one of the most important factors of chess. Aren't positional
considerations and deep analaysis two sides of the same coin, both
leading to the best move or at least to the same evaluation?

Henri


As far as USCF ratings vs. FIDE, the letter-writer gave Pandolfini several
example of players whose USCF ratings were much higher than their FIDE
ratings. Pandolfini didn't have a response to it. I think he is wrong.

As far as computers and positional moves go, there's probably some truth to
what P. writes. Technically, the best absolute move is necessarily the best
positional move. But sometimes Fritz et. al. recommend tactical lines that
are too deep to be useful. Telling me I missed a mate in 14 isn't really
helpful becuase I know I'm never going to see that. Given the limitations of
my calculation ability, how can I decide what move to make? A computer, of
course, doesn't show you how to find the best move, it just shows you what
the best move is (well, much of the time it does). Sometimes this is
helpful, but often it really isn't.

-T


  #7   Report Post  
Old February 1st 04, 07:32 PM
Newby65
 
Posts: n/a
Default Pandolfini and rating

How would Pandolfini know about ratings, considering he hasn't played a
rated game in over 30 years?



Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On



All times are GMT +1. The time now is 11:39 AM.

Powered by vBulletin® Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Copyright 2004-2019 ChessBanter.
The comments are property of their posters.
 

About Us

"It's about Chess"

 

Copyright © 2017