Computer Chess Club Archives


Search

Terms

Messages

Subject: re: About Fernando's, When all has been said let's talk about Rebel 11

Author: Lawrence S. Tamarkin

Date: 13:49:42 01/17/99


Fernando Villegas writes in the computer resource center...
"First thing Rebel 11 should have and that Rebel 10 has not is a tutorial
capable of putting us, the users, in conditions to get a draw once in five
games, that is, to get 10% of the points instead of the almost 0% we get now.
Facts are that the professional consumer that read these pages is, in average,
only an expert level player, at most. Being an "expert" sounds good and it is
not a bad thing, but is not enough to avoid a awful beating from the hands of
Rebel 10 (or 9 and 8) , that regularly crushes us 99% of time or more.
The kind of tutorial I see as desirable and even necessary is not only of a
higher degree of chess acumen that we currently get in specialized packages, but
one in conditions to be all the time looking over our shoulders and giving us
advice and analysis about what is happening. Something of the sort already
exists in Fritz 5 and in M-Chess 8, but I am aiming to a tutor a lot more
engaged with our needs. What about a tutor that keeps a record of our games for
the sake to analyze it looking for patterns of mistakes each time we use our
computer, even with another program? That would be entirely new. Imagine a
tutorial engine operating in the background as a virus program does, so using a
lot more computer power that what is available when we just load the chess
program. So when we were going to use Rebel 11, he already would know something
more about our playing weaknesses. And so his analysis and advices would be far
richer than just give to us a better line than that we pretended to play."

So you are saying it will take over my computer like Noton Utilities? - It will
probably conflict with everything else on the computer making it crash like
never before.  Oh, it says on the side of the box, I only need a 15 gigabyte HD
& 500 mHZ of ram.  No problem in computers - in 2003! Or maybe you are trying to
suggest that the program should be permantly active in the backround, connected
a central super powerful computer on the Internet?  Now, there's an idea for
you!?  Sort of a super chess entity if you will...

"And what about a tutor giving us a set of games or problems from his database
for us to study according our needs? That’s would be, again, really new. Let us
say you are losing games in the endgame when no queens are present anymore and
when you had an even score with the program but then you go astray; then he
would pick a chosen set of games where that conditions are met AND positions
similar to yours in order you to see what should be done. Even the tutor could
deliver a text about the issue as the mentor tutorial program does for each of
his problems."

Does not Chess Mentor already do this to a large degree.  A lot of positons are
given that you train against, the programs keeps score of how well you are
doing, and presents the position on subsequent occassions, until you solve it
flawlessly.  True, there is no playing engine (as in the Chess Assistant
training programs), and that could be a useful improvement in a future version,
but it is excellent instruction.

Graphic Metaphor
"Do we need the monster size of current databases? Maybe a handful of
professional players, at most. There are people -I am one of those people- that
never in his life has given to them more than a cursory glance. We just have not
the time and, what is more, we have not the reason to do so. Anyone believe he
will improve his playing looking at Fisher games in the Sicilian? Are we going
to play a tournament against Rubinstein to see his games and prepare ourselves
for the great day?"

Well that data can be of a very high quality, or a very low quality.  IMO, a
database of very high quality, (annotated games, games arranged in tournament
order & complete data), are the very most usful.  Another factor might be game
collections of player's that perhaps are not the over 2400 varity, but might be
in tournaments that I play in.  If all this is present & accessible from a
database like Chess Base 7, then my ability to prepare for a player, and the
tools I need for sorting out what I need on that future occassion are there too.
 So I say yes, even low rated player's need the giant databases as I see no
other way of having all this material available in case of need.  Of course
having it on various CD's or archived is the encomical way to have it. And I
think the programs' CB7 & Fritz5 are quite sufficient for the detailed study of
that data. (And you can substitute many of the other fine databases, tutorial's,
and chess playing programs, as well).

"But suppose -and I return here to the tutor capabilities issue- that from the
huge mass of games the engine selects specifically the games he thinks we need
to see, games with the same opening we use -say, after ten games-, games where
the score change against our colour as happens in our games with that colour,
etc. That would be a good reason to look at the games. Yes, you can actually do
that with the search mask, fine, but what I am talking is about motivation: you
go to the gym not because you do not know the exercises to keep fitness, but to
be compelled to do so. The same in this case: you can do the search, but
probably you forget to do it, or you do it once in ten times or you just get
bored. A tutorial capable of doing that for you immediately after a game should
be something of a plus to make an extra effort."

Here you are asking for something like mind reading from the program.  I don't
think any program(s) could simulate something like this, and if they could, it
probabley could not be done well.  And if it could be done, it would require
some special attaching electrodes to your skin - And of course, Kasparov would
accuse you of having Benjamin inside your box anyway! -:)   I'm sure a lot more
disk space would be needed too for all of the special searching & heauristic
sensing software.  I think that a chess player/student has an obligation to do a
certain amount of their own research on what they need to improve in their game.
 I don't need a program to pick for me, what it thinks my weeknesses are from
that huge tap of chess data.  It is only important that the program has the
tools necessary for me to work on my game.  Fritz5 comes the closest in this,
with its many analysis tools, and coach funtions.

Your ideas are quite idealistic for future generations of chess programs, and
please forgive me any sarcasm I gave you in critiqing them.  It is only because
I find your point of view so facinating that I gave so much thought to making an
aswerng post.  Also, I'm sure you probably formulated them after considering
what many of the currently available chess software programs already does, but
not in any one package.  Surely, that is why us inkompetent chess addicts! buy
all these different programs after all.  And who knows, perhaps the next Rebel
will even have some feature's that no one can posibly immagine even now.  But
all the feature's - No program will ever have all the features...

mrslug - the inkompetent chess software addict!






This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.