"raylopez99" wrote in message
ups.com...
On Apr 28, 10:42 am, "David Kane" wrote:
"raylopez99" wrote in message

oups.com...



On Apr 28, 2:17 am, "David Kane" wrote:
"raylopez99" wrote in message

Thanks David Kane. My speculation about what you thought was bogus
about the current Elo rating was then largely correct--seems like it's
the sample size being too small--for a moment I thought you had some
other special insight and/or were a crank. I still find ELO quite
useful when done in a normal distribution--play more games, and you
lower the error rate. Of course if you don't play often then you can
have an erroneous rating.

BTW here is the list of soccer (football) ELO ranked champions:
http://www.eloratings.net/world.html

RL

You will note that this rating system is a modified ELO system
that considers the score as well as the result.

In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them! Of course, some hybrid
such as that used in soccer could also have merit.- Hide quoted text -

- Show quoted text -

Interesting, now I see where you're going. So, to further amplify, if
you fail to find the best move possible and, in a mating net, fail to
mate your opponent in the most efficient manner possible, you could in
theory lose points or not win as many points as somebody who mates in
X moves, as opposed to mating in 2X moves. I suppose this is
analogous to losing Elo points if you fail to win, but only draw,
against an opponent who is much weaker than you.

RL

That's an example but likely not a good one. It depends on how strongly
finding the most efficient mate is corrolated with players who win more.
The moves with the most predictive value (I'd guess) would be those
where the outcome hangs in the balance - better players will find them,
weaker players won't. But it's something that would have to be
determined empirically.

With ELO, you gain points if your performance is better
than expected by your rating. With move rating, you'd gain
points if your moves are better than expected by your rating.

I will also comment on the "Tal" argument that players
can play objectively bad moves on purpose in order to
increase their winning chances. This isn't a fundamental
objection at all - all that it means is that the analysis function
will have to be more complex. I.e. if it is determined that
playing objectively inferior moves really gives
superior winning chances, then we'd just have to quantify what
it is about those moves that makes them work, and then adjust
the calculation accordingly to give credit for them. For example,
in the lower levels of scholastic chess if you arrange your Queen
and Bishop in a battery aimed at the castled King, there is a pretty
good chance that it will pay off with a win. The 20-ply "best play"
analysis of the moves has no more relevance to the game than
the phase of the moon. Perhaps at that level we don't need Crafty
analysis at all: we can have just two dimensions: "threatens mate
in 1" and "hangs pieces."

On Apr 28, 3:55 pm, "David Kane" wrote:
"raylopez99" wrote in message

ups.com...





On Apr 28, 10:42 am, "David Kane" wrote:
"raylopez99" wrote in message

groups.com...

On Apr 28, 2:17 am, "David Kane" wrote:
"raylopez99" wrote in message

Thanks David Kane. My speculation about what you thought was bogus
about the current Elo rating was then largely correct--seems like it's
the sample size being too small--for a moment I thought you had some
other special insight and/or were a crank. I still find ELO quite
useful when done in a normal distribution--play more games, and you
lower the error rate. Of course if you don't play often then you can
have an erroneous rating.

BTW here is the list of soccer (football) ELO ranked champions:
http://www.eloratings.net/world.html

RL

You will note that this rating system is a modified ELO system
that considers the score as well as the result.

In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them! Of course, some hybrid
such as that used in soccer could also have merit.- Hide quoted text -

- Show quoted text -

Interesting, now I see where you're going. So, to further amplify, if
you fail to find the best move possible and, in a mating net, fail to
mate your opponent in the most efficient manner possible, you could in
theory lose points or not win as many points as somebody who mates in
X moves, as opposed to mating in 2X moves. I suppose this is
analogous to losing Elo points if you fail to win, but only draw,
against an opponent who is much weaker than you.

RL

That's an example but likely not a good one. It depends on how strongly
finding the most efficient mate is corrolated with players who win more.
The moves with the most predictive value (I'd guess) would be those
where the outcome hangs in the balance - better players will find them,
weaker players won't. But it's something that would have to be
determined empirically.

With ELO, you gain points if your performance is better
than expected by your rating. With move rating, you'd gain
points if your moves are better than expected by your rating.

I will also comment on the "Tal" argument that players
can play objectively bad moves on purpose in order to
increase their winning chances. This isn't a fundamental
objection at all - all that it means is that the analysis function
will have to be more complex. I.e. if it is determined that
playing objectively inferior moves really gives
superior winning chances, then we'd just have to quantify what
it is about those moves that makes them work, and then adjust
the calculation accordingly to give credit for them. For example,
in the lower levels of scholastic chess if you arrange your Queen
and Bishop in a battery aimed at the castled King, there is a pretty
good chance that it will pay off with a win. The 20-ply "best play"
analysis of the moves has no more relevance to the game than
the phase of the moon. Perhaps at that level we don't need Crafty
analysis at all: we can have just two dimensions: "threatens mate
in 1" and "hangs pieces."- Hide quoted text -

- Show quoted text -

The trouble with your system is that you can play a "beautiful" game,
full of profound moves, then, like Kramnik did against the computer
last year, miss a mate in one and lose the game. But your rating
would go up under your system (if you played, sans the one losing
move, 'above' your level).

I'm not sure the average person will understand this proposed system.

RL

"raylopez99" wrote in message
ups.com...
On Apr 28, 3:55 pm, "David Kane" wrote:
"raylopez99" wrote in message

ups.com...





On Apr 28, 10:42 am, "David Kane" wrote:
"raylopez99" wrote in message

groups.com...

On Apr 28, 2:17 am, "David Kane" wrote:
"raylopez99" wrote in message

Thanks David Kane. My speculation about what you thought was bogus
about the current Elo rating was then largely correct--seems like it's
the sample size being too small--for a moment I thought you had some
other special insight and/or were a crank. I still find ELO quite
useful when done in a normal distribution--play more games, and you
lower the error rate. Of course if you don't play often then you can
have an erroneous rating.

BTW here is the list of soccer (football) ELO ranked champions:
http://www.eloratings.net/world.html

RL

You will note that this rating system is a modified ELO system
that considers the score as well as the result.

In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them! Of course, some hybrid
such as that used in soccer could also have merit.- Hide quoted text -

- Show quoted text -

Interesting, now I see where you're going. So, to further amplify, if
you fail to find the best move possible and, in a mating net, fail to
mate your opponent in the most efficient manner possible, you could in
theory lose points or not win as many points as somebody who mates in
X moves, as opposed to mating in 2X moves. I suppose this is
analogous to losing Elo points if you fail to win, but only draw,
against an opponent who is much weaker than you.

RL

That's an example but likely not a good one. It depends on how strongly
finding the most efficient mate is corrolated with players who win more.
The moves with the most predictive value (I'd guess) would be those
where the outcome hangs in the balance - better players will find them,
weaker players won't. But it's something that would have to be
determined empirically.

With ELO, you gain points if your performance is better
than expected by your rating. With move rating, you'd gain
points if your moves are better than expected by your rating.

I will also comment on the "Tal" argument that players
can play objectively bad moves on purpose in order to
increase their winning chances. This isn't a fundamental
objection at all - all that it means is that the analysis function
will have to be more complex. I.e. if it is determined that
playing objectively inferior moves really gives
superior winning chances, then we'd just have to quantify what
it is about those moves that makes them work, and then adjust
the calculation accordingly to give credit for them. For example,
in the lower levels of scholastic chess if you arrange your Queen
and Bishop in a battery aimed at the castled King, there is a pretty
good chance that it will pay off with a win. The 20-ply "best play"
analysis of the moves has no more relevance to the game than
the phase of the moon. Perhaps at that level we don't need Crafty
analysis at all: we can have just two dimensions: "threatens mate
in 1" and "hangs pieces."- Hide quoted text -

- Show quoted text -

The trouble with your system is that you can play a "beautiful" game,
full of profound moves, then, like Kramnik did against the computer
last year, miss a mate in one and lose the game. But your rating
would go up under your system (if you played, sans the one losing
move, 'above' your level).


Huh? Players at Kramnik's level almost never miss mates in one.
That is an indication of playing *way* below his level. Beautiful
GM moves are normal for strong GMs. They're expected and hence
don't change the rating.

On Apr 28, 7:51 am, raylopez99 wrote:

Not true at all. Crafty could easily tell you which programs
far stronger than itself played the most perfect chess.

Wrong.


This is not debatable.

Wrong again.


For instance, the winning program between two chess
programs playing each other by definition will produce at least one
less error than the losing program

And again. You are completely ignoring the *magnitude* and
severity of these errors.


The only way you can get around your erroneous statement is to qualify
"properly" in "properly analyze". If you mean that it is better to
have an even stronger chess program than Crafty to better ("properly")
rate the champions, of course you're right and nobody would disagree
with you.

Still wrong. There are those who will always refuse
to admit that a computer program has sufficient chess
"understanding" to rate the world champions, though
they are gradually declining in numbers.

IMO, the fastest way to make progress here would
be to utilize the very strongest programs for this sort
of game analysis, and give them plenty of time to
look at each position -- far more than the players had.
It is also good to make full use of endgame tablebases.
One more thing: a trio (for instance) of the top-rated
programs, working in tandem, might well do a better
job of evaluating such games than any single program,
because there would be fewer oversights/misjudgments
where the program mistakenly penalizes a good move
which it simply cannot fathom.


-- help bot

On Apr 28, 4:55 pm, "David Kane" wrote:
In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them!

I was replying to this assertion by David Kane.

I don't think the average Joe will accept a rating scheme where
winning a game will lose you points ("then the winner would lose
points and the loser would gain them!")

RL

On Apr 29, 12:58 pm, help bot wrote:
On Apr 28, 7:51 am, raylopez99 wrote:

Not true at all. Crafty could easily tell you which programs
far stronger than itself played the most perfect chess.

Wrong.

This is not debatable.

Wrong again.

From your other posts bot you clearly show you are not qualified to
answer. This is over your head.

SO you are wrong.

Bye

RL

"raylopez99" wrote in message
ps.com...
On Apr 28, 4:55 pm, "David Kane" wrote:
In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them!

I was replying to this assertion by David Kane.

I don't think the average Joe will accept a rating scheme where
winning a game will lose you points ("then the winner would lose
points and the loser would gain them!")

You should go to where the data takes you without
worrying about consumer acceptance. That said,
games with a 1900 beating a 1300 have almost
zero significance in the ELO system and should
not be expected to have much significance in a move-
rating system either.

BTW, the average Joe does not accept the existing
performance-based rating scheme, if you
judge by the small numbers of people with such ratings.
I suspect a good many of those people without ELO ratings
would be happy to have their moves rated, esp. if they
see that it gives accurate answers quickly. That number
might very well dwarf the number who currently
"accept" the ELO rating scheme.

On Apr 29, 3:39 pm, "David Kane" wrote:
"raylopez99" wrote in message

ps.com...

On Apr 28, 4:55 pm, "David Kane" wrote:
In principle,a system that truly rated the moves wouldn't have
to use the result at all. For example, if a 1900 played like a 1700
but beat a 1300 playing like a 1500, then the winner would lose
points and the loser would gain them!

I was replying to this assertion by David Kane.

I don't think the average Joe will accept a rating scheme where
winning a game will lose you points ("then the winner would lose
points and the loser would gain them!")

You should go to where the data takes you without
worrying about consumer acceptance.

You are aware I hope of the theory behind voting, and that game
theorists have determined there is no such thing as a 'free and fair'
vote scheme where more than two parties exist? So you cannot say
"where the data takes you" since your scheme is normative, not based
on the laws of mathematics or science.


That said,
games with a 1900 beating a 1300 have almost
zero significance in the ELO system and should
not be expected to have much significance in a move-
rating system either.

BTW, the average Joe does not accept the existing
performance-based rating scheme, if you
judge by the small numbers of people with such ratings.
I suspect a good many of those people without ELO ratings
would be happy to have their moves rated, esp. if they
see that it gives accurate answers quickly. That number
might very well dwarf the number who currently
"accept" the ELO rating scheme.


You suspect. And I suspect otherwise. Elo's scheme is the granddaddy
of rating, and FIDE with Pres. Ill's of Soviet Muslimlands millions is
promoting ELO, not your speculative scheme. But good luck and I wish
you well. You should get a programmer to write a slick interface and
then market your scheme as shareware--it must might become a fun way
of getting an 'instant' rating from the scholastic crowd. Witness
those books that sold well that purported to rate you (Elo) based on
how accurately you solved certain chess positions. People nowadays
don't want to play 30 plus games before getting an Elo rating (the
minimum number of game needed to get a less than 5% error rate I
believe), so if your scheme correlates well with traditional Elo
rating schemes, as you imply it does, then it's a good way of getting
an 'instant' rating.

RL

"raylopez99" wrote in message
ups.com...
On Apr 29, 3:39 pm, "David Kane" wrote:


You should go to where the data takes you without
worrying about consumer acceptance.

You are aware I hope of the theory behind voting, and that game
theorists have determined there is no such thing as a 'free and fair'
vote scheme where more than two parties exist? So you cannot say
"where the data takes you" since your scheme is normative, not based
on the laws of mathematics or science.


Ratings have predictive value that can be measured objectively.

That in reality ratings are administered with consumer acceptance
in mind rather than accuracy doesn't change that. Many of
the ranking systems of which I am aware (for example, you gave
the soccer example) don't restrict themselves in the way that
chess does. Consumer acceptance doesn't seem to suffer.

In article .com,
raylopez99 wrote:
On Apr 27, 2:49 pm, "Inconnux" wrote:
Then you conceed my point and indeed the point of Crafty rating chess
players.

OK, now if I used Crafty for analysis of Fritz or Rybka games, it would
certainly not agree and would often call their moves 'errors' even though
they are FAR stronger than crafty.

To properly analyze the world champions
you would need to use a program that is atleast equal in strength to
these champions. Crippled Crafty just doesn't cut it... now if they
used Rybka for analysis I wouldn't have any problem
with the study.

J.Lohner

Not true at all. Crafty could easily tell you which programs far
stronger than itself played the most perfect chess. This is not
debatable.

Not only is it debatable, it's not true.

For instance, the winning program between two chess
programs playing each other by definition will produce at least one
less error than the losing program--and Crafty could, at some point,
appreciate this.

Er, how? If Crafty is less able than the losing program, how
can it reliably see the error the losing program couldn't?

The only way you can get around your erroneous statement is to qualify
"properly" in "properly analyze". If you mean that it is better to
have an even stronger chess program than Crafty to better ("properly")
rate the champions, of course you're right and nobody would disagree
with you. But that doesn't mean Crafty's efforts are of no value.

No, they simply aren't of enough value.

Perhaps with a 'properly' written program you might have, in a close
tie, a switch between two players say tied for fifth place in the
pantheon of all-time champions

Perhaps a 'properly' written program would completely rewrite the list
because Crafty's analysis was inadequate.

--
Christopher Mattern

NOTICE
Thank you for noticing this new notice
Your noticing it has been noted
And will be reported to the authorities