OK,
Before this idea is dismissed offhand, please think about it first.

In baseball you have a ifetime batting average and a current bating
average. From one year to the next, a player may have a better average.
Over a career an entirely different result might appear.

What is there were "life-time" chess ratings as well as annual ratings?
Of course, I am thinking of a highly organized and well compensated
group of players here as well. Just an idea for thoughtful serious
discussion.

Rob

"Rob" writes:

What is there were "life-time" chess ratings as well as annual ratings?
Of course, I am thinking of a highly organized and well compensated
group of players here as well. Just an idea for thoughtful serious
discussion.

I don't know how to compute a lifetime rating, do you?
--
Mike Nolan

Mike Nolan wrote:
"Rob" writes:

What is there were "life-time" chess ratings as well as annual ratings?
Of course, I am thinking of a highly organized and well compensated
group of players here as well. Just an idea for thoughtful serious
discussion.

I don't know how to compute a lifetime rating, do you?
--
Mike Nolan

Mike,
The actual math involved, no; I don't know how to do it. But isnt what
someone has right now as a rating exactly that,a lifetime rating? What
would be needed to be decided upon then is an annual rating for play.
Perhaps everyone starts the first tournament with same rating(whatever
that might be) and as they play over the course of a year their annual
rating is caculated. Of course those same match scores would be used
for the lifetime scores as well.

It may not work. It might make the game a bit more exciting though.
Rob

"Sam Sloan" wrote in message
...
On 12 Jun 2005 16:05:16 -0700, "Mark Houlsby"
wrote:

Mike Murray wrote:
AFAIK, common usage generally restricts "Elo" to FIDE ratings,



Wrong. Completely wrong. Totally and utterly wrong. "Elo" refers to
*any* rating system which uses Dr. Arpad Elo's idea, so it applies
*equally* to, say, USCF ratings, FIDE ratings, SCA ratings, DWZ
ratings, ICC ratings, WCN ratings, playchess.com ratings.....


I hope you are aware that Mark Houlsby is a rated 1295 player in his
native Scotland and knows NOTHING about this subject. He just has a
loud mouth.

Sam Sloan

A high rating is not necessary to comment on ratings.

Regards

"Coiled War" wrote in message
...
"Mark Houlsby" No, quite the contrary, it's *not at all likely*. If I
was
under any such threat from people as powerful as that, I would *comply*.
If
there was *any chance* of the authorities' discovering my dissention, I'd
be
in trouble, wouldn't I?


But you are a gutless unprincipled piece of trash, unlike the kind of
people
that grew up in the previous generations, and who took many risks for the
sake of freedom, especially during the Cold War.



That applies to me too.

"David Richerby" wrote in message
...
Chess One wrote:
"Mark Houlsby" wrote:
Mike Murray wrote:
AFAIK, common usage generally restricts "Elo" to FIDE ratings,

Wrong. Completely wrong. Totally and utterly wrong. "Elo" refers to
*any* rating system which uses Dr. Arpad Elo's idea, so it applies
*equally* to, say, USCF ratings, FIDE ratings, SCA ratings, DWZ
ratings, ICC ratings, WCN ratings, playchess.com ratings.....

Exactly.


I don't think this is wrong. An ELO rating is obtained for regularly
timed chess, under regular rules, and with an arbiter present. All
these types of rating need be assessed by the same methods to be
considered ELO equivalents.

No. The chess federations of country X and country Y can both calculate
ratings using Elo's system, for games played under the same conditions but
the ratings are still not directly comparable. If there are players who
have played in both countries, we can extrapolate from their ratings to
compute the *approximate* rating that a country X player would have in
country Y but that's the best that can be done.

Not so. Look at the initial list above, which among others includes ICC
ratings - in addition to games played at full time, there must also be an
arbiter present. ICC ratings are not transposable therefore.

Postal chess is an inflated rating-medium, since the look-ups benefit the
player's strength, and is no test of their otherwise unaided play OTB. It is
also not transposable as an indicator of OTB strength.

Mathematically Elo's idea can be applied to any sport or game, but chess
ratings specifically require real-time play under arbiter supervision to be
ratable. Otherwise they are like-ratings, but not alike enough to transpose
one type [in this case correspondance] into an equivalent OTB. So to say
that one's Elo is 2300 by some transpostion method of extrapolating that
from correspondance 1800 is a nonsense.

It is true that there needs to be 'sufficient X' where X is the pool of
players. to be able to calculate Elo. But this discussion was originally
about transposing correspondance ratings [Taylor Kingstons's whose name
still survives massive thread drift] where any X factor is redundant.

Cordially, Phil


Dave.

--
David Richerby Revolting Beer (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ refreshing lager but it'll turn
your
stomach!

"Rob" writes:

The actual math involved, no; I don't know how to do it. But isnt what
someone has right now as a rating exactly that,a lifetime rating? What
would be needed to be decided upon then is an annual rating for play.
Perhaps everyone starts the first tournament with same rating(whatever
that might be) and as they play over the course of a year their annual
rating is caculated. Of course those same match scores would be used
for the lifetime scores as well.

The mathematicians on the Ratings Committee would probably disagree, but
to me the current ratings system is closer to a weighted average than
anything else, with recent results being very heavily weighted over
previous results.

It may not work. It might make the game a bit more exciting though.

I'm not sure how it could add any excitement. Please elaborate.
--
Mike Nolan

On 13 Jun 2005 11:30:50 -0700, "Rob" wrote:


I don't know how to compute a lifetime rating, do you?

The actual math involved, no; I don't know how to do it. But isnt what
someone has right now as a rating exactly that,a lifetime rating?

No it isn't. It is a kind of moving average where recent games are
rated much more heavily than older ones, and where old enough games
basically have no effect.

The present system is designed to predict future results based on past
results. A rating based on all one's games, though dead easy to
produce, would not do that since what predicts future results is present
strength.

A "lifetime" rating would have little value in predicting future
results.

We do have a system of life time titles, but no one uses the player's
title to predict the outcome of a particular game except in the very
roughest way.

A rating is designed to produce a probabilistic prediction to a faily
high level of accuracy. Lifetime ratings and titles aren't.


Ed Seedhouse,
Victoria, B.C.

Ed Seedhouse wrote:
On 13 Jun 2005 11:30:50 -0700, "Rob" wrote:


I don't know how to compute a lifetime rating, do you?

The actual math involved, no; I don't know how to do it. But isnt what
someone has right now as a rating exactly that,a lifetime rating?

No it isn't. It is a kind of moving average where recent games are
rated much more heavily than older ones, and where old enough games
basically have no effect.

The present system is designed to predict future results based on past
results. A rating based on all one's games, though dead easy to
produce, would not do that since what predicts future results is present
strength.

A "lifetime" rating would have little value in predicting future
results.

We do have a system of life time titles, but no one uses the player's
title to predict the outcome of a particular game except in the very
roughest way.

A rating is designed to produce a probabilistic prediction to a faily
high level of accuracy. Lifetime ratings and titles aren't.


Ed Seedhouse,
Victoria, B.C.

Ed,
Thank you very much. I was not aware that there was a weighting in
favor of more current games played. Does the current system lower a
players rating for inactivity?
Rob

Mike Nolan wrote:
"Rob" writes:

The actual math involved, no; I don't know how to do it. But isnt what
someone has right now as a rating exactly that,a lifetime rating? What
would be needed to be decided upon then is an annual rating for play.
Perhaps everyone starts the first tournament with same rating(whatever
that might be) and as they play over the course of a year their annual
rating is caculated. Of course those same match scores would be used
for the lifetime scores as well.

The mathematicians on the Ratings Committee would probably disagree, but
to me the current ratings system is closer to a weighted average than
anything else, with recent results being very heavily weighted over
previous results.

It may not work. It might make the game a bit more exciting though.

I'm not sure how it could add any excitement. Please elaborate.
--
Mike Nolan

Mike, I was just thinking that someone who had a high rating might not
be currently the strongest player. Yet because they were once strong,
they simply say they are the best because of their historical rating
and not their actual current match play. A situation like that opens up
the opportunity for one person to be the champion and someone else to
claim to be the champion because of an outdated rating. Not knowing
excatly how this stuff works or is caculated, I could be totally wrong
here.
Rob