Main
Date: 28 Jan 2008 13:36:48
From: pialogue
Subject: Grandmaster or Highest Rating by Draw?
Theoretically speaking, if my personal goal was to achieve a draw in
EVERY game that I played (as stupid or ill-advised as that may sound
to most people), then what is the highest chess rating that I could
possibly achieve and is it theoretically possible to be a Grandmaster
having never won OR lost a game?





 
Date: 21 Feb 2008 17:14:21
From: Wlodzimierz Holsztynski (Wlod)
Subject: Re: Grandmaster or Highest Rating by Draw?
On Feb 19, 7:39 pm, [email protected] wrote:

> It obviously is impossible to achieve a higher
> rating than any #1 on a "live" ranking..

:-) (Obviously :-)

A drawing player may reach a rating which
is very close to the rating of the world top
rated player. Then it is enough that the top
rated player loses some games (or even draws,
or draws just one game) against a weaker
player, and the drawing player becomes the
top rated player of the world.

Wlod


 
Date: 19 Feb 2008 19:39:44
From:
Subject: Re: Grandmaster or Highest Rating by Draw?
On Jan 29, 12:31 am, Bjoern <[email protected] > wrote:
> pialogue wrote:
> > Theoretically speaking, if my personal goal was to achieve a draw in
> > EVERY game that I played (as stupid or ill-advised as that may sound
> > to most people), then what is the highest chess rating that I could
> > possibly achieve
>
> Theoretically speaking there is no upper limit. E.g. assume you are
> starting with a rating of, say, 2697. Then one draws a huge number of
> games against a higher rated opponent (e.g. in an endless match versus
> Andand=2799=difference of -102 = score probability of .36) in the same
> rating period. Then if you have n draws (and no other results) the
> rating change is n * 0.14 * 0.5, so to achieve an ELO of X one needs to
> make n=100/7*(X-2697) draws. E.g. n=4329 draws to reach X=3000 Elo. Not
> sure how anyone would manage that many games on a full length time
> control in one rating period, but in theory it is possible.
>
> > and is it theoretically possible to be a Grandmaster
> > having never won OR lost a game?
>
> Yes. As a unrated player get yourself invited to three 9-round closed GM
> tournaments with the opponents rated 2700 on average in each tournament,
> then draw all games. That should do the job.

In theory, yes, but it's extremely unlikely due to the short windows
in the rating period. It obviously is impossible to achieve a higher
rating than any #1 on a "live" ranking..


 
Date: 29 Jan 2008 16:26:12
From: pialogue
Subject: Re: Grandmaster or Highest Rating by Draw?
Thank you all for your journeys into these musings. :o)

Yes, it appears that at some point a middle ground is reached and the
draw person obtains demonstrated equality with the master(s) without
either defeat or victory. It would appear that, under this particular
scenario, the "Draw Master" only reaches a rating as high as the
lowest grandmaster that he draws against because the other
grandmasters are maintaining their ratings relative to everyone else
that they play.

Is anyone good with mathematical equations so as to represent such an
"ascent by draw" with an equation? My initial attempt had x- >0 or as
x approaches zero where x represents "dysfunction" or as the
"dangerous" chess pieces are eliminated such that the opponent no
longer poses a threat and neither do I. Hence a trade-off of my own
"dysfunction" with an equal reduction of my opponent's dysfunction.
This is the reason for the draw in the first place. The "rating"
dynamic comes into play when "comparing" such a non-threatening
strategy against allegedly far superior opponents.

The overall result or goal is that parity is achieved without the
embarrassment or need to "save face" due to having been defeated.

In regard to the Russian chess players, and perhaps the Icelandic
players should be included in this explanation as well, it appears
that since the brain literally heats up when in active use that
alcohol may serve to "cool down" or relax the brain so that the
subconscious or higher conscious calculating power of the brain can do
its job most effectively. Might this help to explain why they are so
good at chess without worrying about ratings? To that end, has anyone
noticed or perhaps even documented the use of alcohol, and the amount
of alcohol used, by any of the better players during matches? From
ANY country for that matter?

Thanks!


 
Date: 29 Jan 2008 07:24:42
From: Taylor Kingston
Subject: Re: Grandmaster or Highest Rating by Draw?
On Jan 28, 6:31=A0pm, Bjoern <[email protected] > wrote:
> pialogue wrote:
> > Theoretically speaking, if my personal goal was to achieve a draw in
> > EVERY game that I played (as stupid or ill-advised as that may sound
> > to most people), then what is the highest chess rating that I could
> > possibly achieve
>
> Theoretically speaking there is no upper limit. E.g. assume you are
> starting with a rating of, say, 2697. Then one draws a huge number of
> games against a higher rated opponent (e.g. in an endless match versus
> Andand=3D2799=3Ddifference of -102 =3D score probability of .36) in the sa=
me
> rating period. Then if you have n draws (and no other results) the
> rating change is n * 0.14 * 0.5,

Maybe the formula has changed in recent years, but I thought the
rating change per game in the event of a draw would be something like
K * P * 0.5, where K is the rating point value of a single game and P
is the probability of the higher-rated player winning. K is usually 10
in a GM-level FIDE game, IIRC, and with a 102-point difference P =3D
0.64; therefore by drawing Anand our 2697-rated hypothetical drawing
master would gain 3.2 points per draw, and Anand would lose 3.2, which
in a 10-game match would come out to 32 Elo points won/lost.

> so to achieve an ELO of X one needs to
> make n=3D100/7*(X-2697) draws. E.g. n=3D4329 draws to reach X=3D3000 Elo.

The K * P * 0.5 formula says only 95 games would be required, though
I wonder about that (see below). With this formula are you
recalculating the ratings after each game?

> Not
> sure how anyone would manage that many games on a full length time
> control in one rating period, but in theory it is possible.

Would FIDE actually rate such a match with D, the rating difference,
remaining constant for all games? Clearly, after, say, the first 20 or
30 draws, our hypothetical drawing master has demonstrated equality
with Anand, and it would not be logical for him to keep gaining rating
points at Anand's expense while demonstrating no superiority over
him.
It would seem more logical to recalculate ratings periodically. In
that case, Anand's initially higher rating would fall, his opponent's
would rise, until they met somewhere near the middle, around 2748.
After that, further draws would result in no rating change at all. In
such a scenario, it would never be possible to reach a higher rating
than one's opponent by means of nothing but draws.

If my mathematical logic or calculations are flawed here (not
unlikely!), my apologies; I raise these points as questions more than
objections.


 
Date: 29 Jan 2008 00:31:42
From: Bjoern
Subject: Re: Grandmaster or Highest Rating by Draw?
pialogue wrote:
> Theoretically speaking, if my personal goal was to achieve a draw in
> EVERY game that I played (as stupid or ill-advised as that may sound
> to most people), then what is the highest chess rating that I could
> possibly achieve

Theoretically speaking there is no upper limit. E.g. assume you are
starting with a rating of, say, 2697. Then one draws a huge number of
games against a higher rated opponent (e.g. in an endless match versus
Andand=2799=difference of -102 = score probability of .36) in the same
rating period. Then if you have n draws (and no other results) the
rating change is n * 0.14 * 0.5, so to achieve an ELO of X one needs to
make n=100/7*(X-2697) draws. E.g. n=4329 draws to reach X=3000 Elo. Not
sure how anyone would manage that many games on a full length time
control in one rating period, but in theory it is possible.

> and is it theoretically possible to be a Grandmaster
> having never won OR lost a game?

Yes. As a unrated player get yourself invited to three 9-round closed GM
tournaments with the opponents rated 2700 on average in each tournament,
then draw all games. That should do the job.


  
Date: 28 Jan 2008 15:37:40
From: J.D. Walker
Subject: Re: Grandmaster or Highest Rating by Draw?
Bjoern wrote:
> pialogue wrote:
>> Theoretically speaking, if my personal goal was to achieve a draw in
>> EVERY game that I played (as stupid or ill-advised as that may sound
>> to most people), then what is the highest chess rating that I could
>> possibly achieve
>
> Theoretically speaking there is no upper limit. E.g. assume you are
> starting with a rating of, say, 2697. Then one draws a huge number of
> games against a higher rated opponent (e.g. in an endless match versus
> Andand=2799=difference of -102 = score probability of .36) in the same
> rating period. Then if you have n draws (and no other results) the
> rating change is n * 0.14 * 0.5, so to achieve an ELO of X one needs to
> make n=100/7*(X-2697) draws. E.g. n=4329 draws to reach X=3000 Elo. Not
> sure how anyone would manage that many games on a full length time
> control in one rating period, but in theory it is possible.
>
>> and is it theoretically possible to be a Grandmaster
>> having never won OR lost a game?
>
> Yes. As a unrated player get yourself invited to three 9-round closed GM
> tournaments with the opponents rated 2700 on average in each tournament,
> then draw all games. That should do the job.

It has been said that in certain areas of Russia, that with wealthy and
influential friends, even a chauffeur or a dog can become a highly rated
FIDE player.
--

Cordially,
Rev. J.D. Walker, MsD, U.C.


   
Date: 29 Jan 2008 07:23:20
From: Ralf Callenberg
Subject: Re: Grandmaster or Highest Rating by Draw?
29.01.2008 00:37, J.D. Walker:

> It has been said that in certain areas of Russia, that with wealthy and
> influential friends, even a chauffeur or a dog can become a highly rated
> FIDE player.

It has also been said, that in certain areas of Russia chauffeurs and
dogs *are* as strong as highly rated FIDE players.

Greetings,
Ralf