Author: KarinsDad
Date: 23:00:28 02/02/99
Go up one level in this thread
On February 02, 1999 at 20:53:09, Dann Corbit wrote:
>On February 02, 1999 at 20:12:15, KarinsDad wrote:
>[snip]
>>Cristobal Digamo was on the winning team in the 1997 US Amateur Team Midwest.
>>His rating at that time was 1351. He won all 4 of his games. His opponent's
>>ratings were 1800, 2004, 1938, and 1786. The ELO predicted chances of him doing
>>this is less than 1 in 200,000 (I didn't figure it out exactly, if someone wants
>>to do that, feel free).
>I got 4e-6 (actually 3.97862227677179e-006), calculated as:
>printf("%.15g\n", win_expectancy(1800.-1351.)*
> win_expectancy(2004.-1351.)*
> win_expectancy(1938.-1351.)*
> win_expectancy(1786.-1351.));
>
>That would be 1 in 251,343.2868052 --> pretty close to your figure.
>
>Since to meet the conditions he must beat the first player AND beat the second
>AND beat the third AND beat the fourth, the probabilities are multiplied. The
>individual probabilities are:
>0.070132762551575
>0.0227768543550127
>0.0329568615673801
>0.0755739585740745
>
>>Do you really think he was a moderately high D class
>>player and defeated an expert, 2 class A players, and a class B player?
>At the time of the tournament, and according to the definitions, that is exactly
>what happened.
It's still bogus.
The "probabilities" are similar to if he beat Kasparov, and then turned around
and beat a class B player. The probabilities are that low (hence, his actual
strength was much higher).
>
>>Within 4 months, his rating was 1604, within 8 months 1667. His current rating
>>is 1780.
>This is obviously his true ability, but his ELO at the time of the wins was
>1351, none-the-less.
Actually, I wouldn't doubt that his true rating belongs up around 1900 or so.
KarinsDad
>[snip]
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.