Author: James B. Shearer
Date: 17:08:59 01/27/99
Go up one level in this thread
On January 26, 1999 at 18:05:29, KarinsDad wrote:
>On January 26, 1999 at 17:44:46, James B. Shearer wrote:
>
>>On January 26, 1999 at 13:49:06, KarinsDad wrote:
>>
>>>On January 26, 1999 at 01:42:38, James B. Shearer wrote:
>>>
>> <snip>
>>>
>>>>So it is conceivable that program x performs at a 2200 level
>>>>against computers on ICC but performs at a 2400 level against humans on ICC.
>>>>(Of course such programs should be balanced by other programs which are better
>>>>against computers so there is no net flow of rating points between the human and
>>>>computer pools).
>>>
>>>Hard to say. I doubt there are many "balancing" programs out there. Since anyone
>>>can play anyone as many times as they want on the internet, an 1800 player (or
>>>computer) can endlessly play a 2400 rated computer and slowly bring it's rating
>>>up (and his rating down).
>>
>> There have to be balancing programs. Consider program x. It will tend to
>>win rating points when playing humans and then lose these rating points when
>>playing computers. It is in effect "pumping" rating points from the human pool
>>to the computer pool. These points must return to the human pool somehow.
>
>Ok, we were talking about two different things. I understand why some of the
>points would return to/from the human pool, but I do not understand why the flow
>to computers from humans would equal the flow from computers to humans. I also
>do not understand why programs would "max" out either high (2400 against humans)
>or low (2200 against computers). Why is that? Is there a rating floor (similar
>to USCF)? Or does the rating system have some safeguards in it that I do not
>know of?
The flow to computers from humans will equal the flow from computers to
humans (assuming equilibrium) because that is the way the rating system works,
your rating rises or falls until it reaches a point where you are losing as many
rating points as you are winning. Such a point will exist because as your
rating rises your wins are worth fewer rating points and your losses cost more
rating points. This is true for a group of players as well as for individual
players. So the ratings of computers as a group will shift relative to the
ratings as humans as a group until the net flow between the groups is zero.
The 2200 and 2400 were the assumed strengths of program x against
computers and humans. There is nothing special about these particular values.
James B. Shearer
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.