Quote:
Originally Posted by arkain101
Right. Here is my translation interpretation.
If:
A is greater than B
B is greater than C
Then A is also greater than C.
Therefore, if C was 1000000, A would still be greater.
|
Perhaps the thread is too old, or the topic too obscure, but I think the point made by the OP is a basic one, and not in any way related to elementary algebra.
Wikipedia on Elo rankings
In a nutshell, Arpad Elo developed a way to rank competitors based on statistics rather than objective analysis of judges. Therefor a player with an Elo ranking of 1 (in chess) would be an absolutely horrible player, where as only five people have topped 2800.
The premise is that a player of Elo ranking one could reasonably be expected to beat a player with a ranking of two every once in a while, and two will beat three occasionally. But as the difference between ranks increases, the chance of the lower ranking player defeating the higher ranking player necessarily reaches zero.
This is an interesting statistical question that I can not begin to answer other than intuitively. Obviously, since a player's selected move depends on skill rather than chance, there will be a point at which the lower ranked player can not hope to win. It doesn't matter how many times I play Kasparov, I will never beat him. (unless of course my skill increases, changing my Elo ranking, and therefor eliminating the 0% chance of success).
However, the OP deals with chess playing programs, not people. So this adds another twist that I am further incapable of dealing with. Do chess programs randomly view all possible moves and determine the best one, or do they operate off of a set of instructions. Is there a particular difference between how a human evaluates the board and selects a move, and how a computer program does so? Do we even really KNOW how humans select moves, or do we just have anecdotal descriptions of how people THINK they come up with their intended move.
Ack, too late at night to be thinking so hard.
Nice topic
