times State of the V:EKN - October 2013

31 Oct 2013 21:23 #56058 by Boris The Blade

When any two players meet, the winner of the contest receives a sum of points from the loser (so it's a zero-sum system, so jhattara is wrong to say "more activity means higher ratings" except in regard to aberrations from new players' initial ratings)

What you fail to see is that those "aberrations" are precisely how the ELO system works for someone in your situation who cannot play enough games to reach the stationary regime. In that early regime, a good player's ELO gain is a direct function (asymptotically linear) of his number of ranked games, just like the current system and the best ELO ranking he can expect is average+, nowhere near the top.

Please Log in or Create an account to join the conversation.

More
31 Oct 2013 21:29 #56060 by fredsct

When any two players meet, the winner of the contest receives a sum of points from the loser (so it's a zero-sum system, so jhattara is wrong to say "more activity means higher ratings" except in regard to aberrations from new players' initial ratings)

What you fail to see is that those "aberrations" are precisely how the ELO system works for someone in your situation who cannot play enough games to reach the stationary regime. In that early regime, a good player's ELO gain is a direct function (asymptotically linear) of his number of ranked games, just like the current system and the best ELO ranking he can expect is average+, nowhere near the top.

No, I totally agree. But then, I was never suggesting we use the ELO system for anything like what Ginés was proposing, either. My interest in it is pretty much sheer curiosity: I'd love to see what numbers it would produce for their own sake. But it certainly wouldn't surprise me if they turn out to be useless for any purpose.

Please Log in or Create an account to join the conversation.

More
Moderators: AnkhaKraus
Time to create page: 0.074 seconds
Powered by Kunena Forum