Streaks

Nybes

Posers and Puzzles

20 Jan 10

KazetNagorra

Germany

Joined: 27 Oct 08
Moves: 3118

08 Feb 10

Originally posted by Palynka
Boring. There are a million reasons why it would not be independent of past performance.

Sure, but it's at least feasible to correct for this one mathematically.

Palynka

Upward Spiral

Halfway

Joined: 02 Aug 04
Moves: 8702

09 Feb 10

Originally posted by KazetNagorra
Sure, but it's at least feasible to correct for this one mathematically.

Mmm... One would have to make assumptions about how the distribution of opponents depends on the player's ranking, but I suppose that's true.

Either way, until we solve the...err...simpler...problem then adding extra layers isn't going to help.

KazetNagorra

Germany

Joined: 27 Oct 08
Moves: 3118

09 Feb 10

Hasn't it been solved yet? Doesn't look so hard, but I find statistics to be rather boring so I'm not going to attempt it.

Palynka

Upward Spiral

Halfway

Joined: 02 Aug 04
Moves: 8702

09 Feb 10

1 edit

Originally posted by KazetNagorra
Hasn't it been solved yet? Doesn't look so hard, but I find statistics to be rather boring so I'm not going to attempt it.

Only numerically, but we haven't found a closed form solution. It's much harder than it looks, that's why it's a good one. It's not statistics, it's probability theory. Not the same thing.