2007 St. Louis Regional

[AdamHeard]

Originally Posted by ChrisH

As I understood the explanation from the guy who designed the system, this is exactly what is supposed to happen. The alliances in a match are supposed to be closely matched in terms of total experience. So a ten year team is paired with maybe a two year team and a rookie. The average experience would then be around four years. They would wind up laying against either a similar alliance or maybe three three or four year teams. Further, the goal of the alogrithm is to have all of the matches have very similar "average ages".

This is an interesting idea and is similar to what AYSO and some other youth sports organizations do in assigning teams. They attempt to spread out the top ranked players so that the teams are close in ability.

I think the problem might be that the distribution of team ages is not uniform. There can be huge gaps that a dumb algorithm will not be able to account for. Another factor is that there is more room for adjustment when you are working with 13 or 14 individuals than when you only have three. Or they might be trying to work with too small a "window" for an acceptable match. There are a lot of ways for something like this to go wrong.

The idea seems nice, but is pretty flawed.

The age of a team doesn't necessarily reflect how good they are.

Also, I wish the spread was a little wider than just adjacent numbers. The LA regional won't be that much fun next week if we have to play against 254 and/or 330 in every match.

I bet 1503 and 1680 are liking this system though, they will probably be playing teams they are consistently better than...