[robocup-legged] questionable proposed rule changes

[Vadym Kyrylov]

As a prospective participant of RoboCup 2007, I am following this discussion with great interest. 
 
I am pretty much surprised by the discussion that concentrates on how to further complicate the gaming environment rather than on improving the quality of game itself. In particular, from what I have seen in the recent 'challenge' competition in Bremen, not a single team was satifactory in such a standard exercise as passing the ball between three robots. To the best of my knowledge, the winner ended in just 1.5 points, meaning that only **one** pass was indeed successful and one more partially successful. The two runners-up have demonstrated even worse preformance; the rest got zeros. (Please correct me if I my memory has failed here.)
 
With such a low quality of ball passing, now I am wondering why the rules should be further complicated. In my opinion, too much has left to be improved in the exising setting. And if we are so concerned about impressing the RoboCup stakeholders by our progress, it looks like the improvement from 1.5 points in 2006 to say 3.0 points earned by the winner in 2007 could me very convincing evidence of the progress being made in this league.
 
I would speculate that further complication of the gaming environment for AIBOs would have same impact as moving a badminton court into the open air in a windy weather. The wind in badminton would have spoiled the play of even top performers, brinning them down to the rookie level. So would the lack of good landmarks such as sufficient number of unique beacons. An implicit proof of my speculation is the substantial number of occurences when the ball in the four-legged league is unintentially sent by robots to the outside. By the way, the number of such outs per game could be a good measure of the progress. Yet this measure would be applicable if only the rules stay unchanged over a few years. 
 
So I am for not changing anything and rather for concentrating on measurable criteria like successful number of ball passes, scoring attempts, and number of outs per game. That would make it possible to measure the progress in a convincing way.
 
Regards,
 
 
Vadim Kyrylov, PhD
Associate Professor 
Department of Applied Technology 
Rogers State University 
Claremore, OK  74017 
USA
Tel: 918-343-7953 
Fax: 918-343-7623 
 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 5567 bytes
Desc: not available
Url : http://lists.cc.gatech.edu/pipermail/robocup-legged/attachments/20060926/4bebef81/attachment-0001.bin