View Single Post
  #1   Spotlight this post!  
Unread 13-09-2007, 12:40
meaubry meaubry is offline
volunteer helper
FRC #6099 (Knights)
Team Role: Mentor
 
Join Date: Jun 2001
Rookie Year: 1997
Location: Shelby Twp, Mi
Posts: 784
meaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond reputemeaubry has a reputation beyond repute
Re: **FIRST EMAIL**/Updated Alliance Pairing Algorithm

Yeah - for open communication and requesting input. Great idea!

As with most experiments or trials to validate that the program works one way or another. I would like to see an outcome based on more controlled variables vs. less controlled variables.

When evaluating the end result it would be nice to see which variables have the most impact in the outcome.

For instance, the variable controlling minimum number of matches between scheduled matches - when varied from 1 through X, may impact the final outcome more than balance between red and blue or # or teams at the event.

Based on posts already documented here in this thread, it sounds like those that have played with the code believe it to yield better results than last years algorithem.

I like that, alot - but as Dave F. has pointed out, if the event algorithem czar decides to use 5 or 6, instead of 2 or 3, for variable d in the code - at a large sized event vs. a small sized event - will the customer complaints from last year re-surface?

I agree that perhaps, strongly suggested values for those variables that may make the biggest difference in a good quality result - should be considered as an alternative, short of hard coding it into the code.

Mike
Reply With Quote