On Mon, 2014-11-10 at 00:55 -0800, Bryce Harrington wrote:
> Something fun was proposed: add randomization to it. So, we pick 10 out 
> of the top 25 committers. The picking is done randomly, where your 
> win-chance is proportional to your commit count (p = your #commits / 
> total #commits)

That's an interesting idea.  I wonder how the randomization would be
handled to ensure fairness?

I like the idea, and curious whether it wouldn't solve your #10 problem. "Sure, I didn't do a lot, but I WON" type of thing. Solves the imposture problem in an interesting way. I don't think that we need Vegas quality randomization, I think the RAND() function from somone's spreadsheet would be totally adequate :-)