Thanks for the reply here.

Pretty much, this is an earlier set of rules.  It lays out the framework for it.  It has gone through multiple revisions since it first was posted.

I will say that one would be expected to play multiple games, to determine a real winner.  A second way to do this is to have it so that you want to be on the winning side at the end of the game.  You can defect late, and still remain.  People on losing side get knocked out of the game.  Eventually you get down to two people.  I have done an analog version (boardgame) using another game besides chess, and it worked.

Being on the winning side the longest can remain a tiebreaker though. What you do want this to test is the ability for people to judge the state of the game, their team, and be able recommend competent moves.  It is something that is also mean to add intrigue to any game pretty much, as spectators watching the crowd will wonder how the crowd will function.  It is also a way to have kibitzing done as an actual game.  And this is ideally suited for a game like chess and Go (abstract strategy game, perfect information, no luck).

As more play happens, more tweaks can be made.  Consider this a germ of an idea with some development behind it.  I do see you have a concern with people being punished for defecting.  I believe it shouldn't be rewarded or punished, just rewarding the best judgment.  One could end up awarding a higher score to the player who started on the losing side.  But, I don't see where the idea is to 'come back'.  It is meant you decide to jump ship before it is too late, as it is going down.