You could use a revised version of the Calhamer system.
1 point for a win
(1/n) points for a draw, where n = number of players in the draw
0
for a defeat
is the original Calhamer system. I'd add in .5*(a/b) for a survive, where a = number of centers and b = victory condition (so in an 18-16 split of Classic, the winner gets 1 point, the survivor gets .44
).
Why .5*(a/b)? Because a 2-way draw would result in .5 apiece with the draw formula and the survive formula. That means that any survival is worse than a 2-way draw -- but it also acknowledges strong second performances. While strong seconds are usually discouraged for gameplay purposes, statistically you want to see strong seconds acknowledged. More data to go off of.