A while back, I came up with an empirical formula to do this. The forumula isn't pretty, but it should work. I fit the bastard around a lot of data points, but there are probably still a few cases that will fall through the cracks. Young players with high leadership ratings caused all kinds of trouble. Anyway, here's what I came up with:
Ldrrat is the leadership rating we are going to calculate for each player in the positional group.
Exp is the number of years experience a player has.
Ldr is the leadership rating for each player in FOF.
For guys with 1 to 3 years experience:
a1 = -9.5
a2 = -1.0
a3 = -0.38
d1 = 0.1
d2 = 0
d3 = 0
Ldrrat = Log(Exp + d) * a * ldr + ldr
For guys with 4 to 7 years experience:
Ldrrat = Log(Exp) * ((0.1 * Exp) - 0.275) * ldr + ldr
For guys with 8+ years experience:
Ldrrat = Log(Exp) * ((0.0075 * Exp) + 0.379) * ldr + ldr
The positional group leader is the guy with the highest Ldrrat. I've got an Excel sheet with some macros programmed to do the dirty work;
get it here. It also will check team chemistry and point out affinities/conflicts (that formula is a little rougher around the edges, but it's solid enough for my taste).