I have a small suggestion for the rating B) Now we have a lot of players, who won 1 Event and have the rating 1. They won f.ex. the tournament in which was 12 pairs. Players, who have won 2 events or was in the first 10 are before them... We could multiply the rating by a special "# of events"-rating. I think, it could be f(k)=(1-1/k)^k*e, where k=1+#of events, e::=LN(e)=1, so when player played a lot of events his rating is more reliable. f(1)=e/4 =~ 0.7 f(2)=0.3*e =~ 0.8 f(10)=0.35*e =~ 0.95 f(20)=0.36*e = ~0.975 f(oo)=1/e*e = 1 F.Ex. we have now 3 players: A:=1 event, rating 1.000 (1 of 12 pairs). B:=2 events, rating 0.974 (2 of 166, 9 of 186). C:=3 events, rating 0.951 (1 of 186, 14 of 70, 7 of 104). MyRating( A )=1.000*0.70 = 0.700 MyRating( B )=0.974*0.80 = 0.780 MyRating( C )=0.951*0.86 = 0.818 Player, who played 1 event can have only 0.7. Thx for read :) If you like, you can add it as a new rating :()