Corrected week 6 rankings

I have adjusted how the ranking system treats games against FCS teams as initially they were counting for far more than a game against an FBS team.

The weighting given to these games when calculating secondary wins and losses is now equal to the average number of games played by FBS teams divided by the number of games FCS have played.  So as the average number of games played is just under 6 but 81  games against FCS teams have been played  the weighting is ~6/81 =.07.  Or in other words my original system was making a win or loss against a FCS count 14 times as much as an ordinary game when secondary and higher level wins and losses  are being calculated.

This has dramatically changed some teams rankings as Virginia Tech have jumped from 112 to 25th compared with the Massey consensus of 24th. Nth Carolina State from 66th to 23rd (Massey Consensus 25th)

Boise have jumped to second, TCU have dropped to third and LSU are now first.

The correlation to the massey consensus is now a respectable .961 which is 33rd best out of the 119 systems whereas previously I was in last place when you excluded the polls. So I think this adjustment has corrected the large and difficult to justify discrepencies in the original version. 
So hopefully next week my rankings will not annoy Josh as much.

Looking at the correlation between my rankings and the others the closest correlation is to the rankings of Hatch and of Wobus (correlation of 0.98).  Although these are two quite different systems and my system isn't like either although it may be closer to Wobus.

JMC