tag:blogger.com,1999:blog-38600807.post-14600560702868316272008-04-23T11:17:00.000-04:002008-04-23T11:17:00.000-04:00

I know you have little enough data to work with as it is, but to get a more accurate trend line, wouldn't it be better to discount anomalies by dropping the highest- and lowest-ranked QB's from each categoty?

Jeff Znoreply@blogger.com