…to simply unplug themselves from the internet and cancel their Sirius and newspaper subscriptions.

Just watch the games and enjoy them. Or not.

He outlines the rating here. Basically, He takes Y/A (adjusted for sacks) to be the biggest factor, while adjusting it with the opponents’ average Y/A conceded (clamped to the range [5.2, 7.2]), then factors INT percentage, then TD percentage as the weakest factor. He believes the traditional rating’s emphasis on accuracy rewards the “dink and dunkers”.

I disagree here, firstly I still don’t understand why “dink and dunk” is considered undesirable a priori by so many “football minds”. Secondly, I think the biggest things a successful QB does are score and protect the ball. I also think some weight has to be given to “flipping the field”. If a drive is unsuccessful at scoring, it should at least hurt your opponent’s field position.

Example: QB A starts at his 20 yard line, tosses a 40 yard pass, the team stalls out running the ball the next three downs and no points are scored, but field position is flipped. The next two drives, however, are three-and-outs, say with two incomplete passes each.
QB B starts at his own 20 yard line, “dinks and dunks” his way to the opponent’s 40 yard line, where the drive promptly stalls, say that of the 40 yards gained on this drive he passed for 20 with 5 attempts. On the next two drives, he performs much the same, dinking and dunking for 20 more yards on 5 more attempts.

QB A has 40 yards on 5 attempts. A tremendous 8 Y/A!! QB B has 60 yards on 15 attempts, an abysmal dink-and-dunk 4 Y/A. However, in the scenario QB A completed 20% of his passes and only flipped the field on one drive. QB B helped flip the field on two of his drives with a far better completion percentage. Since neither QB scored or turned the ball over, I have to say QB B performed better because he gave the opponent worse field position more often.

I guess what all this blather is getting at is that I don’t think Y/A is the most effective way of measuring QB success. I do have to admit that using mostly scoring and turnovers would make a rating that is pretty useless for game-by-game evaluation, it would have to be used in at least several game runs.

The fact that he adjusts the rating somewhat to the opponent is interesting, and it appeals to our desire for “fairness”, if the opponent only played one game against a QB that went absolutely ape (or the opposite in Derek Anderson), this skews the rating in undesirable ways.