I don't necessarily disagree with the idea that teams improve (minus DET who was on the list as the 32nd ranked defense for 3 consecutive years) but I don't think that says much. Maybe a couple of teams did better or worse. Detroit improved and was still last. You can't honestly expect a strong correlation when there's an equally weighted variable in play that you haven't accounted for, not to mention yards don't go on the scoreboard.
Not saying I disagree with the idea, but the article is awful.
Of course, when you have one of the worst defenses in the league, odds are that that the rest of your team isn't in stellar condition either.
What odds? I can readily identify 4 juggernaut offenses from those particular teams in the comparison years. NO, GB, NE (not in chart but mentioned), KC. Not talking good offenses, they were top of the league status. There are teams every year with poor defenses and good offenses, or poor offenses with good defenses.
As NFL fans, we have been conditioned for years to look at volume stats like yards allowed or points allowed as a measure for success in the NFL. How much someone passes or runs for can make for nice anecdotal discussions in the context of fantasy football, but has next to nothing to do with winning in the NFL.
What? Points allowed is no more useful than yards allowed?
If odds say a bad defense is indicative of the quality of an entire team, how could any team who was near the bottom of the league in points allowed (and consequently - because of "odds" - near the bottom in points scored) win anything close to the 6.5 average?
19 of 20 teams reduced their yards allowed (by an average of 827 yards, or 13.2%), and 17 of 20 teams improved their number of wins. The median win improvement is three wins, and that bodes well for Dallas in 2014.
Because I guess when the average doesn't look that good you might as well go for the middle number?
Just to recap:
1. He found no correlation between yards allowed and wins, going as far as saying it has basically nothing to do with winning.
2. Projected that correlation onto just about every other stat, including points allowed.
3. Acknowledged that offenses can perform well enough to carry a horrible defense, but claimed that defensive performance is a measuring stick for the entire team based on "odds".
4. Carried out a defensive only comparison using the previously-established worthless statistic of yards, knowing that it means nothing in terms of wins.
5. Found that teams generally improve in both the statistically worthless area that has no bearing on winning and in wins, and then picked the middle number of 3 additional wins from the list of previous outcomes.
6. Concluded that it's good news for Dallas.