xmarkd400x
2nd Team Getting Their First Start
- Joined
- Sep 17, 2007
- Messages
- 1,745
- Reaction score
- 0
Registered Members experience this forum ad and noise-free.
CLICK HERE to Register for a free account and login for a smoother ad-free experience. It's easy, and only takes a few moments.As much as I appreciate CHFF and their attempts to use actual facts in their reporting, there are some serious problems with that analysis.
Well, for one, they don't clearly define the criteria for better vs. worse. The conclusion points out some ways in which the 'Boys were more impressive than the Pats, but they arbitrarily pick other criteria as more important and decide Pats > Cowboys based on that. That's not objective analysis, it's subjective opinion based on cherry-picked quantitative data.
Their "Quality Opponents" measure is very, very problematic. It only looks at opponent records, regardless of quality of their competition, and doesn't include potentially balancing factors, such as losses vs. poor opponents. A 2-12 team could be ranked higher than a 10-4 team in that system.
There are no controls whatsoever in the measures they use to compare the '92 Cowboys vs. the '04 Patriots, even though more than a decade lies between them. As a data analyst/statistician, the first thing that jumps out is that ALL numbers are higher with the '04 Pats...yards gained AND allowed, pass/rush gained and allowed, etc. Also, despite the fact that the Pats YPG was higher than the Boys, they were ranked lower. This suggests strongly that yardage was up league-wide in 2004 compared to 1994, so comparing straight-up numbers is not going to give a helpful measure of relative ability. (If that's even what they're looking for; they don't say)
It's fine if you want to say "based on these criteria, team X was better," but one could just as easily pick other measures, insert some statistical controls, and demonstrate the opposite. In a solid statistical study, you define your criteria ahead of time, and you use statistical controls to clean up the "noise" in the data.
It concludes the Patriots were better than Dallas, so I think the analysis was spot on - perfect!
Well, for one, they don't clearly define the criteria for better vs. worse. The conclusion points out some ways in which the 'Boys were more impressive than the Pats, but they arbitrarily pick other criteria as more important and decide Pats > Cowboys based on that. That's not objective analysis, it's subjective opinion based on cherry-picked quantitative data.
Their "Quality Opponents" measure is very, very problematic. It only looks at opponent records, regardless of quality of their competition, and doesn't include potentially balancing factors, such as losses vs. poor opponents. A 2-12 team could be ranked higher than a 10-4 team in that system.
There are no controls whatsoever in the measures they use to compare the '92 Cowboys vs. the '04 Patriots, even though more than a decade lies between them. As a data analyst/statistician, the first thing that jumps out is that ALL numbers are higher with the '04 Pats...yards gained AND allowed, pass/rush gained and allowed, etc. Also, despite the fact that the Pats YPG was higher than the Boys, they were ranked lower. This suggests strongly that yardage was up league-wide in 2004 compared to 1994, so comparing straight-up numbers is not going to give a helpful measure of relative ability. (If that's even what they're looking for; they don't say)
It's fine if you want to say "based on these criteria, team X was better," but one could just as easily pick other measures, insert some statistical controls, and demonstrate the opposite. In a solid statistical study, you define your criteria ahead of time, and you use statistical controls to clean up the "noise" in the data.
if you read the article, though, he basically argues himself to a standstill listing various feats from both sides - and then concludes in favor of the Patriots. Although his analysis is interesting, I don't find his conclusion compelling.
I'll add my own two cents: I grew up rooting for the Aikman Cowboys because a couple members of my family did, and switched to the Patriots (as my hometown team) once I realized that being a bandwagoner sucks. And I can say with absolute conviction that as good as he was, Troy Aikman was no Tom Brady.
Your premise that it can be objective is impossible to begin with anyhow. They did not play each other and they did not play simultaneously.
It can only be a subjective-opinion based analysis.