Sunday Morning Quarterback

Tuesday, December 13, 2005

STAT RELEVANCE WATCH, PART THREE
- - - - - - - - - - - - - - -
There are, as they say, lies, damn lies, and statistics. The numbers mean something, yet often we know not what. Here SMQ will look at the final regular season statistics in seven major categories, who succeeded in what, and how those statistics correlated to overall success in terms of final record. SMQ does have a full-time job and does not have a high-powered supercomputer - or even, at the moment, "basic" spreadsheet technology - but his analysis will be driven as deep as his egghead, tinfoil cap curiosity and cell phone calculator will take it. That is to say, quasi-scientific at best.

PART THREE: In which categories did the best team excel? (Read Part One and Part Two)

Method: To determine which numbers have the highest correlation to success, it's wise to look at the best teams and in which categories they tended to excel. SMQ took the top 25 (from his own poll, which does not substantially differ from any of the Official Do Not Disturb polls) and averaged each team's finish in each of the seven major non-scoring statistical categories - rushing offense, passing offense, total offense, rushing defense, pass efficiency defense, total defense and turnover margin - into an average ranking for the top 25 as a whole in each category.

Top 25 - Average Rank Per Category
Rushing Defense: 25.5
Total Defense: 25.6
Pass Efficiency Defense: 27.1
Rushing Offense: 35.9
Turnover Margin: 36.4
Total Offense: 38.1
Passing Offense: 47.9

Clearly, the defensive categories come out way ahead; among the top 25, there were a number of teams with better offensive numbers than defensive - Southern Cal, Texas, Texas Tech, Louisville, Iowa, Oregon, UCLA, Wisconsin - but those eight prove to be the exception to the rather overwhelming general trend (two-thirds of the final poll) of success via defensive prowess. And in the case of USC, Louisville and especially Tecas, the defensive numbers are pretty good, too.

The tendency of good teams exhibiting more defensive success becomes even more pronounced among the top ten:

Top 10 - Average Rank Per Category
Total Defense: 15.9
Rushing Defense: 16.3
Pass Efficiency Defense: 18.7
Rushing Offense: 31.7
Total Offense: 34.4
Turnover Margin: 34.4
Passing Offense: 49.1

Eight of the top ten finished in the top 17 in total defense; only three managed the feat on offense. Only USC, Texas and Notre Dame have better overall offensive rankings than defensive in this elite group, and again, the Trojans' and Longhorns' defenses finished 39th and sixth, respectively.

The top ten also features substantial offensive mediocrity in the form of Virginia Tech (56th), Georgia (55th), LSU (64th) and Miami (61st). On the other hand, only one top ten team, Notre Dame, had a unit even sniffing the bottom half of the defensive stats (64th).

SMQ began this simple study open-minded, with no preconceived hypotheses in any direction. But combined with the tepid conclusion of Part One (top teams in each defensive category had better aggregate records than the top teams in any offensive category) and Part Two (most teams can overcome a weakness in one offensive category more successfully than they can a weakness on defense or in turnover margin), the results are significant in favor of winning with defense.

Of course, this is only a trend, and not a rule. SMQ looked only at one season's results. A for-real scientist would note that, while there is a correlation between defense and success, there is no causation; maybe the defenses are good because the team is winning (teams coming from behind - usually inferior teams - tend to be one-dimensional, etc.). Conversely, offenses may become more conservative when a team is in front and sacrifice stats for ball control and maintaining the lead. There was no effort to account for strength of schedule differences - though all but a couple of the top statistical teams hail from power conferences. It's only logical that a team has to be good in a variety of different elements to be very successful; look at USC and Texas. Part Two showed that the teams with the best all-around statistical performance, on both sides, also tended to be the teams with the best won-loss records, which no single stat category can claim without also noting a few exceptions.

But the correlation is consistent and strong: offense is great, but a team's chances of winning - in 2005, at least - go up significantly higher in relation to the strength of its defense.

Got a gripe with SMQ's methods? Let him hear it below.
- - - - - - - - - - - - - - -
Permalink

9:36 PM

Comments: Post a Comment