The mighty DHI placed in historic context

Ed. Note: During and since this year's playoff run, the one thing everybody wrote to us about was the Defensive Hog Index. After all, it had gone 10-1 picking playoff winners in the two postseasons since its creation. People wanted to know: was this performance of the DHI part of a larger trend, had CHFF unlocked the genetic code of the gridiron, or was it just a recent statistical anomaly? Our assumption, of course, was that 10-1 seemed like an anomaly. But just how much of anomaly was the question.

 

We're working on the historic information now, as part of our wider new database that will yield tons of great info. But Mike Stickles, loyal CHFF reader and occasional contributor, ran many of the numbers for us and found some surprising results.

Good stuff, folks. Here's Mike.

 

The statistician in me cringed when I read the Defensive Hog Index touted as a "consistently reliable indicator of postseason success."

 

Two seasons' worth of playoff games just isn't enough data to properly evaluate consistency.

 

Since all the relevant data for calculating DHI is available back through the 1991 season (the NFL didn't keep third-down conversion stats before 1991), I ran the numbers for all 18 seasons and checked the results against the playoff results for each season. I also compared the performance of DHI against several other stats which I thought might have a comparable predictive value. The results were instructive and – in some instances – surprising.

 

DHI has gone 10-1 this year as a playoff game predictor, and also went 10-1 last year. But from 1991 through 2006, its record ranged from 3-8 to 8-3. For 11 of the 18 seasons, its record was between 6-5 and 7-4.

 

Overall, DHI has gone 120-75 (with 3 ties) – 0.615– predicting playoff games over those 18 seasons. Here's how that compares to simply picking the team with the best record, or the home team (treating the Super Bowls as home team ties since they have all been played at a neutral site):

 

 

Does this mean that DHI has just been lucky for two years, and will soon drop back to mediocrity? Well, not so fast there. A year-by-year analysis showed another factor might be involved.

 

The NFL changed from three divisions per conference to four four-team divisions per conference in 2002.

 

The determination of team schedules, the selection and seeding of playoff teams, and the timing of bye weeks were all affected by the new system. Looking at the year-by-year data, for virtually all of the indicators I checked, their predictive success also changed noticeably at about the same time (plus or minus a year). Here is how DHI compares to the other two indicators both before and after realignment:

 

	W	L	T	Pct (w/o ties)
Defensive Hog Index	52	24	1	0.684
Home team	42	28	7	0.600
Better record in the regular season	39	29	9	0.574

 

DHI has been by far the best predictor since realignment (I tracked nine other individual indicators and tried five other composite indices, and none did better than 0.635).

 

 

But a .684 success rate, while pretty good, isn't quite what you'd call outstanding. Then a little more investigation turned up one other important factor.

 

Some quarterbacks seem totally at the mercy of the quality of their defense. Since 2002, Eli Manning, for example, is a perfect 4-0 in the playoffs when his defense has the better Defensive Hog Index, and a perfect 0-3 when the other team's boasts the better Defensive Hogs.

 

Other quarterbacks have some ability to overcome their defense's shortcomings, but can't do it consistently. Donovan McNabb is a perfect example: he's 4-1 with superior Defensive Hogs, but just 2-3 with an inferior Defensive Hogs.

 

And a few can screw up a game even with the stronger defense behind them. Just look at Brett Favre, who managed only a mediocre 2-3 with superior Defensive Hogs (and lost his only game playing with an inferior defense).

 

 

In the eight years since realignment, 18 quarterbacks have played two or more playoff games against a team with a better DHI; 10 have played three or more such games.

 

Only two of those eighteen have won more of these contests than they've lost:

 

Tom Brady and Peyton Manning.

 

 

Manning has endured a fair bit of criticism for his tendency to wilt in the playoffs. But one reason may be his lack of defensive backup. From 2002-2008, Manning has played in 13 playoff games, but only in the first (a loss to the Jets in 2002) did he have the higher-ranked Defensive Hogs on his side.

 

Since then, he has played 12 straight games against higher-ranked Defensive Hogs, going 7-5 in those games (6-3 against quarterbacks not named Tom Brady).

 

Brady, for his part, wins consistently regardless of the quality of his defense. In the 14 playoff games he's played in that span, he has gone 7-1 with a superior defense (the only loss came against Peyton Manning) and 4-2 with an inferior defense.

 

In the 53 playoff games where neither of these two were playing, DHI has gone 40-12-1, for an 0.769 record (not including the tie). Better than three out of four – now that's outstanding.

 

Year	Def. Hog Index			Home Team			Better Record
	W	L	T	W	L	T	W	L	T
2008	10	1	0	5	5	1	3	7	1
2007	10	1	0	5	5	1	7	4	0
2006	6	5	0	8	2	1	5	3	3
2005	7	4	0	4	6	1	4	6	1
2004	6	5	0	6	4	1	7	4	0
2003	5	5	1	6	4	1	7	3	1
2002	8	3	0	8	2	1	6	2	3
2001	8	3	0	7	3	1	5	4	2
2000	7	4	0	8	2	1	7	1	3
1999	6	5	0	7	3	1	7	1	3
1998	6	5	0	8	2	1	7	2	2
1997	6	5	0	6	4	1	7	3	1
1996	7	4	0	8	2	1	9	2	0
1995	6	5	0	7	3	1	7	1	3
1994	6	4	1	8	2	1	8	2	1
1993	3	8	0	8	2	1	7	3	1
1992	6	5	0	5	5	1	5	3	3
1991	7	3	1	8	2	1	9	1	1

 

This table includes all quarterbacks who have played at least three playoff games in 2002-2008. As noted above, unless you have a great quarterback, it will be extremely difficult to overcome a better cast of Defensive Hogs. Only Brady and P. Manning have winning records in the playoffs when facing teams with super Defensive Hogs.

 

	Own team's DHI is:
	Better or equal		Worse
Quarterback	W	L	W	L
Tom Brady	7	1	4	2
Drew Brees	0	0	1	2
Mark Bulger	0	0	1	2
Jake Delhomme	4	0	1	3
Bret Favre	2	3	0	1
Joe Flacco	2	0	0	1
Rich Gannon	2	0	0	1
Jeff Garcia	1	1	1	2
Rex Grossman	2	2	0	0
Matt Hasselbeck	4	2	0	3
Brad Johnson	3	0	0	0
Eli Manning	4	0	0	3
Peyton Manning	0	1	7	5
Donovan McNabb	4	1	2	3
Steve McNair	1	2	1	1
Chad Pennington	1	0	1	4
Jake Plummer	1	2	0	1
Philip Rivers	3	1	0	2
Ben Roethlisberger	7	1	1	1
Michael Vick	1	1	1	1
Kurt Warner	2	0	1	1

 

By Kerry Byrne