SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed.

Pro-Football-Reference.com ยป Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

Similarity Scores for 2009 teams, part I

Posted by Jason Lisk on October 20, 2009

I'm going to do some team similarity scores like what Doug did a couple of years ago. I'm not going to use the same methodology (not that either is better than the other), because I'm not going to look at specific game results. Rather, I am going to look at a team's overall profile, in terms of wins/losses, points scored and allowed, and yardage for and against.

Like Doug, I'm not putting a tremendous amount of time into deciding how to weigh each factor. I just went with adjustments that generally felt about right, and then made sure the results passed the sniff test.

Anyway, I'm looking at the the overall points scored and allowed through six games, as well as the total net passing yards and total rushing yards for and against, along with team record. Here's the method I'm using to come up with similarity scores for each team who played a 16-game schedule from 1978 to 2008:

  1. Start with 1000 points;
  2. Subtract 5 points for every 1 point difference in Points Scored;
  3. Subtract 5 points for every 1 point difference in Points Allowed;
  4. Subtract 5 points for every 1 point difference in Net Point Differential;
  5. Subtract 1 point for every 10 yard difference in Net Passing Yards (that is, passing yards with sacks yards lost included) on offense;
  6. Subtract 1 point for every 10 yard difference in Net Passing Yards Allowed on defense;
  7. Subtract 1 point for every 10 yard difference in Rushing Yards on offense;
  8. Subtract 1 point for every 10 yard difference in Rushing Yards Allowed on defense;
  9. Subtract 1 point for every 10 yard difference in Net Passing Yards Differential (Offensive Yards minus Defensive Yards allowed);
  10. Subtract 1 point for every 10 yard difference in Rushing Yards Differential (Offensive Yards minus Defensive Yards allowed); and
  11. Subtract 100 points for every difference of 1 win.

I'm going to just quote part of what Doug wrote, because it holds true here.

In some sense, this exercise is just a whole lot of work to get (I'm assuming something very close to) the same results you'd get by running a simple regression of wins versus [scoring margin and yardage differentials]. But I like this method better, because it's not a black box.

You say the Bears should expect to win X games this year. Your friend calls BS: haven't you seen how dominant they've looked? If regression is what you've got, it's tough to give a decent counterargument unless he understands regression. But this method lays the reasoning right out there in a crystal clear way: the 1986 Falcons won their first two games by a combined 41 points and they ended up winning 7 games. The 1994 Seahawks won their first two by scores of 28-7 and 38-9, and they finished at 6-10. That is, of course, the same kind of information that your regression was taking into account, but it's just so much more transparent here.

Also, while I doubt it's actually happening here, this method is theoretically capable of picking out subtle combinations of things that regression wouldn't tell you, because you wouldn't think to ask it.

I'll just add to that last point. While I'm expressly measuring wins, plus points and yards, plus the net differences between offense and defense in those categories, it is probably picking up additional factors. If a team has a good point differential but the yardage doesn't matchup, it might be indirectly measuring teams that have good turnover ratios and good special teams. Certain combinations may indicate a team that has played a tough or easy schedule during the first six weeks, that sort of thing. A team that has the same differential and yardage, but 2 more losses, may indicate something, whether good or bad, going forward, about what happened in some of those individual games, even though I am not directly measuring at the individual game level. Whether those have any meaning, I don't know. They may be just random pockets of information.

My method doesn't know anything about these teams, other than the series of cold, hard numbers input in the formula. It doesn't know that the Jets are quarterbacked by a rookie that has been up and down, though, eerily, three of the top ten comparable teams also featured rookie quarterbacks. It doesn't know that the Vikings were the media darlings before the first game, or that the Broncos were not expected to be this good. It just sees one 6-0 team that it favors to win more games going forward because of the low points allowed and great passing defense. It doesn't know that the same team had a historically bad defense the year before, or that there was a coaching change, or what key changes are responsible for the improved defensive play. It doesn't know that the 2009 Patriots have a future Hall of Famer at quarterback, even though six of the top ten comparable teams were also quarterbacked by a current or future Hall of Famer (in years they did not win a Super Bowl, by the way). It doesn't know that Pittsburgh is the defending Super Bowl champion.

I'm going to start today with all the teams that have played six games, and have won at least three of them. Later this week, I will post the teams that have lost more than half their games and have played six, and next week, I will add the remaining twelve that have had a bye to this point, after they have also played six games.

For each team, I list the weighted average wins of the top ten comparables over the last ten games of the season. I also give a playoff chances percentage. This is also weighted on comparables, and based on the percentage chance of making the playoffs with each win total, based on 1990-2008, when the league has had six playoff teams in each conference. No team has made it with seven or fewer wins, so I assigned 0 to any win total under 8. Teams with 8 wins have made the playoffs 12% of the time, while teams with 11 wins have made it 98% of the time, and no team with more than 11 wins has failed to make the playoffs.

As a result of this, you see some interesting results where some teams have a higher average expected win total, but a lower playoff percentage, depending on the variance of their comparables. I list the teams in descending order of playoff chance percentage. The first column is the similarity score, the second is the comparable team, and the last two columns list the comparable team's record in the first six games, and last ten games.

2009 DENVER BRONCOS
weighted wins (last 10): 6.77
playoff chances: 99%

==============================================================
809	PHI	1981		6-0		4-6
763	NYG	1986		5-1		9-1
747	NOR	1991		6-0		5-5
736	NYG	1990		6-0		7-3
733	CHI	1988		5-1		7-3
723	OTI	2008		6-0		7-3
714	CLE	1994		5-1		6-4
708	JAX	1999		5-1		9-1
706	RAM	1978		6-0		6-4
703	PIT	1978		6-0		8-2
==============================================================

MINNESOTA VIKINGS
weighted wins (last 10): 5.89
playoff chances: 90%

==============================================================
850	KAN	2003		6-0		7-3
779	DAL	1983		6-0		6-4
769	MIN	2003		6-0		3-7
762	MIA	2002		5-1		4-6
756	SDG	1994		6-0		5-5
755	WAS	1983		5-1		9-1
727	DEN	1986		6-0		5-5
712	CIN	1988		6-0		6-4
688	CHI	1985		6-0		9-1
671	RAM	1988		5-1		5-5
==============================================================

NEW YORK GIANTS
weighted wins (last 10): 6.08
playoff chances: 84%

==============================================================
804	NWE	1980		5-1	                6-4
781	MIA	2002		5-1		4-6
771	RAM	1980		4-2		7-3
757	DAL	1986		4-2		3-7
754	MIN	1992		5-1		6-4
744	WAS	1983		5-1		9-1
729	DAL	1995		5-1		7-3
724	JAX	1997		5-1		6-4
724	SFO	1983		4-2		6-4
715	PIT	1979		5-1		7-3
==============================================================

NEW ENGLAND PATRIOTS
weighted wins (last 10): 6.27
playoff chances: 80%

==============================================================
827	BUF	1981		4-2		6-4
823	SFO	1986		4-2		6-3-1
806	KAN	1990		4-2		7-3
782	GNB	2001		4-2		8-2
769	RAM	2003		4-2		8-2
767	BUF	1992		4-2		7-3
764	SFO	1996		4-2		8-2
764	DAL	1999		4-2		4-6
752	SFO	1995		4-2		7-3
747	SDG	2001		4-2		1-9
==============================================================

BALTIMORE RAVENS
weighted wins (last 10): 6.08
playoff chances: 57%

==============================================================
904	DAL	2006		3-3		6-4
865	DEN	2000		3-3		8-2
830	GNB	2003		3-3		7-3
819	SDG	2008		3-3		5-5
806	NOR	2008		3-3		5-5
798	PHI	2008		3-3		6-3-1
791	RAV	1997		3-3		3-6-1
787	CHI	1995		4-2		5-5
783	DEN	1985		4-2		7-3
776	PHI	1988		3-3		7-3
==============================================================

CINCINNATI BENGALS
weighted wins (last 10): 5.39
playoff chances: 57%

==============================================================
916	BUF	1994		4-2		3-7
915	DAL	1991		4-2		7-3
870	CLE	1983		4-2		5-5
852	ATL	1995		4-2		5-5
840	KAN	1997		4-2		9-1
838	DET	2000		4-2		5-5
834	NWE	1984		4-2		5-5
831	CLT	1995		4-2		5-5
830	DET	1999		4-2		4-6
829	PHI	1996		4-2		6-4
==============================================================

PITTSBURGH STEELERS
weighted wins (last 10): 4.79
playoff chances: 45%

==============================================================
903	DAL	2005		4-2		5-5
848	ATL	1994		4-2		3-7
848	PHI	1994		4-2		3-7
841	RAI	1985		4-2		8-2
826	PHI	2005		4-2		2-8
810	PIT	1981		4-2		4-6
810	WAS	2005		4-2		6-4
806	NYG	1985		3-3		7-3
791	NOR	2001		4-2		3-7
789	GNB	1995		4-2		7-3
==============================================================

NEW YORK JETS
weighted wins (last 10): 5.39
playoff chances: 39%

==============================================================
897	PHI	1978		3-3		6-4
866	MIA	1983		3-3		9-1
841	NYJ	1991		3-3		5-5
831	RAV	2008		3-3		8-2
824	CLT	1989		3-3		5-5
818	MIN	1978		3-3		5-4-1
816	SDG	1990		2-4		4-6
813	CLE	2004		3-3		1-9
809	NOR	2000		3-3		7-3
802	SEA	1993		3-3		3-7
==============================================================

JACKSONVILLE JAGUARS
weighted wins (last 10): 4.39
playoff chances: 35%

==============================================================
839	NOR	1985		3-3		2-8
828	WAS	2003		3-3		2-8
821	NYG	1988		3-3		7-3
809	NYJ	2006		3-3		7-3
784	NOR	1998		3-3		3-7
772	TAM	1984		3-3		3-7
769	SEA	2006		4-2		5-5
764	MIA	1989		3-3		5-5
753	PHI	1995		3-3		7-3
750	BUF	1997		3-3		3-7
==============================================================

HOUSTON TEXANS
weighted wins (last 10): 5.43
playoff chances: 31%

==============================================================
876	HTX	2004		3-3		4-6
847	HTX	2007		3-3		5-5
840	WAS	1988		3-3		4-6
839	KAN	1985		3-3		3-7
775	RAM	2004		2-4		6-4
773	NWE	2002		3-3		6-4
764	PIT	1995		3-3		8-2
757	MIN	1985		3-3		4-6
745	CLT	2008		3-3		9-1
741	PIT	1984		3-3		6-4
==============================================================

This entry was posted on Tuesday, October 20th, 2009 at 8:20 am and is filed under Statgeekery. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.