NFL Wrapup: APR Pick Reliability

2008 Navigation Links

Read this for a description of APR

Over the course of the regular season, APR picked 240 NFL games week 2 to week 17. I am counting the Bengals @ Eagles tie game as a push in the following tallies.

Visitor vs Home:

	Visitor Won	Home Won	Total
Picked Right	69	90	159
Picked Wrong	34	46	80
Total	103	136	239

APR picked the home team to win 124 times (90 right + 34 wrong). Since 136 home teams won, that's pretty good evidence that some kind of factor should be added to favor home teams in generating picks.

Picks by Matchup Ratio

The matchup ratio (ρ) for each game is the quotient of the corresponding power indexes for each team (larger divided by the smaller). It is intended to be an indication of "pick confidence": ρ values near 1.0 involve teams judged to be very close in power, while large values indicate more of a mismatch.

Range			Picked Right	# Games	%	PV+HW	PV+VW
	ρ	<1.01	5	9	55.6%	2	3
1.01<=	ρ	<1.05	26	41	63.4%	9	8
1.05<=	ρ	<1.10	29	46	63.0%	11	15
1.10<=	ρ	<1.15	21	41	51.2%	9	10
1.15<=	ρ	<1.20	24	32	75.0%	5	10
1.20<=	ρ	<1.30	21	28	75.0%	5	8
1.30<=	ρ	<1.50	24	29	82.8%	3	11
1.50<=	ρ		9	13	69.2%	2	4

• "PV+HW" stands for "Picked Visitor and Home Won", and indicates the number of games in each range where APR incorrectly picked the visitor to win (and could be picked up by giving home teams more power).

• "PV+VW" stands for "Picked Visitor and Visitor Won", and indicates the number of games in each range where APR correctly picked the visitor to win (and could be lost by giving home teams more power).

• This indicates that simply giving home teams more power would cost more correctly picked games (where the visiting team wins) than it would gain.

• The drop off of accuracy for ρ >= 1.50 is due to the fact that very large values of ρ are only seen at the very beginning of the season, when APR only has a few games to produce power indexes.

NFL Wrapup: APR Strength of Schedule

2008 Navigation Links

Read this for a description of APR.

How This Was Computed

1. The unweighted version of APR (i.e., all games have the same weight) is computed for all regular season games.

2. The 16-game schedules are collated for each team.

3. For each schedule, every team on the schedule is mapped to its power index computed in step 1. This yields a list of 16 power index values for each team (divisional opponents are counted twice; no adjustment is made for home vs away games).

4. The list of power index values is then averaged, yeilding a "Strength of Schedule" power index for each team.

As usual, a higher power index value indicates a more difficult schedule.

The Schedule Power Table

Ordered from weakest schedule (top) to strongest schedule (bottom).

Schedule Power	Team	Playoffs?
1.239	Buffalo Bills
1.244	San Francisco 49ers
1.249	New York Jets
1.251	Miami Dolphins	AFC #3
1.252	New England Patriots
1.262	Arizona Cardinals	NFC #4
1.267	Seattle Seahawks
1.268	Denver Broncos
1.278	Atlanta Falcons	NFC #5
1.281	San Diego Chargers	AFC #4
1.283	New Orleans Saints
1.283	Tampa Bay Buccaneers
1.283	Tennessee Titans	AFC #1
1.284	St. Louis Rams
1.290	Carolina Panthers	NFC #2
1.293	Washington Redskins
1.293	Chicago Bears
1.294	New York Giants	AFC #1
1.295	Philadelphia Eagles	NFC #6
1.296	Green Bay Packers
1.297	Oakland Raiders
1.299	Kansas City Chiefs
1.302	Dallas Cowboys
1.303	Minnesota Vikings	NFC #3
1.311	Baltimore Ravens	AFC #6
1.314	Indianapolis Colts	AFC #5
1.315	Houston Texans
1.321	Jacksonville Jaguars
1.328	Detroit Lions
1.328	Pittsburgh Steelers	AFC #2
1.344	Cincinnati Bengals
1.346	Cleveland Browns

Comments:

• The range of values is actually pretty compact; relatively small differences in schedules can make a big difference in where a team ends up (c.f. the Titans compared to the rest of the AFC South).

• Every AFC East team is in the weakest 8, and all are weaker the Arizona Cardinals and the San Diego Chargers.

• Every NFC West team except the St. Louis Rams is also in the weakest 8.

• Every AFC North team is in the strongest 8.

• Every AFC South team except the Tennessee Titans is also in the stronget 8.

• There's no obvious correlation between schedule power and playoff appearance—the playoff teams seem to be scattered pretty evenly; and both the #1 seeds and the NFC #2 seed are pretty near the middle of the pack.

• The Browns and Bengals placement make it hard for the Lions to argue that they went winless because of injuries, a tough schedule, and a bad team (though to be fair, the Browns got their four wins before Derek Anderson and Brady Quinn were both on IR).

NFL Wrapup: APR Picks by Team

2008 Navigation Links

Read this for a description of APR.

APR's pick standings by team for the 2008 season. Because there are no power indexes until after the first week of play, only 15 weeks of play were picked.

Record	Team(s)	Comments
15-0-0	Lions	Unsurprisingly, APR works best on bad teams that lose a lot, and good teams that win a lot.
14-1-0	Seahawks	ibid
12-2-1	Eagles	12 correctly-picked games is a strong argument that the Eagles deserved their high rank after all (so is making it to the NFC Championship game).
12-3-0	Chiefs, Patriots, Cardinals, Rams	Excellent results, considering the Patriots went from #24 (week 6) to #6 (week 17), and the Cardinals went from #5 (week 5) to #24 (week 16).
11-4	Cowboys, Chargers, 49ers, Titans	The Chargers are another team that seemed to be over-rated based on their record, but 11 correct picks speaks otherwise.
10-4-1	Bengals	They are who APR thought they were.
10-5-0	Ravens, Bears, Jaguars, Steelers	16 teams were picked 10-5 (66.7%) or better, which is pretty good...
9-6-0	Vikings, Browns, Texans, Panthers, Raiders, Saints, Bills, Giants, Colts, Dolphins	...but that means 16 teams were picked 9-6 (60%) or worse, which is not so good.
8-7-0	Packers, Jets, Falcons, Redskins	The Packers (#5 to #18), Jets (#4 to #19), and Redskins (#1 to #22) all had late season collapses, and APR did not adjust quickly enough.
7-8-0	Buccaneers	The Bucs did not travel well. Through week 15, they were 6-0 at home, but just 3-5 on the road. And then, for good measure, they dropped their last two home games.
6-9-0	Broncos	The Broncos win APR's "Upset King of 2008" award. They were 5-4 against teams 8-8 or better, but just 3-4 against teams with losing records, including an inexcusable loss to the 2-14 Chiefs. How do you pick a team like that?

Comments:

• APR uses a simple exponential decay rate to discount older games. I suspect that this gives early games too much weight and late games too little. A more S-shaped curve will probably yield better results.

• Under the current configuration, APR's worst per-team record is for the 1987 Seahawks and 1994 Bills (both 2-13).

NFL: Super Bowl Picks

2008 Navigation Links

Read this for a description of APR.
Read this for a description of SRS.

Picks for the Super Bowl

Based on the Championship rankings. Line value is taken from yahoo.com, and is for comparison purposes only.

Game	APR	SRS	Line
Pittsburgh Steelers @ Arizona Cardinals	Steelers ρ=1.1191	Steelers δ=9.48	Steelers -7

Comments:

• FWDP ranks the AFC North well above the NFC West, also indicating a Steelers victory. FWDP was a lot more interesting last year when it indicated a win for the (otherwise big underdog) Giants.

• The Steelers and Cardinals played last year in week 4. That game resulted in the Steelers' first loss of the season.

• Kurt Warner has 770 yards passing so far in the postseason. 419 of those yards are on passes to Larry Fitzgerald.

• That's better than 12 times the yards of any other Cardinal receiver except Anquan Boldin (106 yards) and Steve Breaston (77 yards).

• The Steelers need to take advantage if and when the Cardinals turn the ball over. Turnovers were the difference in the Cardinals' regular season losses to the Redskins, Jets, and Panthers. The immediate corollary is the Steelers need to pressure Kurt Warner all game.

• The Cardinals have another big weakness: their offense seems to get really conservative when they have much of a lead. In the Divisional-round playoff game against the Panthers, it didn't matter, because Jake Delhomme kept throwing interceptions. But in the Conference Championship Game, the Cardinals let a 18-point lead slip away. Even in the 4th quarter when they were driving for the go-ahead score, they seemed to be content to play for a field goal (which would have given them a very marginal 2-point lead with nearly 3 minutes left for the Eagles to answer).

  With first and goal from the 9, the Cardinals ran on first and second down, gaining just one yard. On 3rd and goal from the 8, the Cardinals settled for what appeared to be another give-up play: a screen pass to running back Tim Hightower. Hightower actually caught the ball outside the 12 yard line. An Eagle defender put a solid hit on Hightower at the 2 yard line, and that should have been the end of the play. But somehow no one wrapped up on Hightower, and he was able to spin away and into the end zone. It was almost an afterthought that the Eagles then also gave up a two-point conversion, which meant a 7-point touchdown would force overtime, instead of clinching a win.

NFL: Playoff Seeds of Super Bowl Teams

The playoff seeding system was started in 1975. Since then, every Super Bowl has featured at least one #1- or #2-seeded team.

In the following table, #1 and #2 seeded teams are bolded. Years marked with a '*' are ones with no #1 seeded team. As usual, the year refers to when the regular season started; the actual Super Bowls are played the following calendar year.

Year	Winner	Loser
1975	Steelers #1	Cowboys #4
1976	Raiders #1	Vikings #1
1977	Cowboys #1	Broncos #1
1978	Steelers #1	Cowboys #2
1979	Steelers #2	Rams #3	*
1980	Raiders #4	Eagles #2	*
1981	49ers #1	Bengals #1
1982	Redskins #1	Dolphins #2
1983	Raiders #1	Redskins #1
1984	49ers #1	Dolphins #1
1985	Bears #1	Patriots #5
1986	Giants #1	Broncos #2
1987	Redskins #3	Broncos #1
1988	49ers #2	Bengals #1
1989	49ers #1	Broncos #1
1990	Giants #2	Bills #1
1991	Redskins #1	Bills #1
1992	Cowboys #2	Bills #4	*
1993	Cowboys #1	Bills #1
1994	49ers #1	Chargers #2
1995	Cowboys #1	Steelers #2
1996	Packers #1	Patriots #2
1997	Broncos #4	Packers #2	*
1998	Broncos #1	Falcons #2
1999	Rams #1	Titans #4
2000	Ravens #4	Giants #1
2001	Patriots #2	Rams #1
2002	Buccaneers #2	Raiders #1
2003	Patriots #1	Panthers #3
2004	Patriots #2	Eagles #1
2005	Steelers #6	Seahawks #1
2006	Colts #3	Bears #1
2007	Giants #5	Patriots #1
2008	Steelers #2	Cardinals #4	*

Comments:

• There's certainly evidence in this list that high playoff seeds aren't as valuable as they used to be:

• There were 8 #1-vs-#1 matchups 1975 to 1993 (19 seasons).

• There were 0 #1-vs-#1 matchups 1994 to 2008 (15 seasons).

• Counting this year, we are on a stretch of 5 Super Bowls not won by a #1 seed. The longest such strech before 2000 was just two years (1979-1980, and 1987-1988).

NFL Championship Round Power Rankings

2008 Navigation Links
Read this for a description of APR

Rank	Team	W-L	Power Index	This week
1 (↑2)	Pittsburgh Steelers	14-4	1.5330 (LW × 1.015)	won by 9 vs Ravens
2 (↓1)	Baltimore Ravens	13-6	1.5317 (LW × 0.975)	lost by 9 at Steelers
3 (↓1)	Philadelphia Eagles	11-7-1	1.4659 (LW × 0.971)	lost by 7 at Cardinals
4 (-)	Indianapolis Colts	12-5	1.4346 (LW × 0.998)	Bye
5 (↑1)	New England Patriots	11-5	1.4307 (LW × 1.005)	Bye
6 (↓1)	New York Giants	12-5	1.4236 (LW × 0.993)	Bye
7 (-)	Tennessee Titans	13-4	1.4134 (LW × 0.996)	Bye
8 (-)	San Diego Chargers	9-9	1.3759 (LW × 1.002)	Bye
9 (↑3)	Arizona Cardinals	12-7	1.3699 (LW × 1.032)	won by 7 vs Eagles
10 (↓1)	Atlanta Falcons	11-6	1.3592 (LW × 1.003)	Bye
11 (↓1)	Minnesota Vikings	10-7	1.3427 (LW × 0.999)	Bye
12 (↓1)	Carolina Panthers	12-5	1.3336 (LW × 1.003)	Bye
13 (↑1)	Miami Dolphins	11-6	1.3032 (LW × 0.998)	Bye
14 (↓1)	Dallas Cowboys	9-7	1.3002 (LW × 0.994)	Bye
15 (-)	Houston Texans	8-8	1.2956 (LW × 0.996)	Bye
16 (-)	New Orleans Saints	8-8	1.2864 (LW × 1.000)	Bye
17 (-)	Chicago Bears	9-7	1.2534 (LW × 0.997)	Bye
18 (↑1)	New York Jets	9-7	1.2407 (LW × 1.001)	Bye
19 (↓1)	Washington Redskins	8-8	1.2350 (LW × 0.992)	Bye
20 (↑1)	San Francisco 49ers	7-9	1.2310 (LW × 1.000)	Bye
21 (↓1)	Green Bay Packers	6-10	1.2287 (LW × 0.998)	Bye
22 (-)	Tampa Bay Buccaneers	9-7	1.2241 (LW × 1.000)	Bye
23 (-)	Cincinnati Bengals	4-11-1	1.2072 (LW × 0.994)	Bye
24 (↑1)	Buffalo Bills	7-9	1.1920 (LW × 1.001)	Bye
25 (↓1)	Oakland Raiders	5-11	1.1909 (LW × 0.999)	Bye
26 (-)	Seattle Seahawks	4-12	1.1729 (LW × 1.004)	Bye
27 (-)	Jacksonville Jaguars	5-11	1.1586 (LW × 0.995)	Bye
28 (-)	Denver Broncos	8-8	1.1510 (LW × 1.000)	Bye
29 (-)	Cleveland Browns	4-12	1.1298 (LW × 0.995)	Bye
30 (-)	Kansas City Chiefs	2-14	1.0764 (LW × 0.999)	Bye
31 (-)	St. Louis Rams	2-14	1.0580 (LW × 1.003)	Bye
32 (-)	Detroit Lions	0-16	0.9995 (LW × 0.998)	Bye

Comments:

• The Steelers regain the top spot (they lost the #1 spot in week 16 after losing to the Titans).

• The Cardinals climb to the #9 spot (the highest they've been since week 9, just before their close win over the 49ers), and 15 spots higher than they were after their week 16 loss to the Patriots.

• There can't be much question that the Cardinals were holding back the second half of the season. But... why not? After they lost to the Giants, a bye was pretty well out of reach. One thing they could do is perpetuate the "same old Cardinals" stereotype. I think every single team they played underestimated them. And even now, Mike Tomlin and the Steelers have to be thinking "just how good are these guys, anyway?" Which is just about how the Cardinals want it, I would imagine.

NFL Championship Game Pick Results

2008 Navigation Links

Read this for a description of APR.
Read this for a description of SRS.
Line picks counted correct when the favored team wins; they do not have to beat the spread to count as a win.

Pick results for the Championship Games.

Eagles (APR, SRS, Line) @ Cardinals

In the first half, the Cardinals just dominated. On offense, they scored at will. On defense, they limited the Eagles to three field goal tries. The second half was a different story. The Cardinals play calling seemed to go conservative on both sides of the ball, while the Eagles started to play with urgency, scoring 3 touchdowns in 3 drives. But once the Eagles regained the lead, it was their turn to be conservative: the Cardinals put together a 14 play touchdown drive that regained the lead, and left the Eagles with just 62 seconds on the game clock to respond, and they could not.

Ravens (APR, SRS) @ Steelers (Line)

The Ravens demonstrated this season that a good defense can cover a multitude of offensive shortcomings, including the inexperience of a rookie quarterback. Holding the Steelers to just 16 points on offense wasn't quite enough; when Joe Flacco started pressing his throws, the interceptions started coming, including a game-sealing pick-6 to Troy Polamalu.

Standings:

	This week	Overall
APR:	0-2	4-6
SRS:	0-2	4-6
Line:	1-1	4-6
Unanimous:	0-1	3-5

New power rankings, and the Super Bowl pick to come.