Sunday, August 31, 2008

Some Historical Analysis of Preseason Power Rankings

I took the preseason data going back to 2001 (the last year NFL.com has quarter-by-quarter scoring data for the preseason), and ran the four preseason ranking algorithms (listed below) on the scores for each season.

Using each power ranking as a means to pick the week one games (i.e. predict the team with the highest power ranking to be the winner), and then compared the pick to the actual game winners. Below is a table summarizing the results for the data:

Year Preseason
Week 3
Pre/half
Week 3
Preseason
Week 4
Pre/half
Week 4
20014-116-96-96-9
20027-97-99-78-8
20039-712-49-77-9
20046-1012-412-412-4
20056-106-107-97-9
200610-610-67-910-6
200710-68-810-610-6
Total52-5961-5060-5160-51

I was happier with these numbers before I ran the 2001 data set. Any way you look at it, it's a pretty sparse data set to be making strong predictions. But it does look like there's reason to think preseason power rankings are useful data.

I'll be doing a more extensive analysis later. And, of course, pick results for the 2008 data will be posted next week.

Saturday, August 30, 2008

Anonymous comments enabled

In order to encourage comments, I've enabled anonymous comments.

I'll go back if the comments fill up with spam, or off-topic stuff, or profanity, but for now, have at it.

Friday, August 29, 2008

NFL Week 1 picks!

Update: Added line picks

The preseason is over!

Because APR is a feedback-based algorithm, and there is nothing to feed back before any games are played, there are no APR-based picks for week 1.

...Unless you use the preseason games. I hereby present picks for week one, based on preseason APR rankings. I've generated four power ranking tables, producing four picks for each game (detailed below). The power rankings are generated as follows:

  • Preseason Week 3: Rankings based on the final scores of all games up to and including week 3

  • Pre/Half Week 3: Rankings based on the half-time scores of all games up to and including week 3

  • Preseason Week 4: Final-score based rankings for week 4

  • Pre/Half Week 4: Half-time score based rankings for week 4

The games with the darker background are those that are not unanimous picks.

Game Preseason
Week 3
Pre/half
Week 3
Preseason
Week 4
Pre/half
Week 4
Line
WAS @ NYGWAS NYGNYGNYGNYG
CIN @ BALCIN CINCINCINCIN
NYJ @ MIAMIA MIAMIAMIANYJ
KC @ NE KC KC KC KC NE
HOU @ PITPIT PITHOUPITPIT
JAC @ TENJAC JACJACJACJAC
DET @ ATLDET DETDETDETDET
SEA @ BUFBUF BUFSEASEAPush
TB @ NO NO NO TB NO NO
STL @ PHIPHI PHIPHIPHIPHI
DAL @ CLEDAL DALDALDALDAL
CAR @ SD CAR CARSD SD SD
ARI @ SF ARI ARISF ARIARI
CHI @ INDIND INDCHICHIIND
MIN @ GB MIN MINGB MINGB
DEN @ OAKDEN OAKDENDENDEN


Are the Patriots really that bad?

The Patriots have had a forgettable preseason, finishing 0-4. What's worse, they were outscored 67-9 in the first half of play, presumably when the first- and second-string players are on the field. Unless they've been holding back in a big way, and not just with Tom Brady, it looks like the Patriots could be in for a major let-down this year. If the Patriots are anywhere near the level they played at last year, there's no way the Chiefs beat them in week 1 (right?). But it's not going to look good if KC is within a score late in the 4th quarter...

Are those other teams really that good?

On the other hand, APR (Pre/half Week 3) has the Bills, Dolphins, Panthers, Cardinals, Saints, Lions, Vikings, and Chiefs all in the top 10. Could all of these teams have had such a major turn around from last year? Could any of these teams have had such a major turn around from last year? Maybe the Pats fans' best hope is that the whole idea that preseason games mean something is just so much hot air. Tune in next week to find out...

Monday, August 25, 2008

NFL: Preseason Power, Week 3

Preseason Power Rankings?!

Yeah, preseason. I've done some (as yet unposted) analysis on preseason games that suggests there's at least some correlation between the way teams perform in the preseason and how they do (at least early) in the regular season.

I was going to do a more in-depth analysis of this during the summer, and present my findings, but computer troubles have derailed that project (for the moment, anyway).

Here's How it Works

Each team's power is computed, based on the APR algorithm. But instead of the game's final score, the game's half-time score is used, instead. Of course, this isn't the steadiest ground in the world: teams don't play for the win like they do in the regular season, and they certainly don't play to be ahead at half time like they play for the win.

On the other hand, good offenses score points, and good defenses get stops, even in preseason. I (somewhat arbitrarily) cut off the second-half results to winnow out the effects of teams playing their second-string and soon-to-be-cut players.

I plan to do a more careful (quarter by quarter and week by week) comparison of preseason to regular season results, but probably not until after the Super Bowl.

The Preseason Half-time Power Rankings for Week 3

RankPowerTeam
1) 1.5256 Buffalo Bills
2) 1.5124 Miami Dolphins
3) 1.4891 Carolina Panthers
4) 1.4540 Arizona Cardinals
5) 1.4314 San Diego Chargers
6) 1.4303 New Orleans Saints
7) 1.3582 Detroit Lions
8) 1.3417 Minnesota Vikings
9) 1.3386 Kansas City Chiefs
10) 1.2998 Seattle Seahawks
11) 1.2982 Tampa Bay Buccaneers
12) 1.2885 Jacksonville Jaguars
13) 1.2881 Oakland Raiders
14) 1.2704 Indianapolis Colts
15) 1.2559 Cincinnati Bengals
16) 1.2381 New York Giants
17) 1.2379 Pittsburgh Steelers
18) 1.2251 Denver Broncos
19) 1.2185 San Francisco 49ers
20) 1.2170 New York Jets
21) 1.2143 Philadelphia Eagles
22) 1.2135 Tennessee Titans
23) 1.2034 Houston Texans
24) 1.1895 Washington Redskins
25) 1.1790 Dallas Cowboys
26) 1.1714 Atlanta Falcons
27) 1.1596 Chicago Bears
28) 1.1476 Cleveland Browns
29) 1.1452 Saint Louis Rams
30) 1.0414 Green Bay Packers
31) 1.0365 Baltimore Ravens
32) 0.8355 New England Patriots