Pythagorean Wins: What Do They Predict for the AFC North?

by Desroko on Apr 8, 2008 12:45 AM EDT Comment 7 comments

Every year, fans (and haters) of certain teams claim that the sqad in question under-(or over-) performed that last season, and that things will get better (or worse) next season.

Turns out, they're often right. These games aren't played in a vacuum, and given the effects of pschology, weather or field conditions, injuries, and cruel Lady Luck, sometimes a team under or overperforms in enough games to significantly affect their record.

Bill James, baseball statistician, invented a formula to try and take some of the randomness out of the game, and find a team's ideal record acording to how many runs/points they scored and allowed that season. With a bit of tweaking this formula can be applied to other sports, including football.

Here's how the 2007 season looks for the AFC North Division*:

Team Pts For Pts Agst Pyth. Wins Actual Wins
Baltimore 275 384 4.992 5
Cincinnati 380 385 7.872 7
Cleveland 402 382 8.480 10
Pittsburgh 393 269 11.376 10

From this we can see that the Ravens, in the immortal words of Denis Green, were exactly who we thought they were. We also find that the Steelers and Bengals both underperformed (the Benglas may be the most average team ever, given that they should have finished 8-8 yet again), and the Browns, confirming my suspicions, were playing over their heads.

Now this is all very interesting, but you can be forgiven for wondering "What use is this?"

As you may guess, teams that significantly deviate from their record tend to see a regression to the mean the next season. That's good news for us and the Bengals, bad news for the Browns, and well, bad news for the Ravens as well, even though they matched their Pythagorean projection.

This isn't perfect - I have a feeling that our brutal schedule is going to prevent us from the 11 or 12 wins that the formula is projecting right now - but barring catastrophe we should be the favorites to win the division again next year.

*This data was compiled by www.imarc.net/examples/pythagorean_wins/index.php?sport=nfl

0 recs | Comment 7 comments

Story-email Email Printer Print

Comments

Once the draft is over,

we'll be in better position to judge whether we'll be back in the 11-12 wins range.

Some interesing tidbits - Dallas and New England both significantly overperformed, as did the Super Bowl champion Giants. The Vikings, Jets, and Dolhpins are the only teams that underperformed worse than the Steelers - the 1-15 Fins would actually expected to win 4 games, if you can believe it.

by Desroko on Apr 8, 2008 12:52 AM EDT reply actions 0 recs

fins

I can believe that. Injuries killed them. Their O Line was decent, but the loss of Trent Green, just about all their RBs, and a number of important defensive players kept them from even reaching that modest goal of 4 wins. But I guess you have a point, hard to believe they were projected to be so bad and still underacchieved.

I have a feeling Parcells will change all of that.

by Blitzburgh on Apr 8, 2008 12:58 AM EDT up reply actions 0 recs

interesting

Thta's prety accurate stuff, at least for that one sample. I'll be curious to see how it fared with the rest of the league.

I guess that Jets loss was win #11 :(

by Blitzburgh on Apr 8, 2008 12:55 AM EDT reply actions 0 recs

The link provided

at the bottom of the diary lists every team. There's some really interesting stuff. It seems to be extremely rare for a team to deviate two or mor egames in any direction, and most saw a deviation of less than one game.

I need to research a bit more before I can give its predictive value for the next season a vote of confidence. Maybe another diary.

by Desroko on Apr 8, 2008 1:00 AM EDT up reply actions 0 recs

btw

Good to see you again Desroko. It's been awhile, at least since you've left a comment or a diary.

Good to have one of the old timers back on the block.

Thanks for this, this is cool.

by Blitzburgh on Apr 8, 2008 12:55 AM EDT reply actions 0 recs

Yeah, I kind of got away from regular posting

late in the season - a combination of frustration and mounting real-world distractions. I lurked.

by Desroko on Apr 8, 2008 12:56 AM EDT up reply actions 0 recs

Very Cool

stats Desroko. It confirms my belief that the Steelers underperformed last year. Part of that can be attributed to a new HC and a new OC feeling their way and making some poor decisions. Part of that can be attributed to injuries. Part of that is the players fault. (i.e. Carter missing Garrard on 4th down)

While the Steelers may very well play up to potential next year I am hoping for 10 wins and not thinking about 11 or 12. Regardless of who they take in the draft, these rookies will not be performing regularly on the field thus the team we fielded last year will compete against the tough schedule the NFL gave us. I'm certainly not counting us out I am just being cautiously realistic and hoping for an invitation to the dance.

When You Run The Ball Good Things Happen

by 5020 on Apr 8, 2008 10:27 AM EDT reply actions 0 recs