Tuesday, November 20, 2007

Patriots 16-0

The Pats certainly look dominant, but can they really go undefeated? Mercury Morris is pulling for them. ESPN anchors respond to this question with their typical yes/no approach, but I like to think I'm a little more sophisticated than that.

The chance at 16-0 is particularly appealing because you can bet both sides of the prop ('Yes' and 'No') for good prices; the Yes is as good as 1-1 (Intertops), the No as good as +230 (VIP).

I'm no NFL handicapper, but NFL game lines are usually pretty accurate. The lines for some future Patriots games are already out, so we can use them to roughly estimate their chances of winning each game:

Week 12: vs. Eagles: 95.0%
Week 13: at Ravens: 91.0%
Week 14: vs. Steelers: 90.0%
Week 15: vs. Jets: 96.0%
Week 16: vs. Dolphins: 96.0%
Week 17: at Giants: 88.0%

The Giants line is interesting; the Patriots are 11-point favorites, but that number is probably influenced by the possibility that New England rests their starters in Week 17. If they lose before Week 17, they'll probably do this, but a run at immortality might convince them to go for broke. In lieu of this possibility, I adjusted the Pats' winning percentage up from 83.0 to 88.0.

Remembering that these numbers are not tested for accuracy, the final verdict: The Pats go undefeated approximately 63.1% of the time, and you should bet both sides of the prop if you have accounts at both Intertops and VIP.

Thursday, November 8, 2007

Adrian Peterson

Adrian Peterson is on pace to set all kinds of records this year, but will he? Sportsbook.com is offering odds that Peterson will break Eric Dickerson's rookie rushing record of 1808 yards. The line is Yes -250/No +190, meaning that the bookies think he has somewhere between a 65.5% and 71.4% chance to do it.

To break the record, Peterson needs 773 yards in his final eight games. It's easy to say that he's on pace for another 1036, so he should break the mark with room to spare. I, on the other hand, think this is another case where we're not giving fair consideration to regression to the mean.

773 yards in eight games is a full-season pace of 1546 yards. How many running backs were projected to top 1546 yards before the season started? Just one, according to KUBIAK: Frank Gore, who's not exactly lighting the world on fire right now.

Furthermore, Gore isn't splitting time. Peterson has Chester Taylor taking ten carries a game from him. Though AP took the rock 30 times Sunday, that looks like an aberration; he should get about 20 carries per game the rest of the way, IF he stays healthy.

With 160 carries, Peterson needs to average over 4.83 yards per carry to break the record. I'll gladly take the Under on that. Feel free to search for a running back who averaged 4.9 yards or better as a starter in his rookie year; you'll be busy for awhile. Furthermore, given Peterson's dominant stats and the Vikings' total lack of passing game, it's very likely defenses will key in on Purple Jesus the rest of the way.

Getting +190, I have to strongly recommend the Under here, which should probably hit 50% of the time or so.

Sunday, November 4, 2007

Tom Brady

(Yes, it's not baseball, but we have to branch out in the offseason before this blog gets condemned.)

What are the odds Tom Brady breaks Peyton Manning's record by throwing for 50 touchdowns this year?

To do so, Brady needs 17 or more TDs in the Patriots' seven remaining games. Of course, it seems unlikely that Brady will play full games in Weeks 16 and 17, but stranger things have happened (like running up the score in a 38-0 game.)

The possible outcomes can be approximated by a Poisson distribution. If we assume Brady's true level of performance right now is 40 TD/year, and that he plays the equivalent of six full games from here on out, his expected number of touchdowns for the rest of the year is 15. The Poisson distribution tells us that 33.6% of the time, he will throw for 17 or more scores.

If you question those assumptions (and who wouldn't) here's the full chart:

True Rate Record %


12 10.1%
13 16.5%
14 24.4%
15 33.6%
16 43.4%
17 53.2%
18 62.5%
19 70.8%
20 77.9%
21 83.7%
22 88.3%