NFL Week 2 Re-Cap Week 3 Power Rankings

The week 2 games of the NFL provided ANOTHER set of surprisingly results. I only predicted 8 of the 16 games correctly. Combine that with the 9 from 16 from week 1 and I am only 17-15 or 53% accurate. For comparison, my model is 64% accurate from the 2002 season to date. I do not think many people would have predicted, prior to the season starting, that: the Chiefs, Dolphins, Buccaneers and Jaguars would all be 2-0 or that the Steelers and Seahawks would be win-less - given all of their schedules. Although give the guys at ESPN some due as they are doing pretty well at picking the games.

I looked into this a bit further to see if it was just my model that was off or if it was a more general function of the results this season. I used bookmakers as a baseline for this comparison as a bad bookie is a broke bookie so they are generally pretty good at getting the lines right. For the bookmaker's prediction I just used whichever team they had as favourite. In some of these case it was very close, but that is also true of the win percentages in my model. 

The table below lists the games and shows when I was correct and when the bookmaker was correct in picking the outright winner - green is good and red is bad by the way. Apologies for the bad quality image. I am struggling to find the best way to put images in Blogger without it ruining the quality. I could, of course, use a better online publishing tool but no one reads this anyway, so it is not like it matters. It transpires that the bookmaker and I were as bad as each other, with exactly the same number of correct picks, just slightly different ones in a few cases. My biased conclusion is: there is no need to worry yet. I will worry if I start losing money and as that has not happened by a long shot, I will just carry on.

But let us put all of that to one side because, although my overall prediction rate might not be going so well, my betting is going just fine. I went 3-2 in week 2 for a net ROI of +33%. Combined with the week 1 results, this gives me an overall win rate of 5-2-1 (bloody Steelers!) and an ROI of +28.5% over the first 2 weeks. This week's bets were:

• Saints at -417 👍
• Dallas at -161 😜
• Tennessee at + 160 😍
• Arizona at +570 🙈
• Seahawks at +180 😭

The Saints was a close one, especially given the odds. The Cleveland Browns are still very badly rated in my model but you cannot go 31-1 over the previous two seasons and not be. Dallas are still over-rated but came through for the win against the even worse New York Giants. Arizona was a bit of a punt with a very small bet, just because the odds were so long. The Seahawks are probably still a touch over-rated in my model, as their team has sort of fallen apart, but at nearly 2 to 1 I do not think that bet was too bad versus Chicago.

Speaking of ratings within my model, the relative positions of each of the teams is listed below. The Chiefs, Buccaneers and Rams are on a charge up the rankings, unsurprisingly. The Bears are another team that seem underrated with the Bills and Raiders joining the over-rated teams. But good things can happen when the bookmakers dismiss a traditionally good team too quickly.

Power Rankings:

1. New England Patriots
2. Philadelphia Eagles
3. Minnesota Vikings
4. Atlanta Falcons
5. Kansas City Chiefs
6. Los Angeles Rams
7. Jacksonville Jaguars
8. Pittsburgh Steelers
9. Carolina Panthers
10. Dallas Cowboys
11. New Orleans Saints
12. Tampa Bay Buccaneers
13. Cincinnati Bengals
14. Los Angeles Chargers
15. Miami Dolphins
16. Green Bay Packers
17. Tennessee Titans
18. Baltimore Ravens
19. Denver Broncos
20. Seattle Seahawks
21. San Francisco 49ers
22. Detroit Lions
23. Buffalo Bills
24. Arizona Cardinals
25. Washington Redskins
26. Oakland Raiders
27. New York Jets
28. Chicago Bears
29. Indianapolis Colts
30. New York Giants
31. Houston Texans
32. Cleveland Browns