Here is the final week’s beatpath graph. The picks were 9-7 this week.

The pecking order amongst playoff teams seems rather clear; the benefits of a vertical graph.

Stay tuned for another graph variant.

Here is the final week’s beatpath graph. The picks were 9-7 this week.

The pecking order amongst playoff teams seems rather clear; the benefits of a vertical graph.

Stay tuned for another graph variant.

My graphs are up: http://www.beatgraphs.com

Moose,

I seem to recall somewhere that you wrote that the Weighted method provided the best overall record in assessing playoffs, is that correct?

And does that mean that the Weighted method predicts an Eagles-Chargers Super Bowl?

I think all weighted did was do the best job of predicting the Super Bowl winner. That’s a pretty small data set, so I’m not reading a ton into it.

Moose, the people* want to see graphs with MIN=>NYG removed!

* Assuming Boga and I represent the people.

Doktarr is correct, the article I wrote here last year referred to only Super Bowl predictions. I went back over my spreadsheet from that research and found a page with some data that would apply to your question. If you consider the final ranks at the end of the regular season, here is what you get:

For the Standard method

20 times the #1 team made it to the Super Bowl

11 times the #1 team won the Super Bowl

For the Iterative Method

20 times the #1 team made it to the Super Bowl

9 times the #1 team won the Super Bowl

For the Weighted Method

14 times the #1 team made it to the Super Bowl

6 times the #1 team won the Super Bowl

So really, the best predictor is the Standard method. However it only gets 11 out of 38 Super Bowls correct. So while IND may have the best chance to take home the Lombardi, that chance is still not very good.

I haven’t run any numbers on predicting the matchup from both conferences, but you can assume predicting two teams is much less accurate. Here are more numbers that I do have.

The following are the average ranks at the end of the regular season for the teams that eventually reached the Super Bowl:

Standard SB Winner: 4.05

Standard SB Loser: 4.74

Iterative SB Winner: 4.92

Iterative SB Loser: 3.97

Weighted SB Winner: 4.66

Weighted SB Loser: 5.97

What’s strange about this is that the Iterative method is worst at ranking the eventual winner at the top, however is best at ranking the runner-up. The method rated 7 SB winners 10th or below including NYG last year. Weighted missed 5 times and Standard only twice.

The last thing to note is that based on current rankings, all three methods have losing records picking the Super Bowl winner. Both Standard and Weighted are 18-20 while Iterative is 16-22. Standard method is in the midst of a bad run, having picked only 3 of the last 11 winners correctly, losing 6 in a row from #1 GB ’97 (defeated by #4 DEN) to #4 OAK ’02 (defeated by #9 TB).

Recently however, there has been no variation between the methods. For the last 7 years each has taken the same side of the matchup, and each has posted a 3-4 record.

Setting the NYG/MIN game to a tie (which essentially eliminates it)

Standard: http://www.beatgraphs.com/images/S_2008_17_AL.png

Iterative: http://www.beatgraphs.com/images/I_2008_17_AL.png

Weighted: http://www.beatgraphs.com/images/W_2008_17_AL.png

MIN drops a little in the Iterative graph, but there’s not much change otherwise.

Yep, that’s pretty much what I expected. Still, I like that iterative graph quite a bit more. If I had to bet on three teams to win the super bowl, I’d probably take the top row of that graph.

According to that graph, the Redskins and the Pats are the two unfortunate victims of a tough schedule and/or being in the wrong division, while the Dolphins and Chargers are lucky to advance.