Saturday, April 25, 2009

Draft Software Working!

After >300 lines of sloppy Perl, I have created a program that does the following:
  1. Ingests as many top 100 draft boards as I can find (I need MOAR!)
  2. Average these boards together to create a consensus top 100.
  3. Compare this "uber-board" against the actual draft to assign scores to each pick. A positive score indicates that the pick was a value (player was picked later than expected based on the consensus big board.) A negative score indicates a reach (player picked earlier than expected based on big board.)
From this, I have some interesting data from the first day of the draft.

The scores for the Lions' picks:
1 Detroit == Matthew Stafford -3.5 (reach)
20 Detroit == Brandon Pettigrew 0.3333 (value)
33 Detroit == Louis Delmas -14 (reach)

Detroit had an average score of -5.7 (reach), ranking #19 of all teams.
The top teams, in order were Philly, Cinci, and San Fran.
The bottom teams were Buffalo, Cleveland, and finally... Oakland (by a lot - score of -37!)

Finally, here's the top ten players still available and their expected position (according to the consensus big board):
1: Michael Johnson (DE) = 37.66
2: Jared Cook (TE) = 51.33
3: Jarron Gilbert (DT) = 56.66
4: D.J. Moore (CB) = 59.66
5: Duke Robinson (OG) = 60
6: Jamon Meredith (OT) = 61.16
7: Rashad Johnson (S) = 62.33
8: Cornelius Ingram (TE) = 62.5
9: Shawn Nelson (TE) = 64.66
10: Shonn Greene (HB) = 67.5

Remember, these scores are derived from the average rank of the top 100 players based on six expert boards. I want to add more to adjust these scores; list some that I'm missing in the comments!

Current source data:
consensus draft (cdsdraft.com)
Rick Gosselin (dallasnews.com)
bleacherreport.com
Scott Wright's draftcountdown.com
drafttek.com's big board
fftoolbox.com - Ricky Dimon

No comments:

Post a Comment