psubills62
04-10-2009, 03:58 PM
OK, so this is going to be a little long, but I think a lot of people who visit this forum should be interested in this. For a while, I've thought there should be some way to objectively rank each draft analyst based on their mock draft. There would be many factors, like some of the following: how many picks they got exactly correct, how far off they were with regards to where a player was drafted, if the team picked the same position (even if it wasn't the same player), etc.
Now before I really get going with this, someone stop me if another person has done this already. But I've started an Excel spreadsheet where I tabulated results from the 1st round of last year for: Kiper, Mayock, McShay, Matt McGuire and Walter (the latter two from walterfootball.com).
Right now, all I've done is go through the first rounds of their mock drafts (that's all I could find) of 2008, and compared them to the actual draft in 2008. So, for example, I went through Kiper's draft, pick by pick, and calculated how many spots removed a player actually was drafted, from where Kiper projected him to be picked.
So, for example, according to this website,
http://mb.jaguars.com/Topic513822-20-1.aspx
Kiper thought that McFadden would go to the Jets at pick 6, when he actually went to the Raiders at pick 4. So that would give Kiper a score of 2 for pick #6 (where he projected McFadden to go). If they got it exactly right, in the same spot (EVEN IF it wasn't the same team), then I gave them a 0. Then I took the average of those values for the first round, and tabulated them.
Obviously, I have to include other factors, but I didn't have the time yet (maybe this summer). For example, I want to include a factor that helps them if a team drafts the same position as an analyst projected, but a different player. Also, a factor that would help that analyst if there was trading that messed up their mock, or if they got a decent amount correct in the LATTER half of the first round (picks 16-32).
Does this make sense so far? I'll be around late tonight to try and answer any clarification questions.
Here's what I'd like from you guys:
1) Let me know who you want me to analyze. I've got Kiper, Mayock, McShay, Matt McGuire, and Walter so far. Is there anyone else you want me to add?
2) I haven't checked for sure, but I'm pretty sure I need ESPN Insider to access Kiper and McShay's mock drafts. When they post their final mock drafts prior to the actual draft, could someone post the whole thing, so I can record it?
3) What other factors should I include?
4) How many rounds should I go? Obviously I'm limited by how many rounds each analyst picks, but if Matt McGuire goes all 7 rounds, and Kiper only goes 3, I'm not sure it's fair to compare them.
I welcome any other comments. The ultimate goal of this is to get a formula that gives one final score for each analyst based on multiple mock draft factors.
Let me know if you support this venture. Here's the "scores" I have so far:
Kiper: 6.29, 9 exactly correct
Mayock: 4.45, 11 exactly correct
McShay: 7.23, 5 exactly correct
Matt McGuire: 13.42, 6 exactly correct
Walter: 8.10, 7 exactly correct
The first score is the average number of spots that their first round picks were off by, as given in the example above with Kiper.
Again, I welcome any comments or discussion about this. Thanks!
Now before I really get going with this, someone stop me if another person has done this already. But I've started an Excel spreadsheet where I tabulated results from the 1st round of last year for: Kiper, Mayock, McShay, Matt McGuire and Walter (the latter two from walterfootball.com).
Right now, all I've done is go through the first rounds of their mock drafts (that's all I could find) of 2008, and compared them to the actual draft in 2008. So, for example, I went through Kiper's draft, pick by pick, and calculated how many spots removed a player actually was drafted, from where Kiper projected him to be picked.
So, for example, according to this website,
http://mb.jaguars.com/Topic513822-20-1.aspx
Kiper thought that McFadden would go to the Jets at pick 6, when he actually went to the Raiders at pick 4. So that would give Kiper a score of 2 for pick #6 (where he projected McFadden to go). If they got it exactly right, in the same spot (EVEN IF it wasn't the same team), then I gave them a 0. Then I took the average of those values for the first round, and tabulated them.
Obviously, I have to include other factors, but I didn't have the time yet (maybe this summer). For example, I want to include a factor that helps them if a team drafts the same position as an analyst projected, but a different player. Also, a factor that would help that analyst if there was trading that messed up their mock, or if they got a decent amount correct in the LATTER half of the first round (picks 16-32).
Does this make sense so far? I'll be around late tonight to try and answer any clarification questions.
Here's what I'd like from you guys:
1) Let me know who you want me to analyze. I've got Kiper, Mayock, McShay, Matt McGuire, and Walter so far. Is there anyone else you want me to add?
2) I haven't checked for sure, but I'm pretty sure I need ESPN Insider to access Kiper and McShay's mock drafts. When they post their final mock drafts prior to the actual draft, could someone post the whole thing, so I can record it?
3) What other factors should I include?
4) How many rounds should I go? Obviously I'm limited by how many rounds each analyst picks, but if Matt McGuire goes all 7 rounds, and Kiper only goes 3, I'm not sure it's fair to compare them.
I welcome any other comments. The ultimate goal of this is to get a formula that gives one final score for each analyst based on multiple mock draft factors.
Let me know if you support this venture. Here's the "scores" I have so far:
Kiper: 6.29, 9 exactly correct
Mayock: 4.45, 11 exactly correct
McShay: 7.23, 5 exactly correct
Matt McGuire: 13.42, 6 exactly correct
Walter: 8.10, 7 exactly correct
The first score is the average number of spots that their first round picks were off by, as given in the example above with Kiper.
Again, I welcome any comments or discussion about this. Thanks!