Help - Search - Members - Calendar
Full Version: Hey Mikey Numbers...
Eagles Forum > Philadelphia Eagles Message Board > Philadelphia Eagles or Football Related Discussion
mcnabbulous
Here is something you'll likely get a kick out of. I'm not sure I agree with using salary cap numbers as the basis, but the result is about in line with what I have in mind.

I'd love to hear what you think:

http://mcnabborkolb.com/blog/2013/4/12/gen...ted-draft-value

http://eaglesrewind.com/2013/04/09/total-p...-the-nfl-draft/
nephillymike
Very interesting.

Good stuff.

I understood most of it. Where he lost me was where he got the cap %'s from. He mentions taking the info for those starting more than 50% of the games then throws out the number 25% and I'm not sure how he uses the 25%.

Assuming that whatever he did resulted in him getting the cap %'s for the top 25% of the players (I guess that is what he wanted to do), then he, I assume, uses these figures to get the %'s of the cap by position.

While I think the analysis is on the right path, I have a few thoughts where it comes up short. I don't know for sure how it would flow in the calculation.

1. Because the QB salaries are so high, I think it overstates their value since he uses salaries. For example, it seems that an average QB would benefit from a higher value adjustment than a PB guard. If you look at some of his shift in the rankings based on this, you see huge jumps in the QB draft rankings. Wilson +22, Barkley +17. Seems way out of line. It should be in there, but I think the adjustment is too big.

2. Relative to the salary adjustment, the risk adjusment is too low.

If I understand the calc correctly, the example below illustrates that point. Assume two prospects with the same rating of 95.0. One a QB and one an OG. From the source it looks like the OG has a miss rate of 33% and a value adjustment of 5.44%. The QB as a risk rate of 50% and a value adjustment of 14.12%. Assume that the std dev for each prospect were identical and I'll ignore them for simplicity.


Prospect..............Rating.......Miss %.....Miss Adj.......Pos. Factor.....Pos Adj..........PVM
Offensive Guard.....95.0..........33%.........(5.74)..........5.44%..........5.17...........94.44
Quarterback..........95.0..........50%.........(7.07).........14.86%.........14.12.........102.05

The difference above yields a difference of 7.61 in PVM

If 95.0 is the top spot in the prospect ratings (real close to Warmack's actual), that 7.6 pt decrease drops him, relative to the QB, a whopping 24 draft slots. (ie., 7.6 points is equal to 24 slots). My gut says that is way too much.

My guess is that the reason for this lies in the salary part being too high.

Looked at another way, if you combine the square root of the risk (ingnoring SD) and the position adjustment, you get a total adjustment factor. A positive number means the position factor exceeds the risk factor:

QB = +7.8
DE = +3.2
OT = +2.6
CB = +1.8
DT = +1.5
ILB= +1.4
TE = +0.8
WR = +0.7
OLB= +0.3
OG = (0.3)
S = (0.6)
OC = (1.7)

The greatest difference is 9.5% between QB and OC. 9.5% of the prospect ratings would move the top player in the pool with a 95.11 down to a realtive 86.07 rating which would put him at #35 relative to the QB. So is a QB worth being drafted 35 slots higher than a Center if they both have the same scout rating??

Seems too steep for me.

Good stuff though. If I get a chance, I'll see if there's a good way to adjust it.
nephillymike
McN, lost you in the numbers??
mcnabbulous
Yes, too much weed smile.gif
Reality Fan
It would be interesting if either method was applied to previous drafts to see if the approach holds any water at all.

This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2017 Invision Power Services, Inc.