So Brian Burns or Sweat it is.

Very nice work Ross. I wish there was more data (more years).

Also it might be worthwhile finding out the average draft position of each position (all players, not just the best). It may reveal that a disproportionate amount of QB’s and Edge players are taken in rounds 1-2.

Might help illuminate any statistical significance or lack there of.

Said it last year and I'll say it again, your analysis is incomplete, leading to false conclusions.

Elite QBs are more valuable than TEs. As a result, more QBs are taken in early rounds, and teams wait to pick TEs.

Look at TEs last year by round:
First - 1
Second - 2
Third - 2
Fourth - 5

Look at QBs last year by round:
First - 5
Second - 0
Third - 1
Fourth - 1

We should expect good QBs to be found early in the draft and good TEs to be found in later rounds, its a self-fulfilling prophecy.

A more interesting analysis would give the probability of finding a top player by position based on the number of picks used per round on the position. This would tell you whether GMs are good at finding elite talent in those rounds.

I think the data is a little skewed, but just got back from Costa Rica vacation so I'm not thinking exactly right yet. So if a first round QB is successful and an edge rusher goes in the first round and both are successful is it because they went in the first round, or because they were good and were selected in the first round. An example, Quentin Nelson, a guard, goes in the 1st round and is successful. A third round guard starts, but not All-Pro. So the best round to pick up a guard is the first round? Or the average round good guards go is the 2nd?
I actually think that each team has a trend and enough research, you will find it and have success determining their picks.

Most horizontal draft boards have individual players ranked per first through 7th rounds. Even priority FAs have a slot. What this year's draft is concluding is there are NO, I repeat NO top 5 draft prospects. There will be players picked in the top 5 slots, but at this time they aren't top 5 talents.
Here are the top 2 guys and they are rated just below the top 5 slots:
1.Bosa
2.Williams
Mid to late 1st round guys:
3.Allen
4.Oliver
5.Ferrell
6.Wilkins
7.White
8.Simmons
9.Williams
10.Hockenson

Those are your top ten players. Figure two maybe three QBs will be selected before 12, and one of those top ten guys will fall to the Packers. The next three positions are:
11.Gary
12.Murphy
13.Jacobs

So one of the top ten or one of the next 3. Those are your draft possibilities. If you knew how Gutsy drafted, or what players MLF wants...you could probably figure who will be selected with that 12th pick.

All the analysis is good and fun and important and intriguing.
Then you have to make a human decision as GM about another human as a draft pick.
If at #12 there is a choice between Burns and Oliver and Hock, you gotta go with the human factor.
Which guy do you most wanna go with into the fray for the next several seasons?
There are no guarantees -- things happen. But at least you win or lose with your guys, so no real regrets.

I guess to say how good or bad the NFL is at judging talent at a position you'd have to look at the whole pool and see the average round drafted for each position. The over/under would indicate that the NFL is relatively good or bad at evaluating a particular position, or at least not reaching for a particular position. Example: not many interior offensive linemen are taken in the first round, so you have to factor that in if you are going to use these numbers to determine how accurate teams' assessments are of a particular position.

Some positions it is also really hard to judge who the top 20% really are - this might somewhat account for the big swing in OTs between the two years.

I am surprised that rb and wr are so high. It always seems that good ones are found in later rounds or even UDFA like that cat from Denver.

Ross - I think at this point you have data but not yet enough for it to provide information. A few more seasons will be helpful. While I would expect that QBs and Edge are usually found at the top of the draft other positions may depend on where the value is in a particular year.

One year the draft may be a good season for CBs, the next year’s draft may be strong for WRs, and so on. Therefore it will take at least a few more seasons to develop a comprehensive view for each position.

However, I appreciate the concept and your effort to bring the information to us. Thanks, Since ‘61

Not many interior OL are taken high becuase it is relatively easy to train one up who was taken lower.

The value of the position is key. Does anyone trade a brady or rodgers for a couple of guards?

You carry two QBs and 5 edge guys.

QB and Edge are two of the most impactful positions, so yeah, they get drafted earlier and that drives the lower numbers. But it shows that later rounds teams can find great players. ie: Bahktiari. I think it is best if Gute can hit it good with the picks. If Gute gets it right on two thirds of the first four rounds and then maybe a couple of good picks in round 5 through 7. If that happens, then Gute hits a home run. Just don't hope another 2016 draft occurs. That doomed Thompson as his formula worked only if he was successful.

Once Thompson lost Dorsey he came apart at the seams.

Another way of looking at this would be calculate the average draft value for each position for all starters on game 3 (use pick used rather than round).

Average draft position of 32 QB starters = 38
Average draft position of 32 TE starters = 76
Average draft position of 64 WR starters = 52
Average draft position of 64 OT starters = 61

The problem with this analysis is interpreting why a number is high. Is it high because there lots high pick drafts or because teams value these positions less and draft for them more heavily after the premium rounds. That's why I like the bust rate approach better. It tells you where and when the risk is relatively high.

overkill Archie

OK, I won't argue. But I think I'll still pop in the game tape and make my decisions that way.

All the nerd analytics and such are very cool and interesting, BOTTOM LINE-IF YOU CAN'T JUDGE TALENT THIS IS ALL BS. Fun BS but BS nonetheless.

Ross, it's a noble endeavor but I'm not sure that you would be able to achieve your goal by purely looking at draft position and player ranking once in the NFL.

There are additional factors to go into this algorithm to make it more accurate and these would be:

1) Relative positional draft strength (perhaps including a draftee score and round projection) : Is this a strong or weak draft class for that position? Did someone go higher than expected? How does this positional draft class stack up against previous year's classes for depth?

2) Team need: Are needy teams going to draft for need over BPA

3) Player fit: Is the draftee "the right tool for the right job" or is a team trying to use a draftee differently to their college history (e.g. a strict 3 technique 4-3 DE as a 3-4 OLB)

Looking at the PFF top 8, the mode is 1st round. The mode is also the number one overall pick.

The median is 17th pick and First round. The average due to Brady (199th pick - 6th round Comp) is the 43rd pick.

I note that 25 of the top 32 QBs per PFF were first round picks, with most of them having been selected in the top 12 (that's from memory). Luck, Goff, Mayfield, Ryan, Wentz, Mariota, Newton, Winston, Smith, Allen, Darnold, Trubisky, and Rosen were all top 5 picks. A couple of 2nd rounders, one third (Wilson), one fourth (Cousins), one sixth (Brady), and Keenum was a UDFA.

I would conclude that franchise QBs generally are found in the first round, and usually the top half.

Using just passer rating, the mode was first round and pick number 7 overall. The average was pick 17. Passer Rating excludes Brady/AR and added Luck and Winston instead. Same conclusion.

Thanks for your research and kudos to you as a writer for inspiring me to take a deeper look into the topic. For starters, here's a chart of your data for the past two years - I find this presentation of the data to be more revealing:

2018 Draft Position of the Top 20%
Position R1 R2 R3 R4 R5 R6 R7 UDFA Total
------------ ---- ---- ---- ---- ---- ---- ---- -------- --------
QB 5 1 1 1 8
RB 4 2 4 3 1 14
WR 5 5 1 2 1 2 16
TE 2 5 1 1 2 3 14
T 9 1 2 1 1 2 16
IDL 6 4 4 3 2 3 22

DL 7 5 4 3 3 2 24
EDGE 14 4 1 2 21
ILB 3 7 2 3 2 1 1 19
CB 11 3 1 1 1 7 24
S 4 3 2 2 3 1 1 1 17

2019 Draft Position of the Top 20%
Position R1 R2 R3 R4 R5 R6 R7 UDFA Total
------------ ---- ---- ---- ---- ---- ---- ---- -------- --------
QB 5 1 1 1 8
RB 3 2 3 1 1 2 12
WR 8 5 6 2 2 23
TE 2 3 4 2 2 13
T 9 4 1 1 1 16
IOL 6 4 4 3 2 1 3 24

DL 6 5 3 2 1 2 3 22
EDGE 10 5 3 1 1 20
ILB 2 6 1 2 3 1 16
CB 11 3 1 1 1 5 22
S 8 4 2 1 3 1 19

I agree that more data is needed before we jump to conclusions. I look forward to producing more of these charts - looking back more years for trends. I also want to see what the charts look like for the top half of the league's starters - how many late round picks make it to solid starter status.

I used a similar process to chart some draft rankings last year and it led to some interesting conclusions (including drafting 3 WR's).

One final note about your data - I found two numbers that used wrong counts. Your 2019 data sheet listed 15 LBs but your stated count was 18. The number of TEs listed was 13, not 14. It appears these will change some of your data points.

Thanks again Ross.