Tony Villiotti - Draft Choice Ranges

[mayoclinic]

Interesting article on the National Football Post from Tony Villiotti of draftmetrics.com. Villiotti suggests that draft choices can be grouped into distinct ranges, in which there is approximately equal probability of a pick being successful:

1. 1-13
2. 14-24
3. 25-46
4. 47-73
5. 74-114
6. 115-187
7. 188 and later

The draft: When 13=1 | National Football Post

Villiotti's claim is that a draft pick in each range has approximately equal "value" - i.e., that there is a roughly equal chance of that pick achieving a certain level of success. The 13th pick has roughly the same chance of success as the 1st pick, and so there isn't much benefit to picking #1 over #13 (consider JJ Watt at #11 in 2011, Haloti Ngata at #13 in 2006, Ben Rothlisberger at #11 in 2004, etc.). So in general, it makes sense to trade up to get into a "higher" draft tier, and to trade back when it keeps you within the same tier.

Obviously, this is an imprecise science, but it's provocative.

The Patriots have generally picked near the top end of the third tier of prospects, in the 25-32 range. They traded up into tier 2 in 2012 to get Chandler Jones at #21, who has been an impact player (as was Vince Wilfork at #21 in 2004).

Again, this is very imprecise. The Pats traded up from 44 to 42 in 2010 to get Rob Gronkowski, a 2 spot trade with enormous implications; and they traded back from 21 to 27 in 2010 and got Devin McCourty, a player they had rated worth taking at 21 in the first place. But it's a useful starting point, IMO.

 

[manxman2601]

Sorry I missed this post.

I'm not a fan of a sweeping narrative approach to analysing the draft. It may be that averages even out over a number of drafts and that, in fact, if you take a 20 year average, that the results do show as the writer suggests. But I don't see any practical application here. I think analysing drafts over a period of years can help in identifying trends, the decline in the value of RBs for example, but I don't think it's so easy to determine absolute value with this holistic approach. In fact, a team is more likely to 'win' the draft by being the best at drilling down as much as possible in a specific draft and ignoring the past.

Edit: thank you for sharing though. Any view is worth considering.

 

[everlong]

I think there's definitely an approximate value by range but I do believe it also varies from draft to draft. Some years the blue chip might be 1-8 and other years 1-16. I think this year that 14-24 and 25-46 ranges might look more like 14-28 and 29-66. With the Pats picking on the cusp it all depends on if one of the top 28 make it to 29. If not it's a certain trade back assuming a partner can be found even for less than conventional value. We have to hope teams fall in love with the QBs. I pray Bortles, Bridgewater, Manziel, Carr and Garoppalo all go before 29.

 

[Get it shawtaay]

Interesting article on the National Football Post from Tony Villiotti of draftmetrics.com. Villiotti suggests that draft choices can be grouped into distinct ranges, in which there is approximately equal probability of a pick being successful:

1. 1-13
2. 14-24
3. 25-46
4. 47-73
5. 74-114
6. 115-187
7. 188 and later

The draft: When 13=1 | National Football Post

Villiotti's claim is that a draft pick in each range has approximately equal "value" - i.e., that there is a roughly equal chance of that pick achieving a certain level of success. The 13th pick has roughly the same chance of success as the 1st pick, and so there isn't much benefit to picking #1 over #13 (consider JJ Watt at #11 in 2011, Haloti Ngata at #13 in 2006, Ben Rothlisberger at #11 in 2004, etc.). So in general, it makes sense to trade up to get into a "higher" draft tier, and to trade back when it keeps you within the same tier.

Obviously, this is an imprecise science, but it's provocative.

The Patriots have generally picked near the top end of the third tier of prospects, in the 25-32 range. They traded up into tier 2 in 2012 to get Chandler Jones at #21, who has been an impact player (as was Vince Wilfork at #21 in 2004).

Again, this is very imprecise. The Pats traded up from 44 to 42 in 2010 to get Rob Gronkowski, a 2 spot trade with enormous implications; and they traded back from 21 to 27 in 2010 and got Devin McCourty, a player they had rated worth taking at 21 in the first place. But it's a useful starting point, IMO.

Thanks for sharing.

I did an analysis like this on my own a while ago, using Pro Bowls alone, which is admittedly a faulty measure.

I reached one similar conclusion - that 1-15 is really where the true value is found. However, found that 16-45 was a pretty flat distribution of impact talent. Then a steady decline into shooting while blindfolded territory by pick #75 or so.

So my conclusion was something like this:

1-15 - The "real first round"
16-45 - The "real second round"
46-75 - The "real third round"
76 - The YOLO round

Obviously it looks like this study was more in-depth than mine. The biggest problem with the Pro Bowls as a measure tactic was the potential for name recognition (see: Meriweather) to dilute the results.

 

[mayoclinic]

Thanks for sharing.

I did an analysis like this on my own a while ago, using Pro Bowls alone, which is admittedly a faulty measure.

I reached one similar conclusion - that 1-15 is really where the true value is found. However, found that 16-45 was a pretty flat distribution of impact talent. Then a steady decline into shooting while blindfolded territory by pick #75 or so.

So my conclusion was something like this:

1-15 - The "real first round"
16-45 - The "real second round"
46-75 - The "real third round"
76 - The YOLO round

Obviously it looks like this study was more in-depth than mine. The biggest problem with the Pro Bowls as a measure tactic was the potential for name recognition (see: Meriweather) to dilute the results.

Looking at it that way, the last "real first round" pick we had was Jerod Mayo in 2008. The Pats reportedly considered Nate Solder in 2011 a top 10 talent, and got him at 17. Otherwise, we've been trying to find 1st round talent in the 2nd round and later: Chandler Jones at 21, Dont'a Hightower at 25, Devin McCourty at 27, Rob Gronkowski at 42, Jamie Collins at 52.

I think that Aaron Donald and Eric Ebron are "real first round" talents. Ra'Shede Hageman has more risk than you would like, but the upside is there. You have to pick your spots and take your shots more carefully when you are consistently picking in the late 20's and later.

 

[BobDigital]

I generally have had my suspicions about these ranges as well. If I had to do it just by my memory I'd give similar ranges as well.

I think the difference from the top 12-15 prospects is more pronounced that the writer suggest though. You need to keep in mind most years about 1-3 of the top 15 or so spots is thrown away cause teams take a risk on a shaky QB prospect which usually fails to work out. That actually brings the numbers down.

 

[stinkypete]

I've generally believed in this "range of values" approach to the draft, but I think it becomes way too subjective to make any kind of specific groupings beyond pick 30.

I see the top 15 or so picks as the "likely impact player, regardless of system" range. Beyond that, all player values are dependent on system. For example:

Who is more valuable, Timmy Jernigan vs. Louis Nix?
The Patriots fan in us jumps up and shouts "Nix" before the "is" sound in "Louis" is spoken. However, Nix is seen to be falling down the draft boards, while Jernigan is mocked anywhere in the top 15 - 25 range.


Similar arguements could be found when comparing ASJ vs. Feidorwicz, Tre Mason vs. Carlos Hyde, etc.

I would be interested to see, from a fan's perspective, a list of consensus rankings by team. How different would a Chicago Bears or Pittsburgh Steelers fan rank this group of prospects? The draft operates the same way.