Is quantity over quality the wrong strategy?

[Klomp]

On draft days, Spielman has rightfully earned a reputation for making deals. This is especially true on Day 3s, when he constantly manipulates the draft board, moving up and down often.

But has that strategy paid off? I'm not so sure.

The drafts are usually strong in the top four rounds, but after that the success rate on picks takes a massive hit. Is the constant trading a reason for it? Should the team simply stand pat more often late in drafts? Or is it simply a product of later round picks having a lower success rate, period?

2017: 3 of 7 no longer with the team
2016: 1 of 5 no longer with the team
2015: 5 of 6 no longer with the team
2014: 6 of 6 no longer with the team
2013: 5 of 5 no longer with the team
2012: 4 of 4 no longer with the team

[gopherwrestler]

I love the gaining of tail end picks. This hit rate is already low. You most likely would have drafted one of these players that you moved around and drafted at your single pick, but when you have multiple you can throw a couple picks at players

the fail rate for players in the 6th and 7th round, I just think it’s smart to acquire multiple if you can get the same player you would have anyway. 

[SemperFeist]

The further you get in the draft, the higher the fail rate gets. On a team with a strong roster, 6th and 7th round picks are basically roster fodder. Maybe you get lucky with one or two, but they likely aren’t players who’ll stick around past their original contract, if they’re lucky. It only makes sense to load up on those day 3 picks and hope a few stick. 

[Heimdallr]

has anyone ever charted how long each player stayed in the league against their draft position? is it linear or a curve? where is the dropoff/value?

[Krauser]

Short answer to the title question: No. Quantity over quality is the right strategy (in general). 

It's usually better to trade down than trade up, but there are times where trading up can be justified (player fills an immediate need). There's some evidence that players traded up for do better than other players taken at the same draft slot by teams who stood pat, but that might be a sunk cost phenomenon (teams decide they "need" the player, and so the player plays more, even if they're not very good, because the need is there). 

Spielman has generally done well in drafting. Supposedly no one can beat the randomness of the draft, but his record from 2012 through 2015 was very solid, limited mostly by injuries (Bridgewater, Floyd) and unexpected regressions (Kalil). 2016 was terrible, but 2017 is very promising so far. 

The strategy of accumulating later round picks is a solid one. The draft value chart (not the outdated trade value per draft slot, but the curve of Approximate Value per draft position) is pretty flat after the first 100 picks: 

That's a plot of the average AV over 5 years for each draft slot. The Career AV per player (not just per slot) shows how much variance there is, especially on the right sided tail of the curve: 

Note that the curves are very similar, but the scale is a little different, since the 2nd chart includes a longer time frame (so the full career average of top picks is over 50, not just 35 as in the 1st 5 years). 

Also note, and more importantly, that the range of the very best players is much higher than the average -- multiple players score >100 CarAV, especially in the top 20 or so picks, and there are many more similarly successful players taken much later in the draft. The one huge outlier in the top right is no doubt Tom Brady, pick 199, with what looks to be the 2nd higher CarAV on the chart at the time this study was done. 

A few Vikings late round outliers:

• John Sullivan, pick 187 in 2008: CarAV 48
• Everson Griffen, pick 100 in 2010: CarAV 50
• Matt Birk, pick 173 in 1998: CarAV 84 --> you can probably find Birk's dot on that 2nd graph above

Because the value curve is so flat, it's better to have multiple shots a little later than one shot a little earlier -- especially near the end of the draft. For instance trading a 3rd for a later 3rd and a 5th is a very good idea -- you drop only slightly down as you move to the right on the curve in the middle of the 3rd round, and you pick up an entirely new lottery ticket around pick 150, with an expected CarAV that is worth more than that. 

The best recent late round Spielman pick is Diggs (already CarAV of 23, tied for about 13th in the draft class, with no one drafted outside the top 100 picks ahead of him). The Diggs pick is an interesting example of this -- Spielman traded down twice in the 3rd round, netting a 5th and a 6th. With 2 picks at the top of the 5th (the other being acquired in trading Matt Cassel to the Bills), he traded down with the first of them (the Falcons traded up for Grady Jarrett, who I would've loved to see in purple), to #146. Spielman's 1st pick in the 5th was the one he got from the trade downs in the 3rd round -- that was used for MyCole Pruitt, who's done nothing in the league. Only after picking Pruitt did he draft Diggs. 

Obviously, most of the trade down scenarios won't work out so well. Vikings traded down one spot before taking Barr in 2014, and their reward was a 5th round pick that they used on David Yankey, who did nothing in the NFL. 

My point in recounting the Diggs pick is that it pays to accumulate more lottery tickets. Somehow, the Vikings had Pruitt over Diggs -- either from need or projecting him to their scheme (IIRC Norv loved him). They traded down twice in the 3rd round and finally selected Danielle Hunter, who's been excellent value. They were willing to trade out of Grady Jarrett, who's been very good in Atlanta. And they still landed Stefon Diggs, who's a much better WR than Treadwell, who they spent a ton of draft capital on the following year. 

There's a lot of data analysis about the NFL draft. If you're interested, I recommend: https://statsbylopez.com/2016/05/02/the-nfl-draft-where-we-stand-in-2016/ as a good summary. 

For @Heimdallr 's question: Scott Kacsmar at Football Outsiders went through the 1994-99 drafts and did this analysis: 

https://www.footballoutsiders.com/index.php?q=stat-analysis/2016/what-does-nfl-draft-really-produce-part-i

https://www.footballoutsiders.com/stat-analysis/2016/what-does-nfl-draft-really-produce-part-ii

 

 

 

[The Gnat]

13 hours ago, Heimdallr said:

has anyone ever charted how long each player stayed in the league against their draft position? is it linear or a curve? where is the dropoff/value?

I had done some analysis at one point to see how likely you were to find a starter at various rounds. I think it was star, starter, stuck around, and gone levels, but somewhere along the lines I lost that research. When I did it, about five years ago now, I want to say that the dropoff in terms of finding starters was basically after round three with a few random spikes later, but nothing consistent.

[Heimdallr]

It seems like if you want to mathematically optimize your draft, you want 1st-2nd round picks, and then a ton of 6th-7th round picks

[perrynoid]

5 minutes ago, Heimdallr said:

It seems like if you want to mathematically optimize your draft, you want 1st-2nd round picks, and then a ton of 6th-7th round picks

Additionally, do your research and aggressively sign the better UDFA before other teams get the chance to.

[Heimdallr]

2 minutes ago, perrynoid said:

Additionally, do your research and aggressively sign the better UDFA before other teams get the chance to.

It would also be interesting to see what the "hit" rate is for priority UDFAs  (guys signed immediately after the 7th round) compared to 5th/6th/7th round draft picks.

Maybe anything after the 2nd-3rd round does not a statistically significant value over UDFAs.

[perrynoid]

3 minutes ago, Heimdallr said:

It would also be interesting to see what the "hit" rate is for priority UDFAs  (guys signed immediately after the 7th round) compared to 5th/6th/7th round draft picks.

Maybe anything after the 2nd-3rd round does not a statistically significant value over UDFAs.

The Vikings have done well with UDFA, perhaps a good strategy paying off in not only an improved roster, but money-wise saves a lot on the cap.

http://www.vikings.com/news/article-1/3-Vikings-Rookies-Go-From-Undrafted-to-53-Man-Roster/45f3ac12-7577-4c8c-b617-44dee17b871e

"Bower, Wilson and Collins are three of 19 players on the Vikings who began their careers as undrafted free agents, joining the likes of  Case Keenum ,  Adam Thielen , Remmers,  Tom Johnson ,  Andrew Sendejo and  Marcus Sherels ."

[Krauser]

1 hour ago, Heimdallr said:

It would also be interesting to see what the "hit" rate is for priority UDFAs  (guys signed immediately after the 7th round) compared to 5th/6th/7th round draft picks.

Maybe anything after the 2nd-3rd round does not a statistically significant value over UDFAs.

Pretty sure there's a higher hit rate for late round picks. But the pool of UDFAs is much bigger -- even with all of Spielman's trading down, he typically has ~5 players drafted in rounds 5-7 per year. Compare that to signing ~12 UDFAs right after the draft. Plus they often add UDFAs from other team's practice squads during the year, and/or bring in guys on tryouts. 

 

1 hour ago, perrynoid said:

The Vikings have done well with UDFA, perhaps a good strategy paying off in not only an improved roster, but money-wise saves a lot on the cap.

http://www.vikings.com/news/article-1/3-Vikings-Rookies-Go-From-Undrafted-to-53-Man-Roster/45f3ac12-7577-4c8c-b617-44dee17b871e

"Bower, Wilson and Collins are three of 19 players on the Vikings who began their careers as undrafted free agents, joining the likes of  Case Keenum ,  Adam Thielen , Remmers,  Tom Johnson ,  Andrew Sendejo and  Marcus Sherels ."

The total of 19 is a little misleading -- many of them were acquired in ways other than signing with the Vikings as UDFAs straight out of college. Remmers, Keenum, Tom Johnson, Lamur, Quigley and Forbath were signed as free agents after playing for other teams. The Vikings traded for Sirles, Easton and Tramaine Brock. They poached Rashod Hill off the Jags practice squad and Sendejo originally signed with the Cowboys, played a few games with them, and spent time on the Jets practice squad before coming to Minnesota. 

Of the UDFAs who originally signed with the Vikings, the main contributors so far are Thielen, Sherels and Anthony Harris. Tashawn Bower, Cayleb Jones, Eric Wilson and Aviante Collins are decent bets to contribute more this year. 

[perrynoid]

Yes, the 19 is a bit misleading, but illustrates the point that UDFA's are a resource oft overlooked by many.  Draft capitol is quite limited, by it's very nature, and it is a good strategy (one seemingly pursued by the Vikings) that investing resources into scouting can pay off when targeting (as you point out) UDFA to sign after the draft is over, as well as plucking people from other teams.  Given his athleticism, I am really looking forward to what Collins might contribute this year.

[The Gnat]

Well, with UDFA's, there are so many things to look at. 

It's beyond the 19, but let's say those 19 were from the Vikings, all signed here originally. The oldest player on the list is 33, so assume you sign 12 UDFA's per year, that's a total of (and again assuming) 132 if you got Tom Johnson (the oldest player) at the age of 22. So looking at that, you are hitting on 14% of those players.

Looking at the first post, there are 33 different players drafted in the last 4 rounds, if that's still what day 3 is, out of those 9 on still on the roster which is a hit rate of 27%.

So, looking at that, it's definitely  more important and you can see the drop off from players drafted in the last four rounds to UDFA's. There would be some logic then based on racking up later round picks, moving down in the 4th and 5th to get a higher volume of players in the 6th and 7th.

[Klomp]

I feel like we've done so well in the 3rd and 4th rounds, that I've wondered if it'd be better to trade up to get more picks in that range, when we can.

[Krauser]

10 minutes ago, The Gnat said:

Well, with UDFA's, there are so many things to look at. 

It's beyond the 19, but let's say those 19 were from the Vikings, all signed here originally. The oldest player on the list is 33, so assume you sign 12 UDFA's per year, that's a total of (and again assuming) 132 if you got Tom Johnson (the oldest player) at the age of 22. So looking at that, you are hitting on 14% of those players.

Looking at the first post, there are 33 different players drafted in the last 4 rounds, if that's still what day 3 is, out of those 9 on still on the roster which is a hit rate of 27%.

So, looking at that, it's definitely  more important and you can see the drop off from players drafted in the last four rounds to UDFA's. There would be some logic then based on racking up later round picks, moving down in the 4th and 5th to get a higher volume of players in the 6th and 7th.

The trouble with that analysis is that you're comparing all UDFAs currently on the Vikings, including from other teams and signed to free agent contracts, to only the draft picks who are currently on the Vikings (excluding other teams). 

Using the original post: 

Quote

2017: 3 of 7 no longer with the team
2016: 1 of 5 no longer with the team
2015: 5 of 6 no longer with the team
2014: 6 of 6 no longer with the team
2013: 5 of 5 no longer with the team
2012: 4 of 4 no longer with the team

2012: Walsh was the starting kicker for the Seahawks last year

2013: Mauti started a couple of games for the Saints and played special teams; Locke punted for the Lions for 5 games

2014: Stephen played quite a bit in 2016-17 for the Vikings, and signed with the Seahawks

2015: Edmond Robinson dressed for a game with the Jets last year, and ended the year on the Cards active roster; Pruitt recently signed with Carolina IIRC

All of those guys would count in the "19 UDFAs" for their teams last year if they were UDFAs and you ignore that they came from another team (as was done in counting up the UDFAs to get 19). 

If you add these 6 late round picks to the 9 still on the team, that's 15 players still in the league last year out of 33, more like 45%. Several more are on practice squads, like Exum in San Francisco, and I think Yankey is still kicking around somewhere too. 

The main criticism of the late round picks is that the draft capital investment can be a self fulfilling prophecy -- teams don't want to lose the benefit of the draft pick, so the player gets more chances than a UDFA who cost less to acquire (note: that includes practice squad signings and direct-to-team UDFAs like Sherels and Sendejo; obviously, sunk costs don't apply equally when the former UDFA signs a pricey FA contract like Remmers).