2022 Pats Free Agent period

[SMU_Sox]

Thanks. This is what I mean - to do a comprehensive study of that would take....a long, long time. And nobody here seems willing/interested in doing that.

EDIT: I have actually started to work on this - I started a huge spreadsheet in the fall but I have put it on hold for a bit. Maybe I'll finish it before this year's draft.

I think you are right that most posters wouldn't be willing or interested to do that. I would though. I am a nerd. You want to do a comprehensive review - sure. I think the best way to do that is to go by AV and draft position while also doing like years played, games started, snaps, etc. I think you could probably find tables online with enough of those data points and then would just have to vlookup to complete a master table.

As for the book though. Yes, I have read it. Read parts of it multiple times. . Wouldn't give you a recommendation if I hadn't read it myself. It's a short read but very dense.

 

[SMU_Sox]

Have you read it? If so, how do they determine draft "success"? What criteria?

File is too big to upload:

View: https://twitter.com/johnalimberakis/status/1486051963695513609?s=21

 

[jezza1918]

Does anyone have any clue how/if these various studies/metrics take into account special teams? Matthew Slater is a great hometown example of someone drafted as a WR in the 5th round, but only started 3 games at WR in his entire career. And because of next to nothing in the way of counting stats his career AV per Football Reference is 0. Yet he is a 2x All Pro and 10x Pro Bowler...my gut tells me he very much outkicked his coverage in terms of draft value.

 

[SMU_Sox]

@BaseballJones take a look at the Chase Stewart AV average by draft slot:




I posted my first link to the book in a post a couple up but check out a few more tidbits: (you can also see my 0 key is missing... as well as my Patriots paper plates)

View: https://twitter.com/JohnALimberakis/status/1486054866598379523?s=20

 

[tims4wins]

I guess my pushback would be that "Starts" is not an objective stat unless you are talking about something like OL or QB. For instance is a guy who plays 70% of the snaps as a nickel CB not a starter? What about James White - he has 13 starts in his career? How the hell does one define a start? First snap of the game? It's all about what package you are in. James White has only started 13 out of 95 games. I guess he was a bad pick.

 

[BaseballJones]

I guess my pushback would be that "Starts" is not an objective stat unless you are talking about something like OL or QB. For instance is a guy who plays 70% of the snaps as a nickel CB not a starter? What about James White - he has 13 starts in his career? How the hell does one define a start? First snap of the game? It's all about what package you are in. James White has only started 13 out of 95 games. I guess he was a bad pick.

I don't know but pro-football-reference has whatever criteria they use, and they list "games" and "games started" for each player.

So what's a better criterion? I'm not saying there can't be one, but what IS one?

 

[SMU_Sox]

I don't know but pro-football-reference has whatever criteria they use, and they list "games" and "games started" for each player.

So what's a better criterion? I'm not saying there can't be one, but what IS one?

AV, Snaps, % of Snaps, Second Contract %, Out of the League after rookie contract binary variable, Contract Value, relative contract value to the position, top 5 contract value binary variable, starts, games played. Those should help. I think snaps and % snaps is way more worthwhile than games played or starts, don't you? And even though AV isn't a great calculation it's still something. It is hard to get a value for line play.

 

[tims4wins]

I don't know but pro-football-reference has whatever criteria they use, and they list "games" and "games started" for each player.

So what's a better criterion? I'm not saying there can't be one, but what IS one?

Snaps, or % of snaps.

 

[BaseballJones]

AV, Snaps, % of Snaps, Second Contract %, Out of the League after rookie contract binary variable, Contract Value, relative contract value to the position, top 5 contract value binary variable, starts, games played. Those should help. I think snaps and % snaps is way more worthwhile than games played or starts, don't you? And even though AV isn't a great calculation it's still something. It is hard to get a value for line play.

Where do you get "snaps" information, especially dating back 10-20 years?

 

[BaseballJones]

Snaps, or % of snaps.

You're a great goal line back and pound it in regularly. But you only come into the game in short yardage/goal line situations. So you play very few snaps and a very small percentage of snaps. Yet you produce a high volume of touchdowns. Successful pick or not?

You're a situational pass rusher - explosive in obvious passing situations, and you produce a high percentage of sacks or pressures when you're in. But you don't play much due to the situation. Successful pick or not?

You're a 7th round pick and you stay in the league for 12 years as a backup OL. You don't play many snaps but you're good enough that teams want to keep you on their active roster. Successful pick or not?

And what objective number of snaps, or percentage of snaps, constitutes a "successful" pick?

 

[SMU_Sox]

Where do you get "snaps" information, especially dating back 10-20 years?

That's a great question. I know after 10 years it can be hard to find. You might have some luck asking Aaron Schatz on twitter about that. Also, I think there are websites that sell play by play data because when I was doing a football gambling model we bought that package and got like 7-10 years worth of play by play data. Honestly though if you are serious about researching this let me get back to you on it with specifics.

 

[BaseballJones]

That's a great question. I know after 10 years it can be hard to find. You might have some luck asking Aaron Schatz on twitter about that. Also, I think there are websites that sell play by play data because when I was doing a football gambling model we bought that package and got like 7-10 years worth of play by play data. Honestly though if you are serious about researching this let me get back to you on it with specifics.

I am serious about this. I want to dig into this more - mainly because I'm motivated by the whole "BB sucks at drafting" narrative. So thanks in advance. I think I'm gonna start following you on Twitter, BTW.

 

[SMU_Sox]

You're a great goal line back and pound it in regularly. But you only come into the game in short yardage/goal line situations. So you play very few snaps and a very small percentage of snaps. Yet you produce a high volume of touchdowns. Successful pick or not?

You're a situational pass rusher - explosive in obvious passing situations, and you produce a high percentage of sacks or pressures when you're in. But you don't play much due to the situation. Successful pick or not?

You're a 7th round pick and you stay in the league for 12 years as a backup OL. You don't play many snaps but you're good enough that teams want to keep you on their active roster. Successful pick or not?

And what objective number of snaps, or percentage of snaps, constitutes a "successful" pick?

2nd contracts are a good indicator of success. For the goal line back he gets that second contract, probably but a cheap one. Same for the backup OL. Same for the pass rusher. AV would probably like the pass rusher. I think your methodology would be pretty critical. This is something where there are going to be exceptions. You won't find one neat way to figure it out. Although... again second contracts help with some of this.

 

[BaseballJones]

2nd contracts are a good indicator of success. For the goal line back he gets that second contract, probably but a cheap one. Same for the backup OL. Same for the pass rusher. AV would probably like the pass rusher. I think your methodology would be pretty critical. This is something where there are going to be exceptions. You won't find one neat way to figure it out. Although... again second contracts help with some of this.

Yes every measurement we've discussed has SOME value. But it's all problematic. You sign as a rookie a four year deal. You are awesome for the first three. Then you get seriously injured, and spend your fourth year trying to come back, but it doesn't work out, and you never play another down in the league again. You don't get the second contract. Successful pick or not?

Or you're a career backup and hardly ever play. But because you're a versatile OLineman, you manage a second contract. But your snap count and snap count % are both really low and you make league minimum given your service time. BUT you do have that second contract.

Successful pick or not? And moreover, if the team can't find another better player to replace you, what does it say about the front office?

 

[SMU_Sox]

Well that’s the problem with a lot of stats analyzing football in general. There’s a ton of individual context that you can’t distill neatly to a data point. You look at multiple ways to slice it, define what you think is on average a success for each criteria or variable you are running it by and do your best to try and get a good big picture on it. It’s not going to be easy. Any variable will have issues handling some exception or situation. You just have to, imo, swallow that and again do your best.

 

[BaseballJones]

Well that’s the problem with a lot of stats analyzing football in general. There’s a ton of individual context that you can’t distill neatly to a data point. You look at multiple ways to slice it, define what you think is on average a success for each criteria or variable you are running it by and do your best to try and get a good big picture on it. It’s not going to be easy. Any variable will have issues handling some exception or situation. You just have to, imo, swallow that and again do your best.

That makes sense to me. Which is why I think the guy in the article using "starts" (and half of his games as a starter) is at least something useful. And it's at least pretty easy to dig out that information if you're not looking to spend half your year researching this.

 

[Shelterdog]

Those RB numbers after the 1st round are surprising. That might have changed in recent years as it seems as if teams' enthusiasm for using a 1st on a RB has waned. I imagine BB has not approached 50% on 2nd-round WRs and DBs.

So there's a myth of terrell davis and jonas grey out there that makes folks think that later running back picks are better than they actually are. Not a ton of running backs get picked in round 1 but you're not going to find that many good RBs after round three. Only two of this year's top 20 RBs were picked after round 3.

 

[Devizier]

As an aside, re: the Patriots, my intuition is that they used to develop undrafted/late round* guys with a relatively high success rate. Is that true, or is that just some sort of retroactively-applied bias speaking?

*Not only Brady, obviously

 

[Cotillion]

So there's a myth of terrell davis and jonas grey out there that makes folks think that later running back picks are better than they actually are. Not a ton of running backs get picked in round 1 but you're not going to find that many good RBs after round three. Only two of this year's top 20 RBs were picked after round 3.

 

[Harry Hooper]

So there's a myth of terrell davis and jonas grey out there that makes folks think that later running back picks are better than they actually are. Not a ton of running backs get picked in round 1 but you're not going to find that many good RBs after round three. Only two of this year's top 20 RBs were picked after round 3.

FYI, I was only looking at the 2nd-round and 3rd-round numbers, but yeah. I am a member of the "You had a good draft if you got 3 good players" camp.

 

[Shelterdog]

FYI, I was only looking at the 2nd-round and 3rd-round numbers, but yeah. I am a member of the "You had a good draft if you got 3 good players" camp.

Right. And one thing that throws folks off is when you look a the good teams they almost always had a bunch of recent drafts where they got 3 good players and you say jesus how come we draft shit players and, I don't know, the Bengals are starting about 10 recent draft picks and the Chiefs have two rookies starting and the 49ers got Deebo in the second and Warner in the third, what gives--and the answer is survivorship bias.

 

[BusRaker]

That's a great question. I know after 10 years it can be hard to find. You might have some luck asking Aaron Schatz on twitter about that. Also, I think there are websites that sell play by play data because when I was doing a football gambling model we bought that package and got like 7-10 years worth of play by play data. Honestly though if you are serious about researching this let me get back to you on it with specifics.

As someone who has been a SQL/MDX/DAX professional for 20+ years I was very interesting at getting this data for my football analysis fix

... until I say the price tag at the gambling data provider websites. Definitely out of hobbyist price range

 

[Super Nomario]

So there's a myth of terrell davis and jonas grey out there that makes folks think that later running back picks are better than they actually are. Not a ton of running backs get picked in round 1 but you're not going to find that many good RBs after round three. Only two of this year's top 20 RBs were picked after round 3.

I think a lot of the thing with RBs is that, unlike most positions, a RB can play effectively right away, so when you hit on an Elijah Mitchell there's a 6th round pick playing as a rookie and putting up numbers. Whereas you can find plenty of, say, late-round / UDFA CBs, like Chris Harris or Richard Sherman or JC Jackson, but they usually don't do much as rookies and come on later.

 

[jmanny24]

I have another question about this, if the hit and miss rates on mid-round picks are generally the same (and what is a big difference in hits and misses looking team by team? Is it 2 more contributors? 3 or more) I'll ask again, why does it seem like NE's picks in those rounds have trouble gaining traction and/or getting on the field. Does that point more to a misevaluation of how these players will fit into the "system" or coaches having trouble getting the best out of these players? The oh well the draft is a crapshoot and hard when you pick so late seems like a lackadaisical copout.

 

[tims4wins]

You're a great goal line back and pound it in regularly. But you only come into the game in short yardage/goal line situations. So you play very few snaps and a very small percentage of snaps. Yet you produce a high volume of touchdowns. Successful pick or not?

You're a situational pass rusher - explosive in obvious passing situations, and you produce a high percentage of sacks or pressures when you're in. But you don't play much due to the situation. Successful pick or not?

You're a 7th round pick and you stay in the league for 12 years as a backup OL. You don't play many snaps but you're good enough that teams want to keep you on their active roster. Successful pick or not?

And what objective number of snaps, or percentage of snaps, constitutes a "successful" pick?

As SMU has noted, there are exceptions to every rule, and no perfect way of looking at things. But I think if you were to put together this analysis (not you you, just someone), you'd weight it by round and by position. 50% of snaps for a RB is very different from 50% for a QB or OL. And of course you'd expect the % of snaps to decrease by round.

Of course % of snaps doesn't account for length of career, so that is a problem too.

You'd probably have to calculate average number of snaps in a season for the league, then figure out on average how many guys take snaps at that position, average length of career, etc. etc. to figure out if a guy was "successful". There is just so much noise though.

I just think "starts" is one of the most arbitrary things because it literally means who the team announces as a starter. I don't even know if it means the first snap.

 

[ZMart100]

I have another question about this, if the hit and miss rates on mid-round picks are generally the same (and what is a big difference in hits and misses looking team by team? Is it 2 more contributors? 3 or more) I'll ask again, why does it seem like NE's picks in those rounds have trouble gaining traction and/or getting on the field. Does that point more to a misevaluation of how these players will fit into the "system" or coaches having trouble getting the best out of these players? The oh well the draft is a crapshoot and hard when you pick so late seems like a lackadaisical copout.

Hit rate by team/GM looks a lot like random chance, though possibly with some outlier poor performers (BB isn't one of them).

 

[slamminsammya]

I like the second contract methodology. You could even get fancy with it and look at how the contract ranks among their position in terms of % of cap, or % of cap - positional median or something like that.

The "starts" is obviously introducing lots of weird bias for positions that shift around with personnel packages but I don't think it actually matters too much in aggregate if you want to get a high level idea of how successful draft picks tend to be.

 

[Super Nomario]

Hit rate by team/GM looks a lot like random chance, though possibly with some outlier poor performers (BB isn't one of them).

Any smart things teams do in the draft gets copycatted to death and ceases to be an advantage quickly. I think the Seahawks legitimately had an advantage with their metrics on SPARQ super athletes in late rounds and with their emphasis on DB size in the early 2010s, but it lasted two years, max, before everyone was looking for those players and the advantage was gone. I don't think you can be smarter than everybody in the draft, not and have that edge last for any length of time.

OTOH ... you can definitely do stupid shit. There's probably less stupid shit than there used to be ... but there is still stupid shit.

 

[slamminsammya]

Hit rate by team/GM looks a lot like random chance, though possibly with some outlier poor performers (BB isn't one of them).

The sad thing to me is that the data to evaluate each team's drafting is so incredibly sparse, and it doesn't necessarily have to be. We would have a much better idea of drafting "skill" if we had access to a team's entire board, and I do suspect it would end up looking far less random if we did so. I have fantasized about getting front offices in some sport to release their draft boards on some delay, say 10 years, so that all sorts of cool analysis could be done.

 

[Saints Rest]

I have another question about this, if the hit and miss rates on mid-round picks are generally the same (and what is a big difference in hits and misses looking team by team? Is it 2 more contributors? 3 or more) I'll ask again, why does it seem like NE's picks in those rounds have trouble gaining traction and/or getting on the field. Does that point more to a misevaluation of how these players will fit into the "system" or coaches having trouble getting the best out of these players? The oh well the draft is a crapshoot and hard when you pick so late seems like a lackadaisical copout.

I think the larger question is why does it seem like the Pats had such a run of lousy drafts? Was it that BB had lost his fastball? Was it poor scouting? Or was it poor player development by positional coaches?

Looking back over the last 20 years, how have the Pats done in terms of drafting?

Year; total picks; total aggregate years of starters; total aggregate Pro Bowlers; total "good picks" (all stats from profootballreference.com, except the latter from my crappy opinion)

• 2021; 8; 1; 0; 3 (Mac,Rham, Barmore)
• 2020; 10; 2; 0; 2 (Dugger, Onwenu)
• 2019; 10; 4; 1; 2 (Harris, Bailey)
• 2018; 9; 8; 0; 2 (Wynn, Michel)
• 2017; 4; 1; 0; 1 (Wise)
• 2016; 9; 17; 0; 1 (Thuney, but HM to Brisset; Roberts, Karras, Grugier-Hill)
• 2015; 11; 20; 0; 4 (Brown, Flowers, Mason, Cardona)
• 2014; 9; 6; 0; 2 (Jimmy G, White)
• 2013; 7; 17; 1; 3 (Collins, Ryan, Harmon)
• 2012; 7; 20; 6; 2 (Hightower, C Jones)
• 2011; 9; 20; 0; 2 (Solder, Cannon)
• 2010; 12; 36, 7; 2* (DMcC, Gronk. * for the murderer)
• 2009; 12; 25; 0; 3 (Chung, Vollmer, Edelman)
• 2008; 7; 7; 10; 2 (Mayo, Slater)
• 2007; 9; 6; 2; 1(Meriweather)
• 2006; 10; 8; 4; 1 (Ghost)
• 2005; 7; 28; 8; 4 (Mankins, Kaczur; Sanders, Cassel)
• 2004; 8; 22; 5; 2 (Wilfork, Watson)
• 2003; 10; 33; 5; 4 (Warren, Wilson; Samuel, Koppen)
• 2002; 6; 18; 0; 4 (Graham, Branch, Green, Givens)
• 2001; 10; 25; 10; 2 (Seymour, Light)
• 2000; 10; 22, 15; 1 (TB12)

So they've rarely drafted high-end stars. Since Hightower and Chandler Jones in 2012, they've only had two pro-bowl caliber seasons, one of which was by a punter. But it seems that most years, they get a couple of solid starters and a couple of role players.

 

[ZMart100]

Daniel Ekuale is back with the team. Unlike most of the future contract guys signed, I think he has a good shot at winning a spot out of camp.
View: https://www.youtube.com/watch?v=_90k6jN1-Go