Fantasy Football - Footballguys Forums
Dr. Dan

Doc's WR1/2 Miss Detector: Breakout Age + Dominator %

Recommended Posts

Posted (edited)

I’ve been trying to determine a predictability score for rookie WR success/failure and I am finding several sites that offer some neat ratings based off of production and break out age. I ran across a post on another forum that stated that over the last 10 years, every top 10-15 ranked WR (minus only 2 exceptions) had a 50th percentile breakout age and 30% dominator rating in their last college year. (This is what I've verified: 79% of the last 10 years falls into this category of anyone finishing in the top 12 wrs in ppr, 91.3% if you lower the dominator rating to 25%)

 

If true, I take this as: not everyone who meets the above criteria becomes a top 10-15 WR, but almost every top 10-15 WR meets the above criteria. Basically, if you're going to draft a WR round 1, you've got a better chance at landing a future WR1 if you can check both of these boxes than if you can't. 

So it got me wondering… what about the 2019 class? What about other classes? Using Player Profiler I created this breakdown:

Rolling 10 years of top 24 finishes and all drafted WRs since 2010

I see this as a great way to rule OUT players rather than to rule them IN when trying to draft a future wr1 or even wr2 in ppr

 

Edited by Dr. Dan
  • Like 11
  • Thanks 2
  • Love 2

Share this post


Link to post
Share on other sites
Posted (edited)
38 minutes ago, wgoldsph said:

Explain "dominator" please

a helpful article

glossary

Quote

College Dominator Rating – first outlined by Frank DuPont in the book Game Plan, the college dominator rating represents a player’s “market share” or his percentage of his team’s offensive production. For example, a 35+% dominator indicates that a wide receiver has the potential to be a team’s No. 1 WR and/or a high caliber contributor. 20-35% indicates a mid-level talent with situational upside. Less than 20% is a red flag.

For wide receivers and tight ends, the dominator rating is the percentage of team receiving production. For running backs, it is the percentage of total offensive production, because running backs are involved in both the running and passing game. College Dominator Rating is not relevant for the quarterback position.

Breakout Age – the younger a person is when he/she first becomes a leader in their respective field, the more likely that person is to go on to become *phenom*enal at their craft. Following this logic, Frank DuPont and Shawn Siegele first examined each wide receiver’s breakout age on RotoViz.com.

The breakout age for wide receivers is defined by their age at midpoint of the college football season when they first posted a Dominator Rating at or above 20% (unless an or other extenuating circumstances prevented the player from playing a full complement of offensive snaps). For tight ends and running backs, a 15% Dominator Rating is necessary to qualify for a breakout designation. Quarterback breakout ages are determined by their age when they first posted a QBR of 50 or higher in a college season, and quarterbacks must average 20 or more action plays per team game played to qualify.

In order to display a more granular percentile rank for comparison purposes, Breakout Ages are listed to the tenth decimal place.

 

 

Edited by Dr. Dan
  • Love 2

Share this post


Link to post
Share on other sites

Ashton Dulin also meets this mark, albeit with inferior competition. You should also look at BMI (26 being the magic number) and speed score has some correlation as well. Breakout age isn’t as set in stone as what you’re suggesting but is massively important. Breakout age compared to draft position is what is relevant. A guy drafted in the first with a breakout age <21.5ish still has a good shot at succeeding. Sub 28% dominator in the last season is pretty much a no go.

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
Posted (edited)
13 hours ago, wgoldsph said:

Explain "dominator" please

Basically just looks at how big of a part of the teams passing game the player received-since college football offenses can vary widely in scheme and success. Oklahoma threw the ball 400 times last year, Georgia Tech threw 126 passes. The dominator is just is a way of evening the playing field.

Edited by Ilov80s

Share this post


Link to post
Share on other sites

Should AJ and DK be skewed higher since they have a teammate that's also at the top of each list, and another talented WR in Lodge? I would assume that they get pushed down the Dominator % just because there's another very talented WR taking targets.  Potentially the same with the Ohio St WRs where there are so many talented players that they all siphon off targets from each other. 

  • Like 3

Share this post


Link to post
Share on other sites
4 hours ago, steelers1080 said:

Should AJ and DK be skewed higher since they have a teammate that's also at the top of each list, and another talented WR in Lodge? I would assume that they get pushed down the Dominator % just because there's another very talented WR taking targets.  Potentially the same with the Ohio St WRs where there are so many talented players that they all siphon off targets from each other. 

Not a bad theory. The fact that both DK and AJ fit the requirements makes their accomplishments even bigger IMO. 

  • Like 2

Share this post


Link to post
Share on other sites

The evidence provided of the 2017 WR draft class  is a small sample size. Too small to draw any conclusions from.

 

Share this post


Link to post
Share on other sites
25 minutes ago, Biabreakable said:

The evidence provided of the 2017 WR draft class  is a small sample size. Too small to draw any conclusions from.

 

People have been using this stuff for almost a decade now. There is plenty of evidence to create minimum thresholds for success. Yes, there are outliers, like Tyreek Hill, but you increase your success by playing within the thresholds. 

  • Like 1

Share this post


Link to post
Share on other sites
Posted (edited)
1 hour ago, Biabreakable said:

The evidence provided of the 2017 WR draft class  is a small sample size. Too small to draw any conclusions from.

 

rotoviz has had articles about it since 2014, John Moore has been doing the Phenom Index for at least that long, which has been a good indicator of success and takes into account breakout age, dominator % and others. Hes used it to go back to as far as 2007 to show it works fairly well. That might be a bigger sample size. it's not 100% accurate but nothing is. These two metrics are the most mentioned when trying to predict success at wr

Edited by Dr. Dan

Share this post


Link to post
Share on other sites

Numbers and stats? What about this “gut feeling” I’m always having? Where do I factor in gut %?

Share this post


Link to post
Share on other sites
50 minutes ago, Snorkelson said:

Where do I factor in gut %?

pant size

  • Like 1
  • Laughing 2

Share this post


Link to post
Share on other sites
4 hours ago, Biabreakable said:

The evidence provided of the 2017 WR draft class  is a small sample size. Too small to draw any conclusions from.

 

The OP states there is data going back the past ten years?

Share this post


Link to post
Share on other sites
1 hour ago, Polka King said:

The OP states there is data going back the past ten years?

To be fair, what I posted was something I read on another forum, which came from a respected poster there however. If I have time I may see if I can go back even a few years to prove this statement right or wrong. I'm willing to bet it holds mostly true. 

  • Like 1

Share this post


Link to post
Share on other sites
4 hours ago, Snorkelson said:

Numbers and stats? What about this “gut feeling” I’m always having? Where do I factor in gut %?

I'm with you, I really can't see any of that actually meaning much.

Share this post


Link to post
Share on other sites
Posted (edited)

Alright... I ran the data for the last 10 years, 2008 through 2018. Any WR who finished in the top 12 for the season in PPR:

Davante Adams: 74th/40.7%
Keenan Allen: 98th/49.5%
Miles Austin: 99th/48.0%
Doug Baldwin (EXCEPTION): 65th/27.0%
Odell Beckham Jr: 70th/34.5%
Anquan Boldin: 67th/41.2%
Dwayne Bowe (EXCEPTION): 38th/35.4%
Antonio Brown: 87th/32.8%
Antonio Bryant: Unable to find
Dez Bryant: 92nd/62.5%
Randall Cobb (EXCEPTION): 63rd/28.2%
Marques Colston: 85th/34.6%
Brandin Cooks: 91st/39.8%
Michael Crabtree: 64th/34.3%
Victor Cruz (EXCEPTION): 19th/39.9%
Eric Decker: 52nd/44.5%
Mike Evans: 89th/30.2%
Larry Fitzgerald: 90th/51.3%
A.J. Green: 61st/42.4%
Josh Gordon (EXCEPTION):82nd/25.0%
Percy Harvin (EXCEPTION): 84th/21.4%
T.Y. Hilton: 93rd/41.9%
Tyreek Hill (EXCEPTION): NA/19.4%
DeAndre Hopkins: 99th/39.3%
DeSean Jackson: 94th/32.1%
Vincent Jackson (EXCEPTION): 47th/70.3%
Alshon Jeffery: 77th/39.7%
Greg Jennings (EXCEPTION):  17th/47.5%
Andre Johnson (EXCEPTION): 60th/28.8%
Calvin Johnson: 66th/55.3%
Steve Johnson (EXCEPTION):  35th/30.3%
Julio Jones: 77th/34.4%
Marvin Jones (EXCEPTION): 80th/26.3%
Jarvis Landry: 71st/40.0%
Brandon Lloyd: 59th/32.2%
Tyler Lockett 66th/44.2%
Jeremy Maclin (EXCEPTION): 83rd/29.5%
Brandon Marshall (EXCEPTION): 27th/42.5%
Rishard Matthews (EXCEPTION): 40th/44.1%
Randy Moss: Unable to find
Jordy Nelson: 57th/49.6%
Hakeem Nicks: 96th/52.2%
Terrell Owens: Unable to find
Sidney Rice: 90th/47.0%
Allen Robinson: 90th/37.9%
Emmanuel Sanders: 80th/34.2%
JuJu Smith-Schuster: 94th/31.9%
Steve Smith: 55th/31.0%
Demaryius Thomas: 74th/70.9%
Adam Thielen: 63rd/45.9%
Michael Thomas (EXCEPTION):25th/39.6%
Mike Wallace (EXCEPTION): 61st/29.6%
Reggie Wayne: 93rd/32.1%
Wes Welker (EXCEPTION): 55th/17.4%
Roddy White (EXCEPTION): 41st/50.5%
Robert Woods: 98th/31.6%

Total with data: 73
Exceptions: 19

Percentage: 79%

 

Data still shows that chances are higher than you have a WR1 with both of these metrics.
You could lower the Dominator threshold a little bit (25%) and capture 91.3% of the players.

Percent of players who had breakout age > 50th percentile: 89% (percent in 2017: 91.6%)

Maybe that's a larger sample size

Edited by Dr. Dan
  • Like 5
  • Thanks 1

Share this post


Link to post
Share on other sites
15 minutes ago, Dr. Dan said:

Alright... I ran the data for the last 10 years, 2008 through 2018. Any WR who finished in the top 12 for the season in PPR:

Davante Adams: 74th/40.7%
Keenan Allen: 98th/49.5%
Miles Austin: 99th/48.0%
Doug Baldwin (EXCEPTION): 65th/27.0%
Odell Beckham Jr: 70th/34.5%
Anquan Boldin: 67th/41.2%
Dwayne Bowe (EXCEPTION): 38th/35.4%
Antonio Brown: 87th/32.8%
Antonio Bryant: Unable to find
Dez Bryant: 92nd/62.5%
Randall Cobb (EXCEPTION): 63rd/28.2%
Marques Colston: 85th/34.6%
Brandin Cooks: 91st/39.8%
Michael Crabtree: 64th/34.3%
Victor Cruz (EXCEPTION): 19th/39.9%
Eric Decker: 52nd/44.5%
Mike Evans: 89th/30.2%
Larry Fitzgerald: 90th/51.3%
A.J. Green: 61st/42.4%
Josh Gordon (EXCEPTION):82nd/25.0%
Percy Harvin (EXCEPTION): 84th/21.4%
T.Y. Hilton: 93rd/41.9%
Tyreek Hill (EXCEPTION): NA/19.4%
DeAndre Hopkins: 99th/39.3%
DeSean Jackson: 94th/32.1%
Vincent Jackson (EXCEPTION): 47th/70.3%
Alshon Jeffery: 77th/39.7%
Greg Jennings (EXCEPTION):  17th/47.5%
Andre Johnson (EXCEPTION): 60th/28.8%
Calvin Johnson: 66th/55.3%
Steve Johnson (EXCEPTION):  35th/30.3%
Julio Jones: 77th/34.4%
Marvin Jones (EXCEPTION): 80th/26.3%
Jarvis Landry: 71st/40.0%
Brandon Lloyd: 59th/32.2%
Tyler Lockett 66th/44.2%
Jeremy Maclin (EXCEPTION): 83rd/29.5%
Brandon Marshall (EXCEPTION): 27th/42.5%
Rishard Matthews (EXCEPTION): 40th/44.1%
Randy Moss: Unable to find
Jordy Nelson: 57th/49.6%
Hakeem Nicks: 96th/52.2%
Terrell Owens: Unable to find
Sidney Rice: 90th/47.0%
Allen Robinson: 90th/37.9%
Emmanuel Sanders: 80th/34.2%
JuJu Smith-Schuster: 94th/31.9%
Steve Smith: 55th/31.0%
Demaryius Thomas: 74th/70.9%
Adam Thielen: 63rd/45.9%
Michael Thomas (EXCEPTION):25th/39.6%
Mike Wallace (EXCEPTION): 61st/29.6%
Reggie Wayne: 93rd/32.1%
Wes Welker (EXCEPTION): 55th/17.4%
Roddy White (EXCEPTION): 41st/50.5%
Robert Woods: 98th/31.6%

Total with data: 73
Exceptions: 19

Percentage: 74%

 

Data still shows that chances are higher than you have a WR1 with both of these metrics.
You could lower the Dominator threshold a little bit (25%) and capture 83.5% of the players.

I've got 73/92 = 79% not 74%. Which only helps the case you're making. Good work. 

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
Posted (edited)
9 minutes ago, barackdhouse said:

I've got 73/92 = 79% not 74%. Which only helps the case you're making. Good work. 

Thank you! eyes crossed after looking all of that up. fixed 

 

So if you lower the threshold to 25% for dominator rating you capture 91% of the field. maybe that's a better number 

Edited by Dr. Dan
  • Like 1

Share this post


Link to post
Share on other sites
7 hours ago, Bojang0301 said:

People have been using this stuff for almost a decade now. There is plenty of evidence to create minimum thresholds for success. Yes, there are outliers, like Tyreek Hill, but you increase your success by playing within the thresholds. 

Since 2014 as stated in the article. 5 years.

I think the logic behind this is pretty sound however I'm not very convinced by the nuts and bolts demonstrated.

The main point of the article was that only 2 of the top 24 WR in 2017 had a breakout age below 50%

That would be more meaningful if you looked at the top 24 of the last 5 seasons. Is 2017 favorable to the argument that break out age is of critical importance? Or is this a normal distribution every year?

I dont know. Maybe looking at a larger sample size makes an even stronger case.

The point is that the author is pushing an idea of analytics but makes a antedoctal argument.

That's fishy.

Then the article goes on to say how Ridley is being over valued and his low college production is a clear sign of why he won't be good in the NFL.

Oops

Share this post


Link to post
Share on other sites
3 hours ago, Polka King said:

The OP states there is data going back the past ten years?

Then why not use that instead of talking about 2017?

Share this post


Link to post
Share on other sites
4 minutes ago, Biabreakable said:

Then why not use that instead of talking about 2017?

At the time I made the post the only thing I could find was that article. I apologize that I didn't have the time to run the data myself until 33 hours later. 

10 years of data supports the 2017 claim, and adding in the dominator percentage narrows it further. 

Share this post


Link to post
Share on other sites
On 4/4/2019 at 12:03 PM, Dr. Dan said:

Players who did not:

Boykin, Miles: 18th percentile age, 30.1% dominator

Brown, Marquis: 31st percentile age, 26.2% dominator

Butler, Hakeem: 30th percentile age, 43.5% dominator

Campbell, Parris: 34th percentile age, 23.5% dominator

Harmon, Kelvin: 31st percentile age, 31.2% dominator

Ridley, Riley: 14th percentile age, 22.2% dominator

Samuel, Deebo: 27th percentile age, 27.2% dominator

McLaurin, Terry: NA percentile age, 17.7% dominator

Thanks for all the leg work on this. Extending beyond the earlier picks, this is also useful to see who might be on the fringe and worth looking into to for the mid and later pick dart throws.  For example, Parris Campbell is 34th percentile and 23% dominator score.  Looking at his situation, he was in an offense with a lot of talented skill players, which might explain a lower dominator score. Similarly, with a log jam of talent at OSU over the years, he may not have had opportunity until he was older.  He may be a player that is higher on someone's rankings accordingly.

Please note I'm not claiming Campbell will necessarily be a mid round pick or that his age and dominator scores are "justified".  I'm only trying to illustrate how this list and method can provide added guidance for the mid rounds of rookie drafts.

Share this post


Link to post
Share on other sites

Agreed that college production is a strong predictor of NFL success for WRs. Other folks who have come up with their own approach to slicing up the data include Football Outsiders with Playmaker Score, Hayden Winks, and me.

Something like the dominator rating is part of most of these systems (the more common name is market share; Playmaker Score uses yards & TDs per team attempt which is very similar).

Breakout age is something different which seems promising. I might go look it up for a bunch of players and think about adding it to my system; it's not currently in my spreadsheets.

  • Like 3

Share this post


Link to post
Share on other sites
4 hours ago, Biabreakable said:

Since 2014 as stated in the article. 5 years.

I think the logic behind this is pretty sound however I'm not very convinced by the nuts and bolts demonstrated.

The main point of the article was that only 2 of the top 24 WR in 2017 had a breakout age below 50%

That would be more meaningful if you looked at the top 24 of the last 5 seasons. Is 2017 favorable to the argument that break out age is of critical importance? Or is this a normal distribution every year?

I dont know. Maybe looking at a larger sample size makes an even stronger case.

The point is that the author is pushing an idea of analytics but makes a antedoctal argument.

That's fishy.

Then the article goes on to say how Ridley is being over valued and his low college production is a clear sign of why he won't be good in the NFL.

Oops

I didn’t read the article. In fact, I don’t put too much stock into what any one person says anymore. Looking at Ridley with what I know today, surface level production was a quality product with concerns of BMI and Age. Final season >28% dominator, breakout age 20.7, speed score 63rd percentile (can be a moving target if a player is a slot reciever). That BMI of 24.9 is very concerning and suggests fragility will occur. Also you’re going to be looking at the Stering Shepard effect with Ridley if he doesn’t take an immediate step this year. By comparison Donte Moncrief is only a half year older than Ridley on his third NFL contract.  People will start claiming he is old before he is even off his rookie contract which will expeditiously lower his value if there isn’t a major breakout before 26. In other words he was and still is a higher risk than DJ Moore, Cam Sutton or Christian Kirk. I would say his rookie year production would have to place him over James Washington though. 

Share this post


Link to post
Share on other sites
On 4/4/2019 at 10:03 AM, Dr. Dan said:

Players who made the cut:

Arcega-Whiteside, JJ: 74th percentile age, 42.7% dominator

Brown, A.J.: 59th percentile age, 32.6% dominator

Hall, Emanuel: 56th percentile age, 35.7% dominator

Harry, N’Keal: 95th percentile age, 43.9% dominator

Humphrey, Lil’Jordan: 54th percentile age, 32.3% dominator

Isabella, Andy: 71st percentile age, 52.2% dominator

Metcalf, D.K.: 72nd percentile age, 32.1% dominator 

 

Players who did not:

Boykin, Miles: 18th percentile age, 30.1% dominator

Brown, Marquis: 31st percentile age, 26.2% dominator

Butler, Hakeem: 30th percentile age, 43.5% dominator

Campbell, Parris: 34th percentile age, 23.5% dominator

Harmon, Kelvin: 31st percentile age, 31.2% dominator

Ridley, Riley: 14th percentile age, 22.2% dominator

Samuel, Deebo: 27th percentile age, 27.2% dominator

McLaurin, Terry: NA percentile age, 17.7% dominator

My formula agrees with the majority of these.

Thumbs up from both: JJ Arcega-Whiteside, AJ Brown, Emanuel Hall, N'Keal harry, Andy Isabella, DK Metcalf
Makes their cut but my formula doesn't like him: Lil'Jordan Humphrey
Don't make their cut but my formula likes him: Hakeem Butler
Don't make their cut, but my formula has him as borderline decent: Marquise Brown, Deebo Samuel
Thumbs down from both: Miles Boykin, Parris Campbell, Kelvin Harmon, Riley Ridley, Terry McLaurin

Going into more detail on the disagreements:

Lil'Jordan Humphrey has terrible athleticism and borderline production. I had some optimism about him earlier in the offseason, but his combine sunk him for me.

Hakeem Butler is an extreme case of a player with a late breakout and then huge production. My inclination is to count this as checking the production box, but I'm open to changing my mind based on historical data. The examples I've found from the first 3 rounds are Ryan Broyles, Terrance Williams, Robert Meachem, Santonio Holmes, A.J. Jenkins, and Anthony Miller. That's 4 failures, 1 modest success, 1 jury still out. Makes me a little more hesitant on Butler, but too small a sample to be a compelling case against him.

Marquise Brown had a 1000 yard season in 2017 which apparently didn't count as a breakout, and 75/1318/10 this year despite missing 2 games, while averaging 13 YPT over those 2 seasons. I think his production was good and this is a problem with relying solely on market share (and apparently not pro-rating).

With Deebo Samuel, I think the main differences are that I'm giving him some credit for his rushing & return production (he has 7 rushing TDs and 4 return TDs over the past 3 years), plus I pro-rate his stats for missed games (2 missed games in 2016 and 1 in 2018). Plus I have a continuous scale rather than sharp cutoffs; I think Samuel comes pretty close to some of the cutoffs that Player Profiler uses. So borderline decent seems appropriate to me.

  • Like 2

Share this post


Link to post
Share on other sites

I find this interesting because the "made the cut" list contains 4 of my top 5. 

Theres definitely more to it than over simplifying with 2 stats to create a hard cut off, but it sure shows some correlation. 

I can always use ZWKs sheet... which I'm more encouraged after hearing it liked Miller a lot. After doing this I was a little bummed that Miller didnt make the cut from 2018s class, but I have high hopes for him to be an exception

 

Share this post


Link to post
Share on other sites

Awesome effort. Thanks!

Especially nice this year for a wr starved superflex I have with picks 8, 14 and 20. One more data point to guide me is always appreciated. 

:hifive:

  • Thanks 1

Share this post


Link to post
Share on other sites
42 minutes ago, Dr. Dan said:

I find this interesting because the "made the cut" list contains 4 of my top 5. 

Theres definitely more to it than over simplifying with 2 stats to create a hard cut off, but it sure shows some correlation. 

I can always use ZWKs sheet... which I'm more encouraged after hearing it liked Miller a lot. After doing this I was a little bummed that Miller didnt make the cut from 2018s class, but I have high hopes for him to be an exception

 

Not really a surprise since the high profile guys are high profile for a reason.   The more useful aspect is discounting highly rated players that don't make the cut and vice versa.

I'm always concerned with the variables beyond those in the analysis.   Somebody mentioned the variation in college schemes and depth of the programs.   Break-out age could hinge on an individual's personal situation.  Injury and college supporting cast could also be a factor.  Since we're talking WRs, caliber of QB play is significant too...  Glancing through the "exceptions" above...  several were paired with very good QBs (Cobb, Jennings, Welker, Thomas, Marshall, Wallace...).   Andre Johnson just missed the "cut" but I would attribute that to talent at the "U" (depth of the program).

Some of the guys on the list made one year appearances in the top "n".  Would be interesting to refine the analysis to players who made the top "n" in multiple years.  

  • Like 2

Share this post


Link to post
Share on other sites
Posted (edited)

How many college WRs met your criteria and didn’t become WR1s (or even pro WRs for that matter)?  I looks to me like you’re working backwards from the conclusion.  If your conclusion is going to be reliable, shouldn’t you work from front to back?  That’s the way you’re intending to use it for future WRs, right?

Edited by Bronco Billy
  • Like 1

Share this post


Link to post
Share on other sites
Posted (edited)
44 minutes ago, Bronco Billy said:

How many college WRs met your criteria and didn’t become WR1s (or even pro WRs for that matter)?  I looks to me like you’re working backwards from the conclusion.  If your conclusion is going to be reliable, shouldn’t you work from front to back?  That’s the way you’re intending to use it for future WRs, right?

I'm sure many. in 2017 there were wr2s that met the criteria. 

My evidence based practice health care mind is looking at this differently than you and most are:

Sensitivity vs specificity. A study with a high sensitivity does a great job ruling something OUT. A study with high specificity does a great job ruling something IN. MRIs are a great example: they have a 96ish percent specificity (if it finds a tear on your MRI theres a 96% chance you have a tear), but it has a very very low (25%ish) sensitivity: if it shows you dont have a tear there is a 25% chance you do not have a tear. Most people dont realize this about MRIs and your MD is unlikely to tell you because it costs 3k. I digress...

 

I see this method as having a high SENSITIVITY: it does a great job at ruling OUT those who are unlikely to become WR1s rather than ruling them IN. That's why I said ablve: not all WRs who meet this threshold become wr1s, but almost all wr1s meet this threshold. 

 

The reason I like this method: there is no fool proof way to predict a hit, so maybe working backwards is the better approach; why not focus on minimizing the misses rather than maximizing the hits? 

We have threads dedicated toward bashing analysists who "get it wrong," and we say "well everyone has misses." My rationale is of you can minimize the misses then maybe you're doing better than those trying to predict the hits. 

 

Maybe we are looking for a method with high specificity, and missing the methods with high sensitivity

This method will give you insight on who to avoid, rather than who to pick. 

If someone took a bet that no more than 1 of the 8 "does not meet" has a top 12 ppr fantasy season, you're probably going to win that bet. Which is interesting because at least half of those guys are considered 1st/early 2nd round picks (Butler, Harmon, Brown are in some peoples' top 3-5)

Edited by Dr. Dan
  • Like 5

Share this post


Link to post
Share on other sites
2 minutes ago, Dr. Dan said:

I'm sure many. in 2017 there were wr2s that met the criteria. 

My evidence based practice health care mind is looking at this differently than you and most are:

Sensitivity vs specificity. A study with a high sensitivity does a great job ruling something OUT. A study with high specificity does a great job ruling something IN. MRIs are a great example: they have a 96ish percent specificity (if it finds a tear on your MRI theres a 96% chance you have a tear), but it has a very very low (25%ish) sensitivity: if it shows you dont have a tear there is a 25% chance you do not have a tear. Most people dont realize this about MRIs and your MD is unlikely to tell you because it costs 3k. I digress...

 

I see this method as having a high SENSITIVITY: it does a great job at ruling OUT those who are unlikely to become WR1s rather than ruling them IN. That's why I said ablve: not all WRs who meet this threshold become wr1s, but almost all wr1s meet this threshold. 

 

The reason I like this method: there is no fool proof way to predict a hit, so maybe working backwards is the better approach; why not focus on minimizing the misses rather than maximizing the hits? 

We have threads dedicated toward bashing analysists who "get it wrong," and we say "well everyone has misses." My rationale is of you can minimize the misses then maybe you're doing better than those trying to predict the hits. 

 

Maybe we are looking for a method with highspecificity, and missing the methods with high sensitivity

This method will give you insight on who to avoid, rather than who to pick. 

 

The point being that there could be all sorts if misses that fall within what you consider criteria for hits that you may not be counting.  I understand what you’re trying to accomplish, but it’s only half the work that needs to be done and the remainder is the dirty work.  Hell, there’s a whole batch of allegedly credible high profile scientists who have done the exact same thing and have drawn some very high impact conclusions without following through, and when the follow through is done the veracity goes right out the window.

  • Like 1

Share this post


Link to post
Share on other sites
3 minutes ago, Bronco Billy said:

 

The point being that there could be all sorts if misses that fall within what you consider criteria for hits that you may not be counting.  I understand what you’re trying to accomplish, but it’s only half the work that needs to be done and the remainder is the dirty work.  Hell, there’s a whole batch of allegedly credible high profile scientists who have done the exact same thing and have drawn some very high impact conclusions without following through, and when the follow through is done the veracity goes right out the window.

You're still looking at it as a method that detects hits. Its a method that detects misses. 

I edited my post to show what I mean: statistics would say no more than 1 of the 8 in my OP for "misses" will have a top 12 ppr wr season. 

 

Yes, its half the work/story and it would be interesting to test the specificity of this, however I'm guessing we will find results all over the place. considering 91% of wr1s meet the criteria of >50% age and >25% DR, I would guess a high percent of wr2 and 3s would meet this as well. And I'm curious how any exceptions would land into the wr2/3 range. 

Seems like way too much work for me to do, although I may try it. 

 

 

 

How I plan to use this is: I likely wont draft anyone in round 1 who doesnt meet a 50% age + 25% DR

A good example is Lil'Jordan Humphrey falls into the "meets criteria" category, but hes on my avoid list for other reasons (you are right- this method is only half the work)

Butler was a potential 1st round target for me. I am now avoiding. Even though there are more factors, my method above tells me of he is to become a wr1 someday. he would be roughly only of the 9% of exceptions from the last 10 years. I dont like those odds. 

 

If you're looking for a way to predict a hit, you're in for a lot of dirty work, as you said. Theres a lot of work to be done.

If you're on the fence between 2 guys, and 1 meets this criteria where another does not, then maybe this helps narrow it down 

  • Like 1

Share this post


Link to post
Share on other sites
2 hours ago, Bronco Billy said:

How many college WRs met your criteria and didn’t become WR1s (or even pro WRs for that matter)?  I looks to me like you’re working backwards from the conclusion.  If your conclusion is going to be reliable, shouldn’t you work from front to back?  That’s the way you’re intending to use it for future WRs, right?

 

On 4/4/2019 at 10:03 AM, Dr. Dan said:

 

 

If true, I take this as: not everyone who meets the above criteria becomes a top 10-15 WR, but almost every top 10-15 WR meets the above criteria. Basically, if you're going to draft a WR round 1, you've got a better chance at landing a future WR1 if you can check both of these boxes than if you can't. 

 

I feel like the bolded in the OP already addressed the concern. Nobody here is making the claim that meeting the criteria will result in a WR1 in the NFL. 

Share this post


Link to post
Share on other sites
14 hours ago, Dr. Dan said:

Alright... I ran the data for the last 10 years, 2008 through 2018. Any WR who finished in the top 12 for the season in PPR:

Davante Adams: 74th/40.7%
Keenan Allen: 98th/49.5%
Miles Austin: 99th/48.0%
Doug Baldwin (EXCEPTION): 65th/27.0%
Odell Beckham Jr: 70th/34.5%
Anquan Boldin: 67th/41.2%
Dwayne Bowe (EXCEPTION): 38th/35.4%
Antonio Brown: 87th/32.8%
Antonio Bryant: Unable to find
Dez Bryant: 92nd/62.5%
Randall Cobb (EXCEPTION): 63rd/28.2%
Marques Colston: 85th/34.6%
Brandin Cooks: 91st/39.8%
Michael Crabtree: 64th/34.3%
Victor Cruz (EXCEPTION): 19th/39.9%
Eric Decker: 52nd/44.5%
Mike Evans: 89th/30.2%
Larry Fitzgerald: 90th/51.3%
A.J. Green: 61st/42.4%
Josh Gordon (EXCEPTION):82nd/25.0%
Percy Harvin (EXCEPTION): 84th/21.4%
T.Y. Hilton: 93rd/41.9%
Tyreek Hill (EXCEPTION): NA/19.4%
DeAndre Hopkins: 99th/39.3%
DeSean Jackson: 94th/32.1%
Vincent Jackson (EXCEPTION): 47th/70.3%
Alshon Jeffery: 77th/39.7%
Greg Jennings (EXCEPTION):  17th/47.5%
Andre Johnson (EXCEPTION): 60th/28.8%
Calvin Johnson: 66th/55.3%
Steve Johnson (EXCEPTION):  35th/30.3%
Julio Jones: 77th/34.4%
Marvin Jones (EXCEPTION): 80th/26.3%
Jarvis Landry: 71st/40.0%
Brandon Lloyd: 59th/32.2%
Tyler Lockett 66th/44.2%
Jeremy Maclin (EXCEPTION): 83rd/29.5%
Brandon Marshall (EXCEPTION): 27th/42.5%
Rishard Matthews (EXCEPTION): 40th/44.1%
Randy Moss: Unable to find
Jordy Nelson: 57th/49.6%
Hakeem Nicks: 96th/52.2%
Terrell Owens: Unable to find
Sidney Rice: 90th/47.0%
Allen Robinson: 90th/37.9%
Emmanuel Sanders: 80th/34.2%
JuJu Smith-Schuster: 94th/31.9%
Steve Smith: 55th/31.0%
Demaryius Thomas: 74th/70.9%
Adam Thielen: 63rd/45.9%
Michael Thomas (EXCEPTION):25th/39.6%
Mike Wallace (EXCEPTION): 61st/29.6%
Reggie Wayne: 93rd/32.1%
Wes Welker (EXCEPTION): 55th/17.4%
Roddy White (EXCEPTION): 41st/50.5%
Robert Woods: 98th/31.6%

Total with data: 73
Exceptions: 19

Percentage: 79%

 

Data still shows that chances are higher than you have a WR1 with both of these metrics.
You could lower the Dominator threshold a little bit (25%) and capture 91.3% of the players.

Percent of players who had breakout age below 50th percentile: 89% (percent in 2017: 91.6%)

Maybe that's a larger sample size

Did a little thought experiment to narrow the focus a bit.

Out of the "exception" players, I count 6 with longevity of WR1 status based on my memory.  Most of the others might have had a good year or two that maybe put them in the conversation but ultimately they returned to good WR2's rather than WR1's.  Those WR's are Roddy White, Michael Thomas, Brandon Marshall, Andre Johnson, Tyreek Hill, Josh Gordon (fell out of WR1 status due to off the field, but I don't doubt he'd be in consideration still if he hadn't had those problems). 

Out of the players that made the critera I count 19 that had longevity of being viewed as a WR1.  Again based on my memory.  Davante Adams, Keenan Allen, Odell Backham Jr, Antonio Brown, Dez Bryant, Brandin Cooks, Mike Evans, Larry Fitzgerald, AJ Green, DeAndre Hopkins TY Hilton, Calvin Johnson, Julio Jones, Randy Moss, Terrell Owens, Jordy Nelson, JuJu Smith-Schuster, Demaryius Thomas, Reggie Wayne.  ***Honorable mentions just to add a few of possibly making the cut in your own view; Robert Woods, Adam Thielen, Alshon Jeffery***

Seems to me when/if they exceptions to the rule produce well, you should probably sell when the longevity of dominance isn't likely to be there.  

  • Thanks 1

Share this post


Link to post
Share on other sites
17 minutes ago, Zyphros said:

Seems to me when/if they exceptions to the rule produce well, you should probably sell when the longevity of dominance isn't likely to be there.  

This was next on my list to take a look at as it's something I noticed as well. Thank you!

Share this post


Link to post
Share on other sites
Posted (edited)

I ran the numbers for WR1 and WR2 for the last 10 years and found:
 

50% Age + >30.0% DR: 62% fell into the category
50% Age + >25.0% DR: 70% fell into the category

I saw the same thing with the exceptions that I did with the WR1 group- the majority of those who made it did not stay long, many only 1 season.

 

 

Davante Adams: 74th/40.7%
Nelson Agholor: 56th/34.2%
Keenan Allen: 98th/49.5%
Robby Anderson: 55th/34.8%
Miles Austin: 99th/48.0%
Tavon Austin (EXCEPTION): 99th/26.3%
Doug Baldwin (EXCEPTION): 65th/27.0%
Odell Beckham Jr: 70th/34.5%
Kelvin Benjamin (EXCEPTION): 6th/29.3%
Benard Berrian: Unable to find
Kenny Britt: 91st/36.1%
Anquan Boldin: 67th/41.2%
Dwayne Bowe (EXCEPTION): 38th/35.4%
Tyler Boyd: 92nd/47.2%
Antonio Brown: 87th/32.8%
John Brown (EXCEPTION): 27th/42.4%
Antonio Bryant: Unable to find
Dez Bryant: 92nd/62.5%
Randall Cobb (EXCEPTION): 63rd/28.2%
Laveranues Coles: Unable to find
Marques Colston: 85th/34.6%
Brandin Cooks: 91st/39.8%
Amari Cooper: 99th/47.2%
Riley Cooper (EXCEPTION): 15th/30.6%
Michael Crabtree: 64th/34.3%
Victor Cruz (EXCEPTION): 19th/39.9%
Eric Decker: 52nd/44.5%
Stephon Diggs: 97th/36.1%
Donald Driver: Unable to find
Braylon Edwards: Unable to find
Mike Evans: 89th/30.2%
Larry Fitzgerald: 90th/51.3%
Michael Floyd: 94th/39.4%
Devin Funchess: 84th/41.4%
Pierre Garcon: 88th/37.7%
A.J. Green: 61st/42.4%
Chris Godwin:  79th/34.9&
Kenny Golladay: 70th/41.8%
Josh Gordon (EXCEPTION):82nd/25.0%
Percy Harvin (EXCEPTION): 84th/21.4%
T.Y. Hilton: 93rd/41.9%
Tyreek Hill (EXCEPTION): NA/19.4%
Santonio Holmes (EXCEPTION): 25th/48.6%
DeAndre Hopkins: 99th/39.3%
Allen Hurns: 70th/28.8%
DeSean Jackson: 94th/32.1%
Vincent Jackson (EXCEPTION): 47th/70.3%
Alshon Jeffery: 77th/39.7%
Greg Jennings (EXCEPTION):  17th/47.5%
Andre Johnson (EXCEPTION): 60th/28.8%
Calvin Johnson: 66th/55.3%
Chad Johnson: Unable to find
Steve Johnson (EXCEPTION):  35th/30.3%
James Jones (EXCEPTION): 11th/44.1%
Julio Jones: 77th/34.4%
Marvin Jones (EXCEPTION): 80th/26.3%
Johnny Knox: Unable to find
Brandon Lafell (EXCEPTION): 19th/45.7%
Jarvis Landry: 71st/40.0%
Brandon Lloyd: 59th/32.2%
Tyler Lockett 66th/44.2%
Jeremy Maclin (EXCEPTION): 83rd/29.5%
Mario Manningham: 84th/47.9%
Brandon Marshall (EXCEPTION): 27th/42.5%
Derrick Mason: Unable to find
Jordan Matthews: 100th/48.3%
Rishard Matthews (EXCEPTION): 40th/44.1%
Robert Meachem (EXCEPTION): 17th/41.8%
Lance Moore: 63rd/33.4%
Randy Moss: Unable to find
Santana Moss (EXCEPTION): 84th/24.8%
Jordy Nelson: 57th/49.6%
Hakeem Nicks: 96th/52.2%
Terrell Owens: Unable to find
Terrelle Pryor: Unable to find
Calvin Ridley (EXCEPTION): 45th/30.1%
Sidney Rice: 90th/47.0%
Allen Robinson: 90th/37.9%
Laurent Robinson: Unable to find
Eddie Royal (EXCEPTION): 56th/21.7%
Emmanuel Sanders: 80th/34.2%
JuJu Smith-Schuster: 94th/31.9%
Cecil Shorts (EXCEPTION): 45th/40.8%
Mike Sims-Walker: Unable to find
Steve Smith: 55th/31.0%
Torey Smith: Unable to find
Golden Tate: 61st/44.3%
Demaryius Thomas: 74th/70.9%
Adam Thielen: 63rd/45.9%
Michael Thomas (EXCEPTION):25th/39.6%
Hines Ward: Unable to find
Nate Washington: 90th/71.8%
Mike Wallace (EXCEPTION): 61st/29.6%
Kevin Walter: Unable to find
Sammy Watkins: 99th/34.2%
Reggie Wayne: 93rd/32.1%
Wes Welker (EXCEPTION): 55th/17.4%
Roddy White (EXCEPTION): 41st/50.5%
Mike William TB: Unable to find
Mike Williams LAC: 99th/34.2%
Tyrell Williams (EXCEPTION): 49th/30.5%
Robert Woods: 98th/31.6%

 

Total 102:
No Data: 18
Total with Data: 84
Exceptions @ >30.0% DR: 32
Percent Met: 62%


Exceptions @ > 25% DR: 25
Percent Met: 70%


I think this supports the rule out claim even further... the drop from WR1 to WR1 and WR2 with meeting the criteria is significant (91% to 70%)

Edited by Dr. Dan

Share this post


Link to post
Share on other sites
Posted (edited)
52 minutes ago, Zyphros said:

Did a little thought experiment to narrow the focus a bit.

Out of the "exception" players, I count 6 with longevity of WR1 status based on my memory.  Most of the others might have had a good year or two that maybe put them in the conversation but ultimately they returned to good WR2's rather than WR1's.  Those WR's are Roddy White, Michael Thomas, Brandon Marshall, Andre Johnson, Tyreek Hill, Josh Gordon (fell out of WR1 status due to off the field, but I don't doubt he'd be in consideration still if he hadn't had those problems). 

Out of the players that made the critera I count 19 that had longevity of being viewed as a WR1.  Again based on my memory.  Davante Adams, Keenan Allen, Odell Backham Jr, Antonio Brown, Dez Bryant, Brandin Cooks, Mike Evans, Larry Fitzgerald, AJ Green, DeAndre Hopkins TY Hilton, Calvin Johnson, Julio Jones, Randy Moss, Terrell Owens, Jordy Nelson, JuJu Smith-Schuster, Demaryius Thomas, Reggie Wayne.  ***Honorable mentions just to add a few of possibly making the cut in your own view; Robert Woods, Adam Thielen, Alshon Jeffery***

Seems to me when/if they exceptions to the rule produce well, you should probably sell when the longevity of dominance isn't likely to be there.  

When exactly did Michael Thomas and Tyreek Hill return to being good WR2's?  And Andre Johnson didn't have WR1 longevity?

Interesting thread.  I value statistical analysis and there's no denying a trend here. 

Edited by skinfanjon

Share this post


Link to post
Share on other sites
Posted (edited)
4 minutes ago, skinfanjon said:

When exactly did Michael Thomas and Tyreek Hill return to being good WR2's?

I believe he is listing those as the 6 exceptions that retained WR1 status vs the other exceptions that fell to wr2 or worse status

 

ETA: I read it the same way the first time too

Edited by Maven25
  • Like 1

Share this post


Link to post
Share on other sites
1 minute ago, Maven25 said:

I believe he is listing those as the 6 exceptions that retained WR1 status vs the other exceptions that fell to wr2 or worse status

 

ETA: I read it the same way the first time too

Ah yeah I see now.  My bad.  Thanks

  • Like 1

Share this post


Link to post
Share on other sites
10 hours ago, Bojang0301 said:

I didn’t read the article. In fact, I don’t put too much stock into what any one person says anymore. Looking at Ridley with what I know today, surface level production was a quality product with concerns of BMI and Age. Final season >28% dominator, breakout age 20.7, speed score 63rd percentile (can be a moving target if a player is a slot reciever). That BMI of 24.9 is very concerning and suggests fragility will occur. Also you’re going to be looking at the Stering Shepard effect with Ridley if he doesn’t take an immediate step this year. By comparison Donte Moncrief is only a half year older than Ridley on his third NFL contract.  People will start claiming he is old before he is even off his rookie contract which will expeditiously lower his value if there isn’t a major breakout before 26. In other words he was and still is a higher risk than DJ Moore, Cam Sutton or Christian Kirk. I would say his rookie year production would have to place him over James Washington though. 

So you don't think 92 targets 64 receptions 821 yards and 10 TD as a rookie is impressive? He needs to take the next step?

His production nothing like Sterling Shephards rookie year. Ridley had 8.9 yards per target as a rookie. Shephard had 6.5

 

Share this post


Link to post
Share on other sites
16 hours ago, Dr. Dan said:

Thank you! eyes crossed after looking all of that up. fixed 

 

So if you lower the threshold to 25% for dominator rating you capture 91% of the field. maybe that's a better number 

Well sure if you move the chains the percentage increases.

While I appreciate your effort looking at a larger sample size you only looked at the top 12 WR which is important, but a higher goal than what the original article was talking about, the top 24 WR in anu given year. You are going to miss a lot of players with such a high cut off. Which seems counter productive to the questions of how much does market share and break out age matter.

If you reduce the threshold further you are just are putting a larger sample of players on one side and isolating the fewer players who did not produce to that threshold. You are creating more forest and losing sight of trees.

Share this post


Link to post
Share on other sites
Posted (edited)
11 minutes ago, Biabreakable said:

 

While I appreciate your effort looking at a larger sample size you only looked at the top 12 WR which is important, but a higher goal than what the original article was talking about, the top 24 WR in anu given year. You are going to miss a lot of players with such a high cut off. Which seems counter productive to the questions of how much does market share and break out age matter.

 

I did this about an hour ago above

 

The reason I cut it at wr1 originally is because if you're drafting a wr round 1, you want a future wr1

Edited by Dr. Dan

Share this post


Link to post
Share on other sites
Posted (edited)
29 minutes ago, Biabreakable said:

Well sure if you move the chains the percentage increases.

 

this is what a predictive model does, tries to narrow into the highest predictability percentage by "moving the chains"

for the top 24, the percentage is probably not significant (62 vs 70%). if anything it shows that if someone falls outside of both of these metrics, they are more likely to end up as a wr2. which may not be a bad thing, but if I'm drafting a wr high round 1 I'm hoping to avoid someone who may never be a wr1

Edited by Dr. Dan

Share this post


Link to post
Share on other sites
2 hours ago, Zyphros said:

Did a little thought experiment to narrow the focus a bit.

Out of the "exception" players, I count 6 with longevity of WR1 status based on my memory.  Most of the others might have had a good year or two that maybe put them in the conversation but ultimately they returned to good WR2's rather than WR1's.  Those WR's are Roddy White, Michael Thomas, Brandon Marshall, Andre Johnson, Tyreek Hill, Josh Gordon (fell out of WR1 status due to off the field, but I don't doubt he'd be in consideration still if he hadn't had those problems). 

Out of the players that made the critera I count 19 that had longevity of being viewed as a WR1.  Again based on my memory.  Davante Adams, Keenan Allen, Odell Backham Jr, Antonio Brown, Dez Bryant, Brandin Cooks, Mike Evans, Larry Fitzgerald, AJ Green, DeAndre Hopkins TY Hilton, Calvin Johnson, Julio Jones, Randy Moss, Terrell Owens, Jordy Nelson, JuJu Smith-Schuster, Demaryius Thomas, Reggie Wayne.  ***Honorable mentions just to add a few of possibly making the cut in your own view; Robert Woods, Adam Thielen, Alshon Jeffery***

Seems to me when/if they exceptions to the rule produce well, you should probably sell when the longevity of dominance isn't likely to be there.  

I had a very similar impression looking through the list.

Great thread @Dr. Dan and others!

Share this post


Link to post
Share on other sites
1 hour ago, Biabreakable said:

So you don't think 92 targets 64 receptions 821 yards and 10 TD as a rookie is impressive? He needs to take the next step?

His production nothing like Sterling Shephards rookie year. Ridley had 8.9 yards per target as a rookie. Shephard had 6.5

 

It was a very impressive rookie year but that target to TD ratio is unsustainable- a bit of an anomaly. Based on his number of targets and now many RZ targets he got, he should have scored 3-4 TDs. 

As for yards per target, Shep was in the slot with a worse QB so it’s not a surprise he was lower. 

Share this post


Link to post
Share on other sites
Posted (edited)

This thread took me a couple of times to concentrate enough to see what you were saying, but now that I'm able to read it with the required concentration, it's an awesome tool. Thanks.  

Edited by rockaction
  • Like 3

Share this post


Link to post
Share on other sites
Just now, rockaction said:

This thread took me a couple of times to concentrate enough to see what you were saying, but now that I'm able to read it with some fortitude, it's an awesome tool. Thanks.  

I think the key is just looking at it as a way to reduce risk. As we have seen with Michael Thomas, Tyreek Hill, Roddy White, etc. it is possible for a WR didn't produce at a young age in college or didn't dominate their college passing game to develop into a WR1 for fantasy. However, it seems one is safer betting on the players that have done it. They won't all work out either but it's a better gamble.

  • Like 1

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.