Mikie2times Posted January 28, 2018 Posted January 28, 2018 (edited) I did a study I posted recently on the forum involving career college passing yards and career college rushing yards for QB's as a % of total yardage gained vs correlation to success. Part of the fault in that study was using subjective material to qualify success. This is the same study, just objective criteria. I included career winning % in games started for each subset of data Busts are defined by any QB that did not start at least 64 games (4 seasons) in his career in the NFL. Active players were excluded from the sample on Busts unless near certain 64 game expectation exists (Derek Carr/62 games, etc). Active player are included toward winning%. Draft pick sample is from 2000 on QB's who had 17.5% or more total yards from rushing in college drafted in 1st or 2nd round Record Starting Games in NFL: 227 and 204 (52.7%) Bust Rate: 8 of 10 (80%) Qualifiers: Lamar Jackson QB's who had between .5% and 17.5% total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 1,362 and 1,289 (51.4%) Bust Rate: 12 of 26 (46.2%) Qualifiers: Baker Mayfield, Josh Allen, Sam Darnold QB's drafted who had .5% or less total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 336 and 396 (45.9%) Bust Rate: 10 of 13 (76.9%) Qualifiers: Josh Rosen, Mason Rudolph Analysis: This is one of the most dominant stat profiles in the study on who to avoid. It's a small sample, but you could see some logic behind the data. If you are running as a means to generate offense in college at a high rate you are also using that escape ability as a means to generate offense in the air through unconventional means. You will be less likely to be successful running the ball at the NFL level. Conversely at very low levels most QB's have negative career rushing yards in college. This could lend itself to poor pocket awareness. All and all the bust rate on outliers as it relates to yards gained on the ground is a staggering 18 of 23 (78%) with Rosen, Rudolph, and Jackson all being in outlier profiles. While the bust rate on non outlier profiles is just 46.2%, with prospects that include Mayfield, Allen, and Darnold. QB's with 9,300 or more career passing yards in college drafted in 1st or 2nd round Record Starting Games in NFL: 1,025 and 877 (53.9%) Bust Rate: 10 of 20 (50%) Qualifiers: Mason Rudolph, Baker Mayfield, Josh Rosen QB's between 5,500 and 9,300 passing yards in college drafted in 1st or 2nd round Record starting Games: 490 and 708 (40.9%) Bust Rate: 19 of 24 (79.2%) Qualifiers: Sam Darnold, Lamar Jackson QB's with less than 5,500 career passing yards in college drafted in the 1st or 2nd round Record Starting Games: 410 and 304 (57.4%) Bust Rate: 1 of 5 (20%) Qualifiers: Josh Allen Analysis: Passing yards/production in the air does matter, especially when you partner it with NFL scouts giving you a high grade. Players that are above 9,300 yards in large part have started since day one in college, have been either consistently productive or excessively productive over 2-3 years. More likely to have stayed a full 4 years improving maturation. Rudolph, Mayfield, and Rosen fit this profile, with Mayfield having the highest career passing yardage numbers out of any 1st or 2nd round pick drafted since 2000. Once you go under 5,500 yards you see a clear trend emerge. Very high draft picks. These tend to be the freak type QB's that either explode on the scene or have measurable's off the charts. It's not common to see this occur but we likely will see Allen fall into this profile. Middle of the rode QB's on the production front have been awful since the 2000 draft, 40.9% winning % and nearly an 80% bust rate on 24 players. Think EJ/JP, they fit in this group. They have the measurable's but generally speaking they weren't able to translate that into elite production on the field (Jackson would be an exception to that statement). Scouts should be very cautious of this combination. So what does the passing data look like when you get out of elite draft pick status? Pretty much the same. ..... QB's with 9,300 or more career passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record Starting Games 500 and 416 (54.6%) Bust Rate: 7 of 12 (58.3%) QB's between 5,500 and 9,300 passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record starting Games: 271 and 380 (41.6%) Bust Rate: 16 of 17 (94.1%) QB's with less than 5,500 career passing yards in college drafted outside the top 10 but selected in 1st or 2nd round Record Starting Games: 107 and 60 (64.1%) Bust Rate: 1 of 2 (50%) Analysis: Middle level bust rate is nearly 100%, that number is spared thanks to Jay Cutler. Final Analysis: The only two QB's to avoid all of the high bust rate levels are Mayfield and Allen. Darnold, Rosen, Rudolph all fit in at least one of the larger bust rate groupings. Jackson fits in 2 of them. I know it's far from a sure fire selection guide. I do find some of the bust rate data to be a tad overwhelming to not give it any merit. Last year Mitch fell into the same grouping as Allen, avoiding any major bust categories but in the very low production buckets. Then Watson and Mahomes fell into the same categories as Mayfield with both those QB's also avoiding any of the major bust categories. Edited January 28, 2018 by KzooMike 8 5
White Linen Posted January 28, 2018 Posted January 28, 2018 Agree or not this is some outstanding analysis. Thanks for the effort.
Bills Fan of St Augustine Posted January 28, 2018 Posted January 28, 2018 Did you apply a factor for QBs coming out early? I would think with that happening more often than historically, it may skew your results?
Bakin Posted January 28, 2018 Posted January 28, 2018 Seems pretty arbitrary but interesting nonetheless. 2
TigerJ Posted January 28, 2018 Posted January 28, 2018 So you're saying the Bills should be all in for Josh Allen?
BillsFan4 Posted January 28, 2018 Posted January 28, 2018 Interesting stuff. Thanks for taking the time to post this.
ghostwriter Posted January 28, 2018 Posted January 28, 2018 Interesting but I'm not into analytics when it comes to football. I know a good one when I see one.
billspro Posted January 28, 2018 Posted January 28, 2018 3 hours ago, KzooMike said: I did a study I posted recently on the forum involving career college passing yards and career college rushing yards for QB's as a % of total yardage gained vs correlation to success. Part of the fault in that study was using subjective material to qualify success. This is the same study, just objective criteria. I included career winning % in games started for each subset of data Busts are defined by any QB that did not start at least 64 games (4 seasons) in his career in the NFL. Active players were excluded from the sample on Busts unless near certain 64 game expectation exists (Derek Carr/62 games, etc). Active player are included toward winning%. Draft pick sample is from 2000 on QB's who had 17.5% or more total yards from rushing in college drafted in 1st or 2nd round Record Starting Games in NFL: 227 and 204 (52.7%) Bust Rate: 8 of 10 (80%) Qualifiers: Lamar Jackson QB's who had between .5% and 17.5% total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 1,362 and 1,289 (51.4%) Bust Rate: 12 of 26 (46.2%) Qualifiers: Baker Mayfield, Josh Allen, Sam Darnold QB's drafted who had .5% or less total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 336 and 396 (45.9%) Bust Rate: 10 of 13 (76.9%) Qualifiers: Josh Rosen, Mason Rudolph Analysis: This is one of the most dominant stat profiles in the study on who to avoid. It's a small sample, but you could see some logic behind the data. If you are running as a means to generate offense in college at a high rate you are also using that escape ability as a means to generate offense in the air through unconventional means. You will be less likely to be successful running the ball at the NFL level. Conversely at very low levels most QB's have negative career rushing yards in college. This could lend itself to poor pocket awareness. All and all the bust rate on outliers as it relates to yards gained on the ground is a staggering 18 of 23 (78%) with Rosen, Rudolph, and Jackson all being in outlier profiles. While the bust rate on non outlier profiles is just 46.2%, with prospects that include Mayfield, Allen, and Darnold. QB's with 9,300 or more career passing yards in college drafted in 1st or 2nd round Record Starting Games in NFL: 1,025 and 877 (53.9%) Bust Rate: 10 of 20 (50%) Qualifiers: Mason Rudolph, Baker Mayfield, Josh Rosen QB's between 5,500 and 9,300 passing yards in college drafted in 1st or 2nd round Record starting Games: 490 and 708 (40.9%) Bust Rate: 19 of 24 (79.2%) Qualifiers: Sam Darnold, Lamar Jackson QB's with less than 5,500 career passing yards in college drafted in the 1st or 2nd round Record Starting Games: 410 and 304 (57.4%) Bust Rate: 1 of 5 (20%) Qualifiers: Josh Allen Analysis: Passing yards/production in the air does matter, especially when you partner it with NFL scouts giving you a high grade. Players that are above 9,300 yards in large part have started since day one in college, have been either consistently productive or excessively productive over 2-3 years. More likely to have stayed a full 4 years improving maturation. Rudolph, Mayfield, and Rosen fit this profile, with Mayfield having the highest career passing yardage numbers out of any 1st or 2nd round pick drafted since 2000. Once you go under 5,500 yards you see a clear trend emerge. Very high draft picks. These tend to be the freak type QB's that either explode on the scene or have measurable's off the charts. It's not common to see this occur but we likely will see Allen fall into this profile. Middle of the rode QB's on the production front have been awful since the 2000 draft, 40.9% winning % and nearly an 80% bust rate on 24 players. Think EJ/JP, they fit in this group. They have the measurable's but generally speaking they weren't able to translate that into elite production on the field (Jackson would be an exception to that statement). Scouts should be very cautious of this combination. So what does the passing data look like when you get out of elite draft pick status? Pretty much the same. ..... QB's with 9,300 or more career passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record Starting Games 500 and 416 (54.6%) Bust Rate: 7 of 12 (58.3%) QB's between 5,500 and 9,300 passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record starting Games: 271 and 380 (41.6%) Bust Rate: 16 of 17 (94.1%) QB's with less than 5,500 career passing yards in college drafted outside the top 10 but selected in 1st or 2nd round Record Starting Games: 107 and 60 (64.1%) Bust Rate: 1 of 2 (50%) Analysis: Middle level bust rate is nearly 100%, that number is spared thanks to Jay Cutler. Final Analysis: The only two QB's to avoid all of the high bust rate levels are Mayfield and Allen. Darnold, Rosen, Rudolph all fit in at least one of the larger bust rate groupings. Jackson fits in 2 of them. I know it's far from a sure fire selection guide. I do find some of the bust rate data to be a tad overwhelming to not give it any merit. Last year Mitch fell into the same grouping as Allen, avoiding any major bust categories but in the very low production buckets. Then Watson and Mahomes fell into the same categories as Mayfield with both those QB's also avoiding any of the major bust categories. Can you add Mike White and Lauletta?
SoCal Deek Posted January 28, 2018 Posted January 28, 2018 Really well done. I think it gives great insight into the kind of analysis that NFL teams actually do. I'm always shocked at how spotty the rate of return is on top QBs...but you have to keep rolling the dice none the less. You can't win consistently in the NFL without a legitimate starting QB. (Dying to find out who the one responder says is the guy he 'knows when he sees him' actually is.)
White Linen Posted January 28, 2018 Posted January 28, 2018 8 hours ago, NewDayBills said: Interesting but I'm not into analytics when it comes to football. I know a good one when I see one. With respect, no you don't. It's impossible to know if a college QB will make it in the NFL. 1
JinxedBill1 Posted January 28, 2018 Posted January 28, 2018 Thanks for this. The other thing to consider is the nerves under pressure and ability to read defenses and perform under pressure. This is what coaches get a sense of once they sit with these kids at the senior bowl and get them in for visits. I wish we had that piece of the puzzle but this type of speculation I'm willing to entertain because it is well thought out and objective. Nice job.
Mikie2times Posted January 28, 2018 Author Posted January 28, 2018 Thank you all for the kind words. I will probably do some more of this type of work based on the feedback. 12 hours ago, Bills Fan of Maryland said: Did you apply a factor for QBs coming out early? I would think with that happening more often than historically, it may skew your results? No, I think that requires it's own analysis to validate that statement then maybe it's own analysis with just QB's who came out early. 12 hours ago, TigerJ said: So you're saying the Bills should be all in for Josh Allen? Mayfield is actually the guy I like the most based on the data. He's part of some good sample sets with neutral rushing yards and high career passing yards. Allen is also a guy I like based on the data. He's part of the strongest data sets but they aren't as large. He is likely to go very early. Only Aaron Rogers and Brock Osweiler made it out of the top 10 in the grouping of QB's drafted 1st or 2nd round under 5,500 yards. That's just rare to see happen, but logical. These guys have very little production, very high draft status, clearly the measurable's. I don't like Jackson, he is part of some very negative data sets. His % of rushing yards is over 30%. Only 2 players have been drafted 1st or 2nd round since 2000 with a higher number (Cam and Vince Young). Jackson could win some games, I just don't see him surpassing the bust credentials. More scared on Darnold, data sets with mid level passing production have almost a 80% bust rate on a large sample of picks. I also don't like the high turnover % at USC. Rudolph and Rosen get small downgrades as being part of some negative data sets with smaller samples, but they also belong to some positive ones. 12 hours ago, Bakin said: Seems pretty arbitrary but interesting nonetheless. It can, where do you establish the cut offs? Does this matter? I believe production should matter in college. If a guy is 1st or 2nd round talent he should have been successful winning in college. That was my biggest argument for Watson last year. He torched Bama for over 400 yards B2B seasons and brought that program back to top 5 status. Production should count for something, if, and only if, scouts determine that production also comes with a high draft grade. You won't get a high grade unless you have the skills. I also think the running data could have some merit. It's logical to assume scrambling QB's create more air production because they can scramble. They buy time and create big plays. Will that work as well in the NFL? Likely not. This deludes all forms of production at the college level as far as determining it's ability to carry over. QB's with very poor rushing %'s get sacked a lot in college. Could that speak to issues with pocket awareness? Issues that will get compounded at the next level? Maybe. I don't know. To your point, it is interesting, maybe that's all it is. 9 hours ago, billspro said: Can you add Mike White and Lauletta? Mike White would fit into the high bust rate stat on Rushing yards (-268). He fits into the +9,300 category in passing production 11,262. Lauletta would be omitted from this list by not playing for an FBS school. I don't know how material this information is when looking at all picks. Maybe a future project. A couple other players I did Tyrod Taylor would have fallen in both major bust categories (Rushing% 23.8% and Career Passing Yards, Between 5,500 and 9,300). Russell Wilson would have fell in all positive categories, 10.8% rushing %/Neutral and 11,720 passing yards +9300.
Formerly Allan in MD Posted January 28, 2018 Posted January 28, 2018 1 hour ago, KzooMike said: Lauletta would be omitted from this list by not playing for an FBS school. But he played exceptionally well against a fine cast of FBS players in the Senior Bowl.
Just Joshin' Posted January 28, 2018 Posted January 28, 2018 4 hours ago, Formerly Allan in MD said: But he played exceptionally well against a fine cast of FBS players in the Senior Bowl. With players who had minimal time playing together and generic schemes. So maybe, maybe not. I think this time of the year people look for reasons to find a diamond. It is hard enough when with the top talent to make the right pick let alone a long shot.
Ayjent Posted January 29, 2018 Posted January 29, 2018 I appreciate the effort that went into this, but the statistics are not related enough to anything that leads to a tangible analysis of the necessary skills to be a successful NFL QB. Rushing yards and passing yards for one's college career are just too broad and as the data shows there is no discernible trend and the sample size is small as recognized. There are several statistics that could be looked at but are not typically available for mass consumption to analyze a QB: First down percentage per play when QB passes/runs Third down conversion rate when QB passes/runs Sacks per drop back (this is only part of the story of pocket presence) Air yards per pass attempt Yards per pass attempt Turnover rate per pass/rush attempt Completion percentage on passes of 10 air yards or more (this is a better indicator of accuracy on intermediate and deep throws) Red zone efficiency adjusted for plays resulting in first down and touchdown conversions where QB passes/runs Passes deflected/batted down at the line of scrimmage? Then you have more qualitative measures: What responsibilities did the QB have pre-snap? Was the player aware of what should and should not be done in game situations? Did the player have awareness in the pocket and move to avoid pressure effectively? When the player broke the pocket how often did the player keep passing options open? Did the player have the ability to throw into tight windows? Did the player have the ability to throw receivers open? Did the player show bad throwing technique/decision making under pass pressure? So as much as people would like to rely on TDs/Ints, rushing yards, passing yards, and QB ratings - the more detailed stats which would likely require someone going through loads of tape with these measures in mind, and then there is still some uncertainty because most of the time you are not sure if what happened during a play was coached or the player going off script. Some things are just hard to quantify - like pocket presence and ability to diagnose a defense pre-snap. My guess is if you did this for the QBs drafted during the past 10 years you'd have a much better predictor of success in the NFL. 2
SoFFacet Posted January 29, 2018 Posted January 29, 2018 Kudos on the effort to set this up, but unfortunately these metrics are too limited and the thresholds are too arbitrary. Evaluators broadly agree that Allen, while having some potential, also has an insane amount of red flags. Probably the most bust-worrisome prospect in years. So I would view any analysis that casts him as the relatively safe option as fairly suspect.
LeGOATski Posted January 29, 2018 Posted January 29, 2018 57 minutes ago, SoFFacet said: Kudos on the effort to set this up, but unfortunately these metrics are too limited and the thresholds are too arbitrary. Evaluators broadly agree that Allen, while having some potential, also has an insane amount of red flags. Probably the most bust-worrisome prospect in years. So I would view any analysis that casts him as the relatively safe option as fairly suspect. Well, the whole point of an analysis like this is to limit the guess work. What if Allen becomes a franchise QB? Is there an analysis out there that everyone can point to and say "that's the standard?" At this point, football has proven to be the most difficult sport for analytics to conquer. There are just too many immeasurable variables involved. On 1/27/2018 at 10:49 PM, NewDayBills said: Interesting but I'm not into analytics when it comes to football. I know a good one when I see one. Haha... who's "the one" then? I'd love to know your past picks.
dave mcbride Posted January 29, 2018 Posted January 29, 2018 On 1/27/2018 at 9:16 PM, KzooMike said: I did a study I posted recently on the forum involving career college passing yards and career college rushing yards for QB's as a % of total yardage gained vs correlation to success. Part of the fault in that study was using subjective material to qualify success. This is the same study, just objective criteria. I included career winning % in games started for each subset of data Busts are defined by any QB that did not start at least 64 games (4 seasons) in his career in the NFL. Active players were excluded from the sample on Busts unless near certain 64 game expectation exists (Derek Carr/62 games, etc). Active player are included toward winning%. Draft pick sample is from 2000 on QB's who had 17.5% or more total yards from rushing in college drafted in 1st or 2nd round Record Starting Games in NFL: 227 and 204 (52.7%) Bust Rate: 8 of 10 (80%) Qualifiers: Lamar Jackson QB's who had between .5% and 17.5% total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 1,362 and 1,289 (51.4%) Bust Rate: 12 of 26 (46.2%) Qualifiers: Baker Mayfield, Josh Allen, Sam Darnold QB's drafted who had .5% or less total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 336 and 396 (45.9%) Bust Rate: 10 of 13 (76.9%) Qualifiers: Josh Rosen, Mason Rudolph Analysis: This is one of the most dominant stat profiles in the study on who to avoid. It's a small sample, but you could see some logic behind the data. If you are running as a means to generate offense in college at a high rate you are also using that escape ability as a means to generate offense in the air through unconventional means. You will be less likely to be successful running the ball at the NFL level. Conversely at very low levels most QB's have negative career rushing yards in college. This could lend itself to poor pocket awareness. All and all the bust rate on outliers as it relates to yards gained on the ground is a staggering 18 of 23 (78%) with Rosen, Rudolph, and Jackson all being in outlier profiles. While the bust rate on non outlier profiles is just 46.2%, with prospects that include Mayfield, Allen, and Darnold. QB's with 9,300 or more career passing yards in college drafted in 1st or 2nd round Record Starting Games in NFL: 1,025 and 877 (53.9%) Bust Rate: 10 of 20 (50%) Qualifiers: Mason Rudolph, Baker Mayfield, Josh Rosen QB's between 5,500 and 9,300 passing yards in college drafted in 1st or 2nd round Record starting Games: 490 and 708 (40.9%) Bust Rate: 19 of 24 (79.2%) Qualifiers: Sam Darnold, Lamar Jackson QB's with less than 5,500 career passing yards in college drafted in the 1st or 2nd round Record Starting Games: 410 and 304 (57.4%) Bust Rate: 1 of 5 (20%) Qualifiers: Josh Allen Analysis: Passing yards/production in the air does matter, especially when you partner it with NFL scouts giving you a high grade. Players that are above 9,300 yards in large part have started since day one in college, have been either consistently productive or excessively productive over 2-3 years. More likely to have stayed a full 4 years improving maturation. Rudolph, Mayfield, and Rosen fit this profile, with Mayfield having the highest career passing yardage numbers out of any 1st or 2nd round pick drafted since 2000. Once you go under 5,500 yards you see a clear trend emerge. Very high draft picks. These tend to be the freak type QB's that either explode on the scene or have measurable's off the charts. It's not common to see this occur but we likely will see Allen fall into this profile. Middle of the rode QB's on the production front have been awful since the 2000 draft, 40.9% winning % and nearly an 80% bust rate on 24 players. Think EJ/JP, they fit in this group. They have the measurable's but generally speaking they weren't able to translate that into elite production on the field (Jackson would be an exception to that statement). Scouts should be very cautious of this combination. So what does the passing data look like when you get out of elite draft pick status? Pretty much the same. ..... QB's with 9,300 or more career passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record Starting Games 500 and 416 (54.6%) Bust Rate: 7 of 12 (58.3%) QB's between 5,500 and 9,300 passing yards in college drafted outside the top 10 but selected in the 1st or 2nd round Record starting Games: 271 and 380 (41.6%) Bust Rate: 16 of 17 (94.1%) QB's with less than 5,500 career passing yards in college drafted outside the top 10 but selected in 1st or 2nd round Record Starting Games: 107 and 60 (64.1%) Bust Rate: 1 of 2 (50%) Analysis: Middle level bust rate is nearly 100%, that number is spared thanks to Jay Cutler. Final Analysis: The only two QB's to avoid all of the high bust rate levels are Mayfield and Allen. Darnold, Rosen, Rudolph all fit in at least one of the larger bust rate groupings. Jackson fits in 2 of them. I know it's far from a sure fire selection guide. I do find some of the bust rate data to be a tad overwhelming to not give it any merit. Last year Mitch fell into the same grouping as Allen, avoiding any major bust categories but in the very low production buckets. Then Watson and Mahomes fell into the same categories as Mayfield with both those QB's also avoiding any of the major bust categories. Great analysis, Kzoo; much appreciated. That said, I fell out of love with this sort of thing when Brian Brohm happened. By pretty much every measure, his college stats projected him to Canton.
Ayjent Posted January 29, 2018 Posted January 29, 2018 Just now, LeGOATski said: Well, the whole point of an analysis like this is to limit the guess work. What if Allen becomes a franchise QB? Is there an analysis out there that everyone can point to and say "that's the standard?" At this point, football has proven to be the most difficult sport for analytics to conquer. There are just too many immeasurable variables involved. You are right, but no matter the sport there is some degree of human element (determination and ability to improve, ability to to succeed with different coaching staffs, fit with other players) in any analysis of future potential. Football analytics are still developing, and I'd say that there is still a considerable way for them to go and that they can improve. The eye test in my opinion is merely people saying certain stats of evaluation are just not accurate indicators of success and they are right. However, if you start to break down the players that "look" good and what they do and don't do, you start recognizing that there are stats and qualitative evaluations that can tell a better story.
Mikie2times Posted January 30, 2018 Author Posted January 30, 2018 8 hours ago, dave mcbride said: Great analysis, Kzoo; much appreciated. That said, I fell out of love with this sort of thing when Brian Brohm happened. By pretty much every measure, his college stats projected him to Canton. Thank you, of note, Brain Brohm lives here QB's drafted who had .5% or less total yards from rushing in college drafted in 1st or 2nd round Record Starting Games 336 and 396 (45.9%) Bust Rate: 10 of 13 (76.9%) Qualifiers: Josh Rosen, Mason Rudolph 2
.thumb.png.1b7fb4086290c3afc001a7900b31fab0.png)
Recommended Posts