Bob Magaw
Footballguy
i did a study last year of the 2006 & the first half split of 2007, looking at some of the top scorers on offense & defense, breaking them down positionionally by (among other things) age & draft pedigree... i plan to do this again soon for the past three seasons...
without looking at it, it seemed that some positions (QB) showed far less of a correlation between pedigree and playing at a high level (brady a 6th rounder, romo a UFA, etc)...
at other positions (LB), michael boley was the only non-day one (when this meant after three rounds) pick out of the 17 non-UFA top 20 LBs at the 2007 half-season split...
this led to thinking about possible implications of the data...
but first, lets make clear what this linkage study (at least in its inception) ISN'T trying to accomplish...
obviously the sample size noted above of just a season & a half is laughably small... i'll look back three years soon, and leave to other interested researchers & number crunchers to backtest some of these ideas, in case it might help for extrapolation and predictive purposes... i may try to go back further, & plan to track it going forward... the league seems to go through cyclical changes in trends, & it is unclear of what use it would be to study stats from decades ago when for instance the spread offense wasn't in as wide a use... & to what degree yielded defensive stats from a different era would be relevant to now (perhaps the idea of whether this last point is valid, & if so, what time segments & measures would be useful or non-useful for such a study, could almost be a whole separate topic & subject for a later thread)...
in the above study, i looked at top 20 primary defensive positions (DL/LB/DB)... my rationale was that in 10 team leagues, that would yield two starters at each position, and also once the scoring tables got into the 30s-40s, distinctions began to get increasingly smaller & imo negligible... i wanted to study difference makers...
on offense, i went with top 10 QBs, 20 RBs, 30 WRs & 10 TEs (again, with the rationale that this would yield the full contingent of starters in standard leagues)... but the point being, and this goes for offense & defense, the sample looked at for potential pedigree - starting/suceeding linkages (and possible differences by position) was the top 10-30... so this study makes no claims on that set comprised of ALL drafted players...
of what use it than...
since with each succeeding rookie draft that is the place usually begun from in evaluating who to prioritize in the draft... attempting to forecast who will stick, make the roster & have a long career & who won't... the above study doesn't address (that said, i'd like to see such a study conducted over time)...
perhaps a way this info could be used, is in dynasty applications with vets, when trying to grade & rank players by incorporating pedigree type information... IF there were a linkage such as alluded to above, it might be helpful for parsing and reducing top 30-40 lists by postion, possibly winnowing out lesser pedigreed prospects, as a sort of tie-breaker, in a reverse engineering sense... the caveat is that like with all incorporation of historical & statistical data into player evaluation, it should be combined & OVERLAYED on top of things like measurables & scouting info, film breakdown, etc, on a CASE-BY-CASE basis...
some other general thoughts about positions, that may have use for redraft & dynasty purposes, offense & defense...
i was trying to think of a more comprehensive but simpler overarching framework to help understand &, if it isn't too pretentious, attempt an explanation why some positions catch on faster than others (& maybe putting such a study on more rigorous statistical & historical footing)...
one thing i noticed recently is a kind of mirroring between OL/DL, RB/LB & WR/DB (QBs & TEs are sort of outliers to the framework, for reasons to be explained)...
the OL doesn't matter for offensive scoring purposes in MOST leagues (except indirectly in that attending closely to best & worst OLs can be helpful in breaking down the prospects of skill position players by team), though DL does in IDP variants... the line of course are responsible for either blocking or escaping blocks on each other, attacking the QB & RB or thwarting those attempts... so there is another kind of connection & linkage... one thing respective OL & DL seem to have in common is that it generally takes a few years to hit their stride... this can be explained by the fact that lineman, in addition to matters of technique & scheme comprehension (where it is a given that the vast majority of incoming rookies are undoubtedly raw), they need to undergo an almost metamorphosis of physical development & maturation, before they can even hold their own, let alone star... hopefully this part isn't too controversial...
one point of possible convergence with RB/LB was that there seemed to be a lot of examples where young players did extremely well (for instance last season, look no further than adrian peterson & patrick willis)... the usual explanation for RBs, and it seems reasonable to me, is that RBs are a position where attributes such as pure, natural talent and instincts appear to weigh more heavily... one other obvious, but maybe neglected point, imo, is that grabbing a ball that is handed to you by a QB is inherently easier to initiate a play where they are off to the races than the case of a pass play, which needs a highly synchronized sequence of events to take place... while flipping a pass to the RB in the flat may not seem like a big deal, throwing an intermediate pass across the middle which involves a QB passing the ball over some levels of the defense and in front of others, and squeezing it between several defenders, to a WR who needs to escape the jam, run a precise route, concentrate in traffic & often run according to precise timing, in some cases with the QB releasing the ball before the WR makes his break... decidedly IS a big deal... no wonder it usually takes QBs & WRs a few years to make an impact...
LBs may mirror RBs on the defensive side, in that they seem to favor talent & instinct, & they can produce at a high level relatively faster... one tie-in i wanted to make between the results & data from the study above at LB position, with this relative positional difference model... while the time study is small so far, after it is expanded, it will be interesting to see if there is a linkage between pedigree and starting/success... IF so, it would be pretty much a tautology to point out that this suggests the scouts seem to have a good handle on what scouting attributes tend to lead to success... it would be a different point to suggest that IF this is the case, we should respect pedigree information (at least for those positions in which such a breakdown is useful)...
WRs have already been covered to an extent... DBs may fall somewhere between DL & LB in terms of their rate of development arc & trajectory... i would expect them to take a little longer to catch on than LBs, but perhaps quicker than DL... while many can compete physically in less than the 2-3 years it can take many lineman to compete at a high level, they have some disadvantages relative to LBs... coverage schemes & responsibilities for young CBs & safeties can be very complex at the pro level, & may take a year or two to digest... one disadvantage DBs may have relative to WRs (when compared to RBs vs. LBs) is that if a WR isn't jammed, they can build up a lot of speed downfield, & can better leverage the fact that they know where they are going & DBs don't... with case of LBs, their straightest path to the RB may often be shorter than that of a safety trying to get across the field to cover a WR... and the RBs have to fight through more traffic & are impeded more than WR (where they can't be downfield, BY RULE
), making the LBs job easier than that of DB...
QBs & TEs were called outliers to this model & i promised an explanation...
QBs are different on one level because they are the nerve center of the offense (courtesy of george plimpton in paper lion), and have to know what EVERYBODY is doing, in a way that, for instance, RBs & LBs don't have to... another point about possible reasons why there might seem to be a possible greater linkage & "dependence" on pedigree for success with some positions more than others... it was suggested that scouts might have a better handle on predicting how elite LBs would fare in the transition from college to pro that is so difficult for QB... in the case of QBs, the ability to be calm under pressure, process information quickly, the ability to see the field, the timing to squeeze the ball through tighter windows and more quickly, the physical courage & toughness to withstand a beating... these are very difficult to reduce to a test that would parallel the 40, vertical & broad jump, cone & shuttle drills, benchpress, which are tests yielding physical measurables of speed, quickness, strength, explosiveness, etc... many/most of the cited QB traits are mental & intangibles...
TE may be different than other positions in that they are sort of a hybrid OL/WR position, that has to be physical enough to block in line, but athletic enough to split the safeties on the deep seam route... this is a generalization, as there are receiving specialist TEs like scheffler that aren't very good blockers, as well as blocking specialists... it is important to make a distinction between poor blocking TEs that get to play & get scoring opps, & those for which blocking failures are an obstacle to playing time... this must differ by team, coaching philosophy, scheme, etc...
i will try & relay more of what i found in previous study by positions below, & plan to write separate articles on offense & defense this summer following up here...
* age is also critically important in dynasty (& to a lesser degree in redraft)... we all know by osmosis RBs are near done at 30... but it would be helpful to have more rigorous data on potential differences in ALL the positions (even breaking out DT/DE, MLB/ILB & OLB, S & CB on defense)... it didn't bear on this immediate subject, and is a separate area, but it will be included in my study to follow, on offense as well...
i had already done the pedigree study, but recently some thoughts coalesced around the offense/defense position parallel model... hopefully the attempt to bring together two broad thoughts doesn't become a jumbled morass or thicket where just covering either by themselves would have been clearer... but sometimes i think it is useful & helpful to look at the "BIG PICTURE"... for those with the inclination & time to do so...
without looking at it, it seemed that some positions (QB) showed far less of a correlation between pedigree and playing at a high level (brady a 6th rounder, romo a UFA, etc)...
at other positions (LB), michael boley was the only non-day one (when this meant after three rounds) pick out of the 17 non-UFA top 20 LBs at the 2007 half-season split...
this led to thinking about possible implications of the data...
but first, lets make clear what this linkage study (at least in its inception) ISN'T trying to accomplish...
obviously the sample size noted above of just a season & a half is laughably small... i'll look back three years soon, and leave to other interested researchers & number crunchers to backtest some of these ideas, in case it might help for extrapolation and predictive purposes... i may try to go back further, & plan to track it going forward... the league seems to go through cyclical changes in trends, & it is unclear of what use it would be to study stats from decades ago when for instance the spread offense wasn't in as wide a use... & to what degree yielded defensive stats from a different era would be relevant to now (perhaps the idea of whether this last point is valid, & if so, what time segments & measures would be useful or non-useful for such a study, could almost be a whole separate topic & subject for a later thread)...
in the above study, i looked at top 20 primary defensive positions (DL/LB/DB)... my rationale was that in 10 team leagues, that would yield two starters at each position, and also once the scoring tables got into the 30s-40s, distinctions began to get increasingly smaller & imo negligible... i wanted to study difference makers...
on offense, i went with top 10 QBs, 20 RBs, 30 WRs & 10 TEs (again, with the rationale that this would yield the full contingent of starters in standard leagues)... but the point being, and this goes for offense & defense, the sample looked at for potential pedigree - starting/suceeding linkages (and possible differences by position) was the top 10-30... so this study makes no claims on that set comprised of ALL drafted players...
of what use it than...
since with each succeeding rookie draft that is the place usually begun from in evaluating who to prioritize in the draft... attempting to forecast who will stick, make the roster & have a long career & who won't... the above study doesn't address (that said, i'd like to see such a study conducted over time)...
perhaps a way this info could be used, is in dynasty applications with vets, when trying to grade & rank players by incorporating pedigree type information... IF there were a linkage such as alluded to above, it might be helpful for parsing and reducing top 30-40 lists by postion, possibly winnowing out lesser pedigreed prospects, as a sort of tie-breaker, in a reverse engineering sense... the caveat is that like with all incorporation of historical & statistical data into player evaluation, it should be combined & OVERLAYED on top of things like measurables & scouting info, film breakdown, etc, on a CASE-BY-CASE basis...
some other general thoughts about positions, that may have use for redraft & dynasty purposes, offense & defense...
i was trying to think of a more comprehensive but simpler overarching framework to help understand &, if it isn't too pretentious, attempt an explanation why some positions catch on faster than others (& maybe putting such a study on more rigorous statistical & historical footing)...
one thing i noticed recently is a kind of mirroring between OL/DL, RB/LB & WR/DB (QBs & TEs are sort of outliers to the framework, for reasons to be explained)...
the OL doesn't matter for offensive scoring purposes in MOST leagues (except indirectly in that attending closely to best & worst OLs can be helpful in breaking down the prospects of skill position players by team), though DL does in IDP variants... the line of course are responsible for either blocking or escaping blocks on each other, attacking the QB & RB or thwarting those attempts... so there is another kind of connection & linkage... one thing respective OL & DL seem to have in common is that it generally takes a few years to hit their stride... this can be explained by the fact that lineman, in addition to matters of technique & scheme comprehension (where it is a given that the vast majority of incoming rookies are undoubtedly raw), they need to undergo an almost metamorphosis of physical development & maturation, before they can even hold their own, let alone star... hopefully this part isn't too controversial...
one point of possible convergence with RB/LB was that there seemed to be a lot of examples where young players did extremely well (for instance last season, look no further than adrian peterson & patrick willis)... the usual explanation for RBs, and it seems reasonable to me, is that RBs are a position where attributes such as pure, natural talent and instincts appear to weigh more heavily... one other obvious, but maybe neglected point, imo, is that grabbing a ball that is handed to you by a QB is inherently easier to initiate a play where they are off to the races than the case of a pass play, which needs a highly synchronized sequence of events to take place... while flipping a pass to the RB in the flat may not seem like a big deal, throwing an intermediate pass across the middle which involves a QB passing the ball over some levels of the defense and in front of others, and squeezing it between several defenders, to a WR who needs to escape the jam, run a precise route, concentrate in traffic & often run according to precise timing, in some cases with the QB releasing the ball before the WR makes his break... decidedly IS a big deal... no wonder it usually takes QBs & WRs a few years to make an impact...
LBs may mirror RBs on the defensive side, in that they seem to favor talent & instinct, & they can produce at a high level relatively faster... one tie-in i wanted to make between the results & data from the study above at LB position, with this relative positional difference model... while the time study is small so far, after it is expanded, it will be interesting to see if there is a linkage between pedigree and starting/success... IF so, it would be pretty much a tautology to point out that this suggests the scouts seem to have a good handle on what scouting attributes tend to lead to success... it would be a different point to suggest that IF this is the case, we should respect pedigree information (at least for those positions in which such a breakdown is useful)...
WRs have already been covered to an extent... DBs may fall somewhere between DL & LB in terms of their rate of development arc & trajectory... i would expect them to take a little longer to catch on than LBs, but perhaps quicker than DL... while many can compete physically in less than the 2-3 years it can take many lineman to compete at a high level, they have some disadvantages relative to LBs... coverage schemes & responsibilities for young CBs & safeties can be very complex at the pro level, & may take a year or two to digest... one disadvantage DBs may have relative to WRs (when compared to RBs vs. LBs) is that if a WR isn't jammed, they can build up a lot of speed downfield, & can better leverage the fact that they know where they are going & DBs don't... with case of LBs, their straightest path to the RB may often be shorter than that of a safety trying to get across the field to cover a WR... and the RBs have to fight through more traffic & are impeded more than WR (where they can't be downfield, BY RULE
), making the LBs job easier than that of DB...QBs & TEs were called outliers to this model & i promised an explanation...
QBs are different on one level because they are the nerve center of the offense (courtesy of george plimpton in paper lion), and have to know what EVERYBODY is doing, in a way that, for instance, RBs & LBs don't have to... another point about possible reasons why there might seem to be a possible greater linkage & "dependence" on pedigree for success with some positions more than others... it was suggested that scouts might have a better handle on predicting how elite LBs would fare in the transition from college to pro that is so difficult for QB... in the case of QBs, the ability to be calm under pressure, process information quickly, the ability to see the field, the timing to squeeze the ball through tighter windows and more quickly, the physical courage & toughness to withstand a beating... these are very difficult to reduce to a test that would parallel the 40, vertical & broad jump, cone & shuttle drills, benchpress, which are tests yielding physical measurables of speed, quickness, strength, explosiveness, etc... many/most of the cited QB traits are mental & intangibles...
TE may be different than other positions in that they are sort of a hybrid OL/WR position, that has to be physical enough to block in line, but athletic enough to split the safeties on the deep seam route... this is a generalization, as there are receiving specialist TEs like scheffler that aren't very good blockers, as well as blocking specialists... it is important to make a distinction between poor blocking TEs that get to play & get scoring opps, & those for which blocking failures are an obstacle to playing time... this must differ by team, coaching philosophy, scheme, etc...
i will try & relay more of what i found in previous study by positions below, & plan to write separate articles on offense & defense this summer following up here...
* age is also critically important in dynasty (& to a lesser degree in redraft)... we all know by osmosis RBs are near done at 30... but it would be helpful to have more rigorous data on potential differences in ALL the positions (even breaking out DT/DE, MLB/ILB & OLB, S & CB on defense)... it didn't bear on this immediate subject, and is a separate area, but it will be included in my study to follow, on offense as well...
i had already done the pedigree study, but recently some thoughts coalesced around the offense/defense position parallel model... hopefully the attempt to bring together two broad thoughts doesn't become a jumbled morass or thicket where just covering either by themselves would have been clearer... but sometimes i think it is useful & helpful to look at the "BIG PICTURE"... for those with the inclination & time to do so...
Last edited by a moderator:
