ZWK
Footballguy
If you could only look at statistic to summarize a RB prospect's college performance in order to predict his NFL success, what would it be?
I would want to see his yards after contact (capped). Being hard to tackle - able to beat defenders and pick up extra yards after a defender has a chance to bring him down - is one of the most important abilities for a RB to have, and I suspect that it carries over pretty well from college to the NFL (at least for RBs who have the size & athleticism to play in the NFL). Many college RB stats (like yards per carry) are heavily influenced by contextual factors like scheme, blocking, and the quality of the defense, but Capped YAC seems likely to depend more on the RB and less on the context. Give me that (along with a RB's size and workout numbers) and I think I could make pretty good predictions of who will succeed in the NFL.
Unfortunately, yards after contact are not tracked as a standard statistic.
But this offseason, while I've been watching RB game videos, I have been tracking yards after contact and missed tackles. I have used my video watching plus counting, plus a little bit of number crunching, to make a Hard To Tackle Rating estimating RB prospects' elusiveness.
This post contains my method, my elusiveness rankings for 20 RB prospects, and a comparison with Greg Peshek's elusiveness numbers. If you just want to see my numbers, they're in this spreadsheet.
My Method
I've been watching cutup videos (showing every carry & reception for a full game) with a pen in hand, writing down three things for every play:
Missed tackles. If a defender has a chance to tackle the RB and the RB gets by him (in a way that lets him keep going) that counts as a missed tackle. I count missed tackles each play (using my judgment and attempting to apply a consistent standard on borderline plays), and I use missed tackles per touch as one of my main elusiveness metrics (average missed tackle rate is about 30%).
Yards after tackling opportunity. I note where the RB is when a defender first has a chance at him, and where the ball ends up being spotted at the end of the play - the difference is the RB's yards after tackling opportunity for that play. I call this "Yards After Contact" since that's a simpler & more standard name, even though in some cases the RB gets by the defender untouched. My second main elusiveness stat is capped yards after contact (with a minimum of zero yards and a maximum of ten yards).
Space vs. traffic. A play gets marked down as happening in "space" if the RB's first encounter with a defender who had a chance at him happened in space, meaning that there was just one defender & the RB had some room to maneuver. Otherwise (if there were multiple defenders with a chance to make the play, or there was limited room to maneuver because of other bodies in the area) I mark it down as being in traffic. About 57% of the plays I've tracked were in traffic, 43% in space. I calculate missed tackle rate and capped yards after contact separately for traffic plays and for space plays, since it is much easier for a RB to break tackles & pick up YAC in space than in traffic.
Skipped plays. If the RB doesn't have a real chance to pick up YAC, and the play has a below average result in terms of YAC / missed tackles, then I leave that play out of my calculations. Unless they involve 3+ yards after contact, I typically skip touchdowns (not the RB's fault if he finds the endzone before he has a chance to break a tackle), plays that begin inside the opponent's 5 yard line (which are almost guaranteed to look bad in terms of YAC), and receptions where the RB is going to the ground as he catches it (it's not a lack of elusiveness that prevents YAC after a diving catch).
My Stats
Missed Tackle Rate
Missed tackles divided by total touches
Capped YAC, aka YATOCAT (Yards After Tackling Opportunity Capped At Ten)
Average yards after contact, with a floor for each play of 0 yards and a ceiling for each play of 10 yards
3+ Extra Rate
On plays without a missed tackle, percent of plays where the RB had 3+ yards after contact
Space Missed Tackle Rate
Missed tackles on plays in space divided by number of plays in space
Space Capped YAC
Average yards after contact on plays in space, with a floor for each play of 0 yards and a ceiling for each play of 15 yards
Traffic Missed Tackle Rate
Missed tackles on plays in traffic divided by number of plays in traffic
Traffic Capped YAC
Average yards after contact on plays in traffic, with a floor for each play of 0 yards and a ceiling for each play of 8 yards
Traffic 3+ Extra Rate
Percent of plays in traffic where the RB had 3+ yards after contact (included plays with a missed tackle)
Hard To Tackle Rating
A weighted sum of the other eight stats (with a small strength of schedule adjustment), which has an average a bit over 6. It is roughly 50% yards after contact, 30% missed tackles, and 20% 3+ Rate (and also about 1/3 traffic, 1/3 space, 1/3 all plays mixed together).
RB Rankings
There are 20 RBs who I have charted for at least 3 games & 50 touches. Their detailed stats are in this spreadsheet, and their overall Hard To Tackle Rating is below:
8.45 Dri Archer
7.53 Tim Flanders
7.33 Carlos Hyde
7.20 Jeremy Hill
7.13 Devonta Freeman
6.98 James Wilder
6.90 Lache Seastrunk
6.84 Antonio Andrews
6.68 Charles Sims
6.65 Isaiah Crowell
6.49 Ka'Deem Carey
6.35 Jerick McKinnon
6.29 Andre Williams
6.22 Tre Mason
6.06 Bishop Sankey
6.00 Silas Redd
5.57 James White
5.36 Marion Grice
5.24 Damien Williams
4.61 Terrance West
Comparison with Greg Peshek's Metrics
Greg Peshek also tracks elusiveness metrics for college RBs. There are a few differences in our methods, but they produce similar results. For the 9 RBs who are in my list of 20 and in his list of 10 RBs charted, the correlation between my Hard To Tackle Rating and his Extra Yardage score is r=0.52. For the 12 RBs who we have both done some charting for (those 9 plus Storm Johnson, Antonio Andrews, and Marion Grice), the correlation between my Hard To Tackle Rating and an equally weighted standardized sum of Peshek's Btk% and YacCap is r=0.68. Some of the most notable differences in our methods:
I would want to see his yards after contact (capped). Being hard to tackle - able to beat defenders and pick up extra yards after a defender has a chance to bring him down - is one of the most important abilities for a RB to have, and I suspect that it carries over pretty well from college to the NFL (at least for RBs who have the size & athleticism to play in the NFL). Many college RB stats (like yards per carry) are heavily influenced by contextual factors like scheme, blocking, and the quality of the defense, but Capped YAC seems likely to depend more on the RB and less on the context. Give me that (along with a RB's size and workout numbers) and I think I could make pretty good predictions of who will succeed in the NFL.
Unfortunately, yards after contact are not tracked as a standard statistic.
But this offseason, while I've been watching RB game videos, I have been tracking yards after contact and missed tackles. I have used my video watching plus counting, plus a little bit of number crunching, to make a Hard To Tackle Rating estimating RB prospects' elusiveness.
This post contains my method, my elusiveness rankings for 20 RB prospects, and a comparison with Greg Peshek's elusiveness numbers. If you just want to see my numbers, they're in this spreadsheet.
My Method
I've been watching cutup videos (showing every carry & reception for a full game) with a pen in hand, writing down three things for every play:
Missed tackles. If a defender has a chance to tackle the RB and the RB gets by him (in a way that lets him keep going) that counts as a missed tackle. I count missed tackles each play (using my judgment and attempting to apply a consistent standard on borderline plays), and I use missed tackles per touch as one of my main elusiveness metrics (average missed tackle rate is about 30%).
Yards after tackling opportunity. I note where the RB is when a defender first has a chance at him, and where the ball ends up being spotted at the end of the play - the difference is the RB's yards after tackling opportunity for that play. I call this "Yards After Contact" since that's a simpler & more standard name, even though in some cases the RB gets by the defender untouched. My second main elusiveness stat is capped yards after contact (with a minimum of zero yards and a maximum of ten yards).
Space vs. traffic. A play gets marked down as happening in "space" if the RB's first encounter with a defender who had a chance at him happened in space, meaning that there was just one defender & the RB had some room to maneuver. Otherwise (if there were multiple defenders with a chance to make the play, or there was limited room to maneuver because of other bodies in the area) I mark it down as being in traffic. About 57% of the plays I've tracked were in traffic, 43% in space. I calculate missed tackle rate and capped yards after contact separately for traffic plays and for space plays, since it is much easier for a RB to break tackles & pick up YAC in space than in traffic.
Skipped plays. If the RB doesn't have a real chance to pick up YAC, and the play has a below average result in terms of YAC / missed tackles, then I leave that play out of my calculations. Unless they involve 3+ yards after contact, I typically skip touchdowns (not the RB's fault if he finds the endzone before he has a chance to break a tackle), plays that begin inside the opponent's 5 yard line (which are almost guaranteed to look bad in terms of YAC), and receptions where the RB is going to the ground as he catches it (it's not a lack of elusiveness that prevents YAC after a diving catch).
My Stats
Missed Tackle Rate
Missed tackles divided by total touches
Capped YAC, aka YATOCAT (Yards After Tackling Opportunity Capped At Ten)
Average yards after contact, with a floor for each play of 0 yards and a ceiling for each play of 10 yards
3+ Extra Rate
On plays without a missed tackle, percent of plays where the RB had 3+ yards after contact
Space Missed Tackle Rate
Missed tackles on plays in space divided by number of plays in space
Space Capped YAC
Average yards after contact on plays in space, with a floor for each play of 0 yards and a ceiling for each play of 15 yards
Traffic Missed Tackle Rate
Missed tackles on plays in traffic divided by number of plays in traffic
Traffic Capped YAC
Average yards after contact on plays in traffic, with a floor for each play of 0 yards and a ceiling for each play of 8 yards
Traffic 3+ Extra Rate
Percent of plays in traffic where the RB had 3+ yards after contact (included plays with a missed tackle)
Hard To Tackle Rating
A weighted sum of the other eight stats (with a small strength of schedule adjustment), which has an average a bit over 6. It is roughly 50% yards after contact, 30% missed tackles, and 20% 3+ Rate (and also about 1/3 traffic, 1/3 space, 1/3 all plays mixed together).
RB Rankings
There are 20 RBs who I have charted for at least 3 games & 50 touches. Their detailed stats are in this spreadsheet, and their overall Hard To Tackle Rating is below:
8.45 Dri Archer
7.53 Tim Flanders
7.33 Carlos Hyde
7.20 Jeremy Hill
7.13 Devonta Freeman
6.98 James Wilder
6.90 Lache Seastrunk
6.84 Antonio Andrews
6.68 Charles Sims
6.65 Isaiah Crowell
6.49 Ka'Deem Carey
6.35 Jerick McKinnon
6.29 Andre Williams
6.22 Tre Mason
6.06 Bishop Sankey
6.00 Silas Redd
5.57 James White
5.36 Marion Grice
5.24 Damien Williams
4.61 Terrance West
Comparison with Greg Peshek's Metrics
Greg Peshek also tracks elusiveness metrics for college RBs. There are a few differences in our methods, but they produce similar results. For the 9 RBs who are in my list of 20 and in his list of 10 RBs charted, the correlation between my Hard To Tackle Rating and his Extra Yardage score is r=0.52. For the 12 RBs who we have both done some charting for (those 9 plus Storm Johnson, Antonio Andrews, and Marion Grice), the correlation between my Hard To Tackle Rating and an equally weighted standardized sum of Peshek's Btk% and YacCap is r=0.68. Some of the most notable differences in our methods:
- We have different definitions of missed tackles & yards after contact. The biggest explicit difference here is in how we define a missed tackle - he has a much narrower definition and only counts about 1/4 as many broken tackles as I do. And on all of the plays that require judgment calls, he is using his judgment and I am using my judgment.
- Peshek charts every play of every game in 2013 (for his 10 RBs); I only chart a few games for each RB and I include some games from previous years if that's what's available on YouTube.
- I distinguish between plays in space & plays in traffic, calculate elusiveness stats separately for the two types of plays, and then combine the numbers. Peshek just averages together all plays.
- We use different formula for calculating the overall elusiveness metric. I use a weighted sum of 8 stats, adjusted for strength of schedule. I'm not sure how he calculates his Extra Yardage score, but it's some kind of combination of Capped Yards After Contact and Broken Tackle %.