For obvious reasons, there aren’t many inroads into the Sabermetric side of college football, which is more naturally inhabited by, you know, drankin’ and vitriol, etc. The limited sample size, vast differences in schedules and wildly varying numbers based on opponent strength don’t help, either. But Missouri fan Bill Connelly, aka "The Boy" at Rock M Nation, has spent the offseason adding some of his own ideas to the increasingly relevant formats developed by Football Outsiders in an effort to build a similar system for the amateurs. You can start with parts one, two and three.
Bill’s most immediately applicable contribution, though, might come from the latest edition, posted in the "FanPosts" section to the right, especially his look at success of passing downs vs. non-passing downs, and what those results mean for offenses strategically.
To get there, you have to get through some of the wonkier elements of the earlier posts, which are focused on finding value in individual plays, based on how likely those plays are to contribute points. One of the ways he did this was to develop a "points per play" average, based on how likely a team was to score from any particular point on the field (i.e., what was the average number of points scored on drives when the ball was at each yard-line). Like so:
Anyone can tell you the chances of scoring increase as the offense moves closer to the defense’s goalline, but this shows that yardage becomes more valuable along that track. That is, all 20-yard gains are not created equal: a 20-yard gain from your own 1-yard line to your own 21-yard line is worth almost no value on the scoreboard –– less than half a point, on average. But a 20-yard yard gain from your own 41-yard line to your opponents’ 39-yard line is worth about a full point. And it only increases from there: it’s worth a full point, on average, to move the ball from your opponents’ four a few feet to the one. Yards gained on the offense’s side of the 50 count as much as any others in the box score, but they’re much less likely to lead to points on the scoreboard.
These averages vary, though, depending on down, which Bill also charted. If you add up all the gains in any given game according to the values in the "free-floating" chart above, then add them up according to the down-dependent chart, then average those two numbers together and subtract turnover costliness (calculated here at –3.17 points per turnover), you get something pretty close to the actual point totals of a game –– by his count, Bill’s numbers reflected the Missouri-Kansas score last year exactly.
Another part of the measure is a standard "success rate," based on down and distance:
Success Rate is the standard for a lot of the stats I use. It’s a slightly-altered version of a Football Outsiders measure. On any given play, there’s a roughly 44% chance of ‘success.’ Here are the rules according to down:
1st Down: 50% of necessary yardage. If it's 1st-and-10, you need 5 yards for 'success.' Football Outsiders use 40 percent for 1st down, but with the games I've entered, that led to a 1st down success rate of about 51%. Bumping the requirements to 50% led to the 44% rate for which I was aiming.
2nd Down: 65% of necessary yardage (rounded up to the nearest yard, of course). If it's 2nd-and-10, you need 7 yards for 'success'. 2nd-and-15?10 yards. This makes sense, really, because to succeed regularly on 3rd downs, you need to stay at 3rd-and-5 or less. Getting most of the way there on 2nd downs sets you up infinitely better for 3rd down. Football Outsiders uses 70%, but the success rate for that was around 42%. Weakening the requirements slightly got me into the range I was looking for.
3rd and 4th Downs: 100% of necessary yardage. I figure this requires no explanation.
- - -
But Bill wanted one all-encompassing number, similar to the widely-used OPS in baseball, that could quantify a player’s or team’s "average" play in a way that correlated strongly with scoring points. So he added "success rate" to "points per play" –– again, following the OPS method in baseball, which adds on-base average and slugging average (bases by batter per at bat) –– and emerged with a number he calls "S&P." Read more about this in his glossary.
Okay. All of that is a necessary prologue to the best part of Bill’s latest entry to the series, which breaks down the success rate, points per play and S&P numbers in a myriad different ways –– in blowouts, in close games, by down, by quarter, by half, etc. Most interesting by far is his statistical assessment of success rates on "passing downs" vs. "non-passing downs," when a passing down is defined thusly:
• 2nd-and-8 or more
• 3rd-and-5 or more
• 4th-and-5 or more
In those situations, Bill’s methods provide some pretty stark results:
Plays in favorable down-and-distance were almost twice as successful as plays run in passing downs. Later on, Bill finds the same ratio for sacks –– passing downs were almost twice as likely to result in a sack than "on-schedule" downs.
In real-world terms:
Check the down and distance on those plays, and also here, here, and here, for starters.
I welcome this, personally, as an empirical base that bolsters my usual emphasis on keeping the entire playbook open: outside of talent, predictability is the number one killer of offenses, and defenses that stop the run and make offenses one-dimensional are, well, see above. Usually, I invoke this intuitively (or with YouTube videos). Now, there’s research and stuff that puts the difference in high-contrast black and white. Kudos, Bill.
Loading comments...