Monday, January 19, 2015

College National Championship Postgame Analysis

All hail the college football playoff.  Instead of being subjected to an Alabama vs. Florida State BCS "championship" game (and don't even try to tell me that any other teams could have gotten into that game; Florida State for all its shortcomings was the undefeated defending champion, and Alabama is Alabama - you know, from the SEC) we got let those two go play an imaginary bronze-medal game.  Meanwhile the real championship was decided on the field (unless you like TCU).  I'd still prefer a ten-team tournament of conference champions, but wow, was this better than before.

That the underdog Buckeyes came away with the win wasn't a huge surprise.  In the last four years, Ohio State's recruiting classes have ranked #11, #4, #2, and #3 according to rivals.com (Oregon's were #9, #16, #22, and #26).  Oregon was also missing key contributors at tight end, offensive line, cornerback and receiver, though OSU was also down to third quarterback Cardale Jones (more on him later).  The surprise was the dominating fashion with which OSU dispatched a team that is used to dominating its own opponents.  The Buckeyes averaged 6.4 yards per play, possessed the ball for over 37 minutes (a statistic with little meaning against Oregon, but still), and won by 22 despite losing the turnover battle 4-0 (ignoring a meaningless Oregon interception on the last play).  Tailback Ezekiel Elliott rushed for a 6.8 yard average on 36 attempts while using only half a jersey.

Meanwhile, the vaunted Oregon offense was held largely in check.  Although they piled up 465 yards of offense and didn't turn the ball over, they were just 2-14 on third and fourth down, and scored just 20 points on five trips inside the red zone.  Surprisingly, a Ducks offense that is known for innovation seemed strangely low on ideas.  No Oregon non-quarterback attempted a pass, no designed laterals were thrown, no double reverses occurred.  Byron Marshall, a potential matchup advantage as a former tailback playing receiver, had only one rush.  Normally, Oregon's base offense and blistering pace are enough to keep most defenses off balance.  But Ohio State coach Urban Meyer is a spread guru himself and a close friend of the Oregon program, so the Ducks could have anticipated less confusion than normal.  Ohio State, being out of academic session during the 11 day layoff after the Sugar Bowl, was also allowed unlimited practice time instead of the normal 20 hour weekly limit.  Under these circumstances, a few trick plays would seem to be a useful asset.  OSU certainly knows about having some tricks ready; in their Sugar Bowl win over Alabama they used a reverse pass to jump start their comeback.  Even the great Nebraska and Florida State teams of the 1990s, who could beat anyone when everything went as expected, always had a trick play ready just in case it didn't.

More surprising than Oregon's offensive rigidity, though, was their stubborn insistence on playing their base defense.  Despite consistently yielding chunks of yards to the Buckeye ground game, Oregon's defenders remained in their base 3-4 Cover 4 alignment.  With their frontal defenders undersized and outnumbered, the Ducks were unable to generate either a consistent pass rush (their one "sack" of the game was actually a Winstonian scramble-fumble by Jones) or any serial success against the run.  

To be fair, there is wisdom in dancing with the girl who brought you to the prom.  The Ducks defensive philosophy all season has been to bend without breaking, limiting big plays by keeping the play in front of the defense, playing zone to keep the defenders facing the play, forcing the opposition to execute long drives, and stiffening in the red zone.  Although the opposition may gain yards against this strategy, over the course of a long drive a penalty, string of incompletions, or turnover would often get the Ducks D off the field.  Oregon's tall but light defensive ends are liabilities against the run but ideal for rushing the passer when the Ducks have the lead, which they usually do.  The problem was that Ohio State's running game was so successful, they were able to score quickly without much passing.  Despite the Buckeye playbook narrowing at times to just inside zone, sweep, and counter, the Ducks never adjusted and never stopped allowing chunks of yards on the ground (though they were in position to pick up the four turnovers that kept them in the game).  So, with the obviously unfair advantage of hindsight, what should the Ducks have done instead?

Bear Down

The team that gave Ohio State it's only loss on the season offers a starting place for ideas.  Virginia Tech, who would go on to finish 7-6 and fifth in the ACC Coastal Division, beat the Buckeyes 35-21 in Columbus on September 6th.  The Hokies slowed down the Buckeye ground game with a "Bear" front, a five-man defensive line that puts a nose guard and two tackles over the offensive center and both guards.

(Image from elevenwarriors.com)
The Bear front is named for the Chicago Bears, who used a defense called the 46 Bear under Buddy Ryan in the 1980s.  (The 46 is not the number of linemen and linebackers, it was just the jersey number of Doug Plank, Chicago's strong safety at the time and a key cog in the defense.)  The 46 is a truly innovative scheme that creates several problems for the offense, with the three interior linemen only its most obvious structural element.  Virginia Tech doesn't use the true 46 Defense here, but just covering the three inside linemen causes two big problems for the Buckeye-style run game:

1) The inside zone (and outside zone) play relies on double teams wherever the defense leaves a "bubble," or uncovered lineman.  The offensive linemen want to double their nearest defender, and then let one of the doubling linemen scrape off to a linebacker, creasing a crease for the runner to find and exploit.  But against a five-man defensive line, no lineman can help double team because each has a defender of his own to deal with.  As Homer Smith wrote, "if there are no linebackers, there is no zone play."


Inside zone versus 3-4 front.  For simplicity, this diagram omits extra blockers such as tight ends and h-backs, which Ohio State often uses.  These extra blockers may create favorable matchups but do not give the offense any extra numerical advantage, because they would each bring a corresponding defender into the play.  In fairness to Oregon, its Cover 4 scheme expects the two safeties (not pictured) to help fill against the run.  Still, the bubbles in the front are obvious.

Against the Bear front, the offensive and defensive lines have the same number of players, so the zone play has no bubble to exploit.  The result is that there is no natural crease for the runner.

2) The off-tackle counter and power plays rely on a double team at the point of attack, combined with a kickout block from a backside pulling lineman.  Against a bear front, the center can't block both the nose guard in front of him and the defensive tackle over the pulling guard, so he either has to call for help on the nose, creating a cascade effect that eliminates the double team, or the guard has to stay home.

The counter play uses a block down, kick out power scheme.  The running lane is created by the double team at the point of attack and the kickout block of the end man on the line.

The backside defensive tackle is a threat to disrupt the power play by shooting the gap and either dropping the runner for a loss or interfering with the pulling guard.  The center has to block back to protect this guard, but also has a nose guard to deal with himself.  Against a Bear the center can make a "TNT" call as shown to down-block the whole front and maintain the integrity of the play.  But this adjustment removes the double team at the point of attack and leaves the Mike free to run to the ball.

Of course, adding a player to the forcing unit means removing one somewhere else.  The Hokies took that player from the contain, playing Cover 0 or Cover 1 and daring Ohio State to beat them over the top.  The Buckeyes of September 6th couldn't.  Quarterback J.T. Barrett did not consistently beat the man coverage with throws downfield, in part because his offensive line could not give him time against the five-man rush.  He created a few plays with his feet, but not consistently enough to bring the Buckeyes back.  But two things changed between September 6th and January 12th: Barrett broke his ankle and was replaced by the big-armed Jones, and Devin Smith emerged as one of the more dangerous deep threats in the game.  Ohio State's offensive line also improved at pass protection over the course of the season.  By the time the championship game came around, it had become unlikely that Oregon, Virginia Tech, or anyone else could have stopped the Buckeyes with the same gameplan that worked early in the season.

Playing with Fire

If Cover 0 and Cover 1 had become too risky against the suddenly bombs-away Bucks, but sitting back in a 2-shell was overly conservative, then a 3-3 fire zone offers a middle ground.  At first glance, this coverage looks hopelessly unsound.  Not only do only three defenders have to handle the entire undercoverage (also true of Cover 4, btw) but because the undercoverage and contain both consist of odd numbers, the weakpoints in the two levels align, creating inviting lanes for receivers to run right through.  


Technique makes the coverage work.  The slot defenders do not spot-drop directly to the flat on the snap (although a few DCs, such as Manny Diaz, do teach spot-dropping in fire zone coverage), leaving their receiver free to run the seam.  If the inside receiver runs straight up the seam, the slot defender collisions him and "carries" the seam route until the depth where the free safety can cover it.  He then drops diagonally outside to cover the curl route before going to the flat (there are different ways to teach this coverage, and also exceptions to these rules for things like crossing receivers).

More fundamentally, the 3-3 coverage is a percentage play by the defense.  It allows five frontal defenders to stop the run and pressure the passer, while keeping a three-deep shell to prevent big plays over the top.  Unlike the Cover 0 and Cover 1 coverages used by the Hokies, it also keeps defenders in zone coverage where they can face the play.  This puts them in position to make a quick tackle if the offense completes a short pass, or to make the interception in the event of a tipped ball or overthrow.  You can't get something for nothing though, and this defense does allow large windows for the short passing game.

But Cardale Jones, for all his success as a short-yardage runner and long-yardage thrower, had not yet proven his ability to make consistently good decisions in a ball-control short passing game.  Furthermore, Oregon's odd front defense should adjust naturally to a Bear.

Just one of several ways to create a Bear front from a base 3-4.

All of which makes it so surprising the Ducks never made this adjustment, or any clear defensive adjustment, especially in the second half as the diminishing clock steadily reduced the downside of taking a risk.  With the benefit of hindsight, there are probably many things that Oregon could have done differently in this game.  Firing up a zone blitz coverage is just one.

Addendum: Barrett and Jones

With Cardale Jones announcing that he will return to school next year instead of cashing in his instant stardom by leaving for the NFL, Ohio State has an embarrassment of quarterback riches on its hands.  Braxton Miller, the two-time Big 10 offensive player of the year who missed the championship season with a shoulder injury, may decide to change schools rather than positions.  Since he is eligible to play immediately as a graduate transfer, he will be a highly desired free agent, possibly even by Oregon.  All J.T. Barrett did after Miller went down was lead the Buckeyes to an undefeated regular season and Big 10 championship.  Then there's Cardale Jones, who came on to win the Big 10 title game and both playoff games to bring home the national title.

If Miller came back to OSU as a halfback, the potential two-quarterback formations could drive opposing defenses... nuts.  But he probably wants to play quarterback and may not be willing to risk spending his final year as a backup, so he seems likely to transfer.  That leaves Barrett and Jones, which offer an interesting contrast in styles.  Barrett is the faster runner and more experienced player, but Jones offers much greater size and the bigger arm.  Although Meyer might normally prefer to put as much speed on the field as possible, the fullback-sized Jones, like former Meyer QB Tim Tebow, can bulldoze in short yardage situations.  This allows a hybrid runner/receiver like Percy Harvin or Chris Rainey to fill the halfback spot.  Such "four and a half receiver" formations put extreme stress on the defense and allow Meyer's offense to reach new heights.  Combined with Jones' arm, this would make him my choice.

Sunday, January 11, 2015

Patriots Bring the A-11 to the NFL

On Saturday, the Patriots were in trouble against the Ravens.  Trailing in the third quarter and struggling to move the ball against the Baltimore defense, New England coach Bill Belichick reached into his bag of tricks.  He ran several plays with only four offensive linemen, confusing the defense into leaving a receiver wide open:


(image stolen from SB Nation)

Before this play running back Shane Vereen (circled) declared himself as an ineligible receiver to the referee, who told the defensive captain.  As an ineligible receiver, Vereen can't go downfield, so he runs backward instead.  Tight end Michael Hoomanawanui, aligned where the offensive left tackle would normally be, sprints down the seam uncovered by the surprised defense.  The result is a fifteen yard pass from Tom Brady and a first down.

Ravens coach John Harbaugh was none too pleased.  Calling the play and the preceding substitution "clearly deceptive," he insinuated that the referees did not understand what had happened, and that the NFL may change some rules in the offseason to prevent this kind of thing happening in the future.  He even took an intentional penalty to prevent the Patriots from doing it on one play.

There is nothing illegal about what the Patriots did.  Only those players who align in the backfield or as one of the two end men on the line of scrimmage are eligible to catch a pass; the five "covered linemen" are not.  (On this play, Vereen is "covered" by Julian Edelman at the bottom of the screen, so he is aligned as an ineligible receiver, while Hoomanawanui is the end man on the left of the formation, so he is eligible.)  In the NFL, the offense normally has five players with ineligible numbers (50-79) on the field, to help the referees identify who can and can't catch a pass.  If they align with more or fewer than five such numbers, they must tell the referee who is or isn't eligible.

At first glance, this looks similar to plays you see from time to time where the offense uses an unbalanced formation to put a tackle out wide and hide a tight end at the tackle spot (LSU, Alabama, and Auburn have all done this recently).  But there's a critical difference.  On those plays, the offense still has the usual five ineligible numbers on the field, albeit in unfamiliar places.  On this play, there were seven players on the Patriots offense (counting Brady) with jersey numbers indicating they could catch a forward pass.  Vereen declared himself ineligible, just like when an extra offensive lineman comes on the field and declares himself eligible.  But when one lineman declares himself eligible, the defense can find him easily.  It's somewhat harder when a player reports as ineligible, leaving just a few seconds to determine by elimination the remaining five players who are eligible.  This is what got Harbaugh so worked up.

Belichick explained the formation this way after the game: "We ran it three times, a couple different looks. We had six eligible receivers on the field, but only five were eligible. The one who was ineligible reported that he was ineligible. No different than on the punt team or a situation like that."  Actually, when you put it that way, that sounds kinda familiar...


All Eleven


The A-11 offense was a short-lived high school offense based around a similar premise.  This scheme (the name is short for "all eleven eligible") used a spread formation with three groups (called "pods") of three linemen each, with two backs in the backfield side by side, all of them wearing the jersey numbers of eligible receivers.

(image from Wikipedia)

Before the snap, all players except the center would stand in the backfield.  Just before the snap, six of the players would step onto the line of scrimmage, "covering" five of them and making them ineligible receivers.  After the required one second, the ball would be snapped and the five eligible receivers would be free to go downfield, with the other five staying near the line.  Without enough time to adjust to the shift, the defense would often leave receivers running uncovered, just like Hoomanawanui.
(image from fishduck.com)
Normally, it wouldn't be legal to have eleven players on the field in eligible receiver numbers in a high school game.  But the A-11 creators took advantage of a loophole, the "scrimmage kick exemption."  Because the quarterbacks in the backfield were at least seven yards off the line of scrimmage, the A-11 was technically a kicking formation.  And because high school teams often use defensive players on kick coverage units, it makes sense that there was a rule relaxing the normal regulations around eligible receiver numbers on kicking downs.  But after two years of the A-11 turning the game upside down, the traditionalists had seen enough, and the National Federation of High School Associations voted in 2009 to close the loophole, restricting the scrimmage kick exemption to fourth downs.  (In the NCAA the exemption exists only on obvious kicking situations, presumably fourth downs or late in the half.)  The A-11 still exists as a (relatively) conventional unbalanced line formation, but the dream of "all eleven" football is dead, at least until fourth down.

But now, it looks like the Patriots may have resurrected it with their "punt team" offense.  The NFL is the only league that allows players to declare themselves eligible or ineligible.  Theoretically, there is nothing stopping the Patriots from going into their game with the Colts next weekend with eligible numbers on all of their players.  Before the snap, five of the players could inform the referee that they are ineligible, then scramble to the line and snap the ball.  

The creators of the A-11 hoped to change the game, but they couldn't overcome the powerful interests who wanted to keep it as it was.  Time will tell if the Patriots are powerful enough to succeed where they failed.

Monday, December 22, 2014

A Basic Rocket Offense for Youth Football

The youth team I most recently helped coach runs the wishbone.  I like the wishbone.  It has lots of lines of force, nice symmetry, and easily goes unbalanced or breaks into other formations.  But it's hard for kids under the age of about 11 or 12 to master its most powerful weapon, the triple option.  With the 7-8 and 9-10 year old teams, we didn't run the option at all - the wishbone still gives plenty of ways to run the ball with power and misdirection, so we did that instead.  But running the wishbone without the option feels a little like using a wrench to pound a nail - it works but it feels like there might be a better way.

As an assistant coach, my job is to help the head coach implement his vision, so none of the following is intended as criticism.  I also fully understand that running the wishbone with the younger kids gets them familiar with the formation so they are ready to run the option from it when they get older.  But if the goal were just to have a team of young kids run the ball with power and misdirection, there are simpler ways to do it, like this:

The Rocket Sweep


Monday, November 24, 2014

The Four Triangles

Football theory essay question:

triangle read combines a vertical stretch to defeat Cover 2 with a horizontal stretch to defeat Cover 3, providing options regardless of the defensive coverage (many of these plays will also include crossing routes to defeat man coverage).  There are four possible ways to orient a right angle along the axes of a grid (|_, _|, |¯ , ¯|), so there are four possible ways to orient a triangle read.  But are all four of these triangles in use in football today?  Why or why not?

Triangle 1: Lower Outside Apex

This is the classic triangle read pass, used in plays like Mesh (Air Raid version and Coverdale & Robinson version), Stick, and Double Quick Outs.  The receiver in the flat does double duty: he is the low man in the high-low stretch on the cornerback that is a classic way to attack 2-Deep Zone, and he's also the outside man in the inside-out stretch on the outside linebacker or sky safety that is a staple of attacks on 3-Deep.

 Mesh, Coverdale and Robinson version

Monday, October 20, 2014

NCAA Champion's League Update

After seven chaotic weeks of football, the job of the College Football Playoff Selection Committee looks harder than ever.  Four undefeated FBS teams remain (Florida State, Mississippi, Mississippi State and Marshall), but only two of the "Power Five" conferences have one. Upsets have blemished many of the preseason favorites, while suspensions and questions surround others.  The selection process promises to be adventurous, which is great TV but not great for the Committee or the sport.  Fortunately, there is a better way. 

Sunday, September 21, 2014

The NCAA Football Champions League

College football has a championship problem.  Fans are unhappy with the subjectivity of the poll system, but they still love the bowl games.  Rather than simply choosing a champion as it does in other sports, the NCAA has allowed the conferences to enter into a four-team playoff to be contested this fall for the first time.  This can only lead to an expanded playoff in future years, as there will definitely be several teams with a legitimate claim to the fourth spot (if not to the top spot).  And the selection committee approach promises to be both controversial and entertaining, which probably wasn't what the association was going for.

Wednesday, September 10, 2014

More on Defending the Read Option

Chris Brown wrote a characteristically excellent little post about an additional wrinkle for the read option: using a downfield pass as the "pitch phase" in the triple option.

 

In theory, having a downfield pass as the third option instead of a backward pitch or a stalk-block doesn't add any additional effectiveness to the play, since the receiver is accounting for the cornerback one way or another.  And none of these adds any additional numbers to the playside in the way that the pitchback coming across the formation in the wishbone does, for example.  Numerically, this play is the same as a quarterback sweep, not that that's a bad thing.  Probabilistically, though, there's an extra dimension here. 

Saturday, August 30, 2014

The NFL Sign-and-Trade

The Texans' Andre Johnson, tired of his team's perpetual rebuilding, has requested a trade.  Houston can hardly afford to trade him though, because while getting rid of Johnson would save them having to pay his $15.6 million salary, it would also add almost $12 million in dead money to their salary cap (ignoring for now the post-June 1 split rule), leaving a net savings of only around $3.7 million (hat tip to the excellent website overthecap.com).  Replacing anything close to Johnson's production for only $3.7 million would be impossible, so Houston is forced to keep their unhappy player. 

The Johnson example shows how the rules governing NFL player contracts help make trades rare, especially for star players with large contracts.  When they do happen, trades tend to involve players near the end of their contracts who are at risk of being cut outright.  With some careful contract structuring, however, a trade could allow two teams to capitalize much more on their individual resources in a way that would benefit both teams, and the traded player.

Sunday, January 26, 2014

One From the Vault: Restart the Revolution by Joseph Mellor

The article below is one I found on the internet years ago, saved on my computer, lost, and recently found again.  It doesn't appear to be available online anymore, so as a service to the internet football community, I'm reprinting it here.  Joseph Mellor, whoever you are, if you don't want your work on my site just let me know and I'll take it down.


The main thesis of the essay is that the Run and Shoot Offense, particularly at the NFL level, was unfairly maligned by casual fans and lazy reporters, and remains a devastating scheme that should be given another chance.  Mellor backs this argument up with convincing enough logic and facts, but near the end of the article he answers his own question of why the Run and Shoot died out: it didn't.  Its principles have been incorporated, Borg-like, into every NFL offense today.  Like the timing routes of the West Coast Offense or the zone run game of Alex Gibbs and the Redskins, once the effectiveness of the Run and Shoot's self-adapting routes had been demonstrated, every team in the league copied them.  No NFL team runs the "pure" Run and Shoot offense of Mouse Davis anymore, and even June Jones, now head coach at SMU, has adapted his offense to base out of the shotgun and use less motion than he used as a player at Portland State or as the coach of the Falcons.  But remnants of the scheme certainly live on in almost any modern passing offense, and the "pure" offense still has a dedicated following as a High School offense.

Thursday, January 9, 2014

Reminder: NFL (and College) Overtime is Still Not "Fair"

I don't mind tie games. If two teams are evenly matched and play equally well, a tie is the logical way for a game to end (outside of elimination games, of course).  Ties also make for more interesting playoff races as more teams remain in playoff contention later into the season.  Most football fans, however, disagree with me, and see ties as unsatisfying cop-outs that should be quarantined in the world of soccer.  Overtime reliably serves to break a tie, but it has to be designed carefully or it may reward the wrong things.  In football today, no level of football can claim an overtime system that is totally "fair," but there are fair alternatives to the current systems.

Sunday, December 22, 2013

How to Run Out the Clock

Holding the ball, the lead, and a first down late in the game, your team wants to run out the clock and end the game. How much time can they safely run off without giving the ball back to the opposing team?

In the NFL, timing rules ensure that exactly 40 seconds elapse between plays, so a team with a first down just inside the two-minute warning can kneel three times to end the game without needing a fourth down snap, if the defense is out of timeouts. This assumes the offense uses the entire 40 second play clock between snaps (reasonable since a delay of game penalty is basically harmless), but it also assumes the kneel plays themselves take no time at all, so the two assumptions tend to average out.  If the defense does have a timeout or the first down snap occurs just before the two minute warning, then the challenge for the offense is a little tougher, and the time taken by the plays themselves becomes important.

Monday, November 11, 2013

Basic football theory: time, space, force, and... pawn structure?

Football and chess both involve the placement and movement of pieces with diverse capabilities in a limited playing space, so it's not surprising the basic principles governing the two games are similar.  In fact, while chess masters and pro football coaches may have many hundreds of favorite play sequences memorized, most people can get pretty far in both games just by keeping these basic concepts in mind.

Tuesday, October 15, 2013

Football risk management and the downfield pitch

In his book Coaching Run-And-Shoot Football, coach Al Black offers the shovel pass as one of the constraint plays the run-and-shoot offense uses to punish defenses playing for the pass, along with the draw and speed option. 

Run and shoot shovel pass away from trips

At the end of his description of the quarterback's responsibilities in the shovel play, coach Black throws in this tidbit: "After the pitch he should continue downfield to be in position to receive a pitchout from the receiver he tossed the ball to."

Very interesting.  In any option play, the quarterback occupies a defender who would otherwise have to be blocked or faked, reducing the minimum number of unoccupied defenders to one (the counterpart of the ballcarrier) from two (the counterparts of the ballcarrier and the QB). But if the ballcarrier (here, the slot receiver) can also occupy a defender with an option, then the number of free defenders is reduced to zero. Put another way, a shovel pass give-and-go creates a triple option (keep, pitch, or pitch and pitch back) with only two runners, meaning two defenders can be left unblocked, the nine remaining defenders can be blocked or otherwise occupied by nine offensive players, and nobody will be left to tackle the ball.  Touchdown, right?  And furthermore, revolutionary offensive concept, right?

Sunday, October 6, 2013

Endgame Fail Costs Tennessee vs. Georgia

In football as in chess, endgame tactics are distinct from the opening and the midgame.  In particular, there are times when it is better to possess the ball while tied (or even trailing) in the score than to be defending with the lead.  The Tennessee Volunteers' failure to manage this situation very likely cost them the game yesterday, though Georgia did their best to bail them out with some tactical mistakes of their own.

Having already converted two fourth downs on an epic drive that began on their own 20, Tennessee had a first and goal at the Georgia 7 yard line with 2:41 remaining in the game and the score tied, 24-24.  The appropriate tactic here is to to kneel the ball three times and kick the field goal, costing Georgia it's final timeout and leaving perhaps 30 seconds remaining at the kickoff.  For Georgia, the appropriate tactic is to allow Tennessee to score the touchdown. Although they apparently didn't realize this on first down, on second down their defenders looked surprisingly bewildered by a simple off-tackle run, and Rajion Neal took it in for the score with 1:54 remaining.


NCAA football and NFL football are not exactly the same, and there are important differences in end-game timing, most importantly the clock stoppage for first downs in college.  But the two games are still pretty analogous, so the Advanced NFL Stats win probability calculator can give us a rough idea of the relative value of the two endgame states Tennessee had to choose from.  Georgia got the ball back on their own 25 with 1:54 remaining down seven points, a situation which has historically led to an NFL victory 11% of the time.  If they had started with 30 seconds remaining down three points, they would have been in a situation which has resulted in NFL victory only 6% of the time.  Furthermore, the calculator does not take timeouts into account, and Georgia preserved a crucial final timeout by allowing the Tennessee touchdown.  The timing rules in the NCAA certainly give a college team a better chance of scoring a field goal with 30 seconds and no timeouts than an NFL team would have, but the same rules also help a team trying to score a touchdown with two minutes and a timeout to work with.  The relative difference between the two situations is arguably comparable in both leagues.

Advanced metrics aside, simple reasoning also indicates that being down three with 30 seconds allows much less margin for error than being down seven with two minutes to go.  Including the field goal attempt, 30 seconds might allow time for five or six offensive plays to move approximately 40 yards and attempt a 50 yard field goal, all of which only forces overtime.  A single incomplete pass or completion in bounds drastically reduces the offense's chances, and a sack or tackle short of the first down after the first play might end the game outright.  Finally, an NCAA defense can commit pass interference on any deep pass, knowing it only gives up 15 yards.

As it turned out, of course, Georgia drove down to score the tying touchdown with five seconds left, then won the game on a field goal in the first overtime period.  On the way, however, they made some strange endgame decisions of their own.  With 58 seconds remaining, Georgia QB Aaron Murray completed a 32 yard pass to Brendan Douglas, who chose to gain an extra five yards or so rather than go out of bounds and stop the clock.  Just over 30 seconds were left by the time the Bulldogs ran their next play, an inexplicable three yard checkdown completion that cost them their final timeout.

Tennessee helped the Bulldogs' final drive with two penalties, but the penalty they didn't commit may have cost them their best remaining chance to win.  Murray's scoring pass to Rantavious Wooten with five seconds left was well defended, but in this situation, Tennessee's defensive backs should have treated this like a two point conversion attempt, committing any penalty necessary to prevent the catch, since the maximum cost would be a retry from half the distance to the goal, a one-yard penalty in this case.  In fact, the final play, and arguably the two plays prior beginning at the two yard line with 16 seconds remaining, would have been the ideal situation for Buddy Ryan's "Polish goalline" tactic.  Tennessee could have put their entire 122 man roster on the field for the final three plays, ensuring Georgia would have only one real attempt at the tying score with no time remaining.  Georgia would accept the penalty of too many men on the field and move the ball half the distance to the goal after each play, but only after taking precious seconds attempting to score against a 10 to 1 numerical disadvantage.  Had they used this tactic, Tennessee could well have been penalized for a "Palpably Unfair Act," which gives the officials wide latitude to penalize the defense, prevent the clock from running, or even award a touchdown or a forfeit victory, though it's hard to predict what the officials might have chosen to do so late in a close conference game.

Sunday, September 15, 2013

Three Ways to Unleash Yale's BEAST

The "BEAST" formation (yes, in all caps) is the name that coach Steve Calande gave to the Yale formation when he rediscovered it around the year 2000:



Very little information on the original Yale formation and how it was used is available online.  Based on the number of backs in the backfield, it may have been developed around the time the rule requiring at least seven linemen was phased in between 1895 and 1910, as a way of recapturing some of the brute force the offense lost when the flying wedge was outlawed in 1894.  By the time the forward pass started gaining popularity after 1912, the formation's lack of split receivers may have started its decline.  The football books of the time tended to focus on human interest stories about star players rather than play diagrams (like most football books today), so I haven't been able to find any specific information on this offense in old sources.  If anyone has any information about which teams used the formation (Yale, presumably), when, and how, please let me know.

Once it had faded away, the formation remained largely forgotten for decades, with the exception of a few holdout Single Wing teams, particularly Notre Dame Box teams, who used it as a changeup.  A related system, the Fat Offense, was created by coach Bruce Eien as a changeup to his spread offense, though the Fat uses a balanced line and a wingback, and it's possible that Eien got the idea from Calande.

Saturday, August 31, 2013

Top Priority of the Summer: Defending the Read Option

With the read option play exploding onto the scene during the 2012 NFL season, defending the play has become one of the hottest topics of the offseason.  Articles by Grantland contributor and football blogger extraordinaire Chris Brown, Bleacher Report's Matt Miller, Greg Bedard of Sports Illustrated and others have all taken a swing at the topic in recent months.  These articles each highlighted various tactics and techniques, yet none took the opportunity to explain how the bidirectional structure of the read option makes it fundamentally different from other option plays, and how this influences the tactical arms race between offense and defense.


Picking Poison

"Read option" is not a precisely defined term.  In general, the term usually refers to plays where the quarterback and halfback run in opposite directions, and a defender to the side of the quarterback's run path is left unblocked as the option key:
As with all option plays, the offense is making a bargain with the defense: the offense gets to leave a defender unblocked, letting the quarterback occupy a defender with an option read who would otherwise have to be blocked or occupied with a fake.  The flipside of this arrangement is that the defense can dictate who carries the ball.  To defend an option play, the defense designates at least one defender for each potential ballcarrier (note this is a necessary but not sufficient condition for defending the play - many TV commentators suggest that the option can be stopped just by ensuring a defender is assigned to each of the dive, keep, and pitch, but there are still plenty of ways for an option play to succeed even if every defender knows his assignment, not least of which is that the defenders with those assignments might get blocked).  The defense also arranges these assignments to force the ball to the runner they perceive as least dangerous, and to funnel this runner to an area where they have more defenders, starting a cat and mouse game of adjustments and counter-adjustments between the two teams.

Sunday, August 18, 2013

The One-Play Offense

Simplicity is a virtue in offensive football.  If the appearance of complexity can be maintained for the defense, a simple offense leaves more practice time to focus on core plays, fundamental skills, and other facets of the game.  Vince Lombardi often wrote about the value of a simple game plan, executed well.  Paul Johnson condenses his weekly list of available plays until it can be written on a Post-It note. Hal Mumme, when he metamorphosed the West Coast Offense into the Air Raid, so valued simplicity that he made his offense solely right-handed, shedding all strong-left formations like so much spent snakeskin.

An offense consisting of six plays, or even ten, would be extremely simple by modern standards, but the Holy Grail of offensive simplicity would be a system consisting of a single play.  When the aforementioned Mumme was coaching at Southeastern Louisiana University, I attended a clinic there where he mentioned just such an offense.  The (apocryphal?) story he told was of a High School coach in Kentucky (Mumme was previously head coach at the University of Kentucky) who had great talent on his team but no time to practice - all his best players prioritized their commitments to their school and club basketball teams, and played football when time allowed.  His response was a single-play offense, and the single play was the Air Raid "Mesh":

Saturday, August 10, 2013

The Bears' Spread Offense, 1950 edition

Don Faurot's 1950 book on the Split-T formation1 is a classic of offensive philosophy.  The offense he invented ("football's newest development") was both highly successful in its own right and led directly to the creation of the triple option, one of the major evolutionary moments in football history.  But the book is more than just the introduction of a new offense or the seeds of a future breakthrough.  In over 350 pages it also covers the kicking game, the leadership challenges of coaching and administration, football watching tips for novice fans, and defensive strategy for combating the popular offenses of the 1940's.  The defensive chapter in particular shows how wide open some of these older offenses were, and how much today's spread offenses owe to schemes developed before World War II, or earlier.

On page 273 he describes how to defend against the "Bears' Spread," which he attributes to legendary Chicago Bears player/coach/owner George Halas.  Here's the formation, and the defense he suggests:



Whoa.  Now that's a spread formation.  The tailback is ten yards deep, 3 or 4 yards deeper than a modern quarterback in the shotgun, but otherwise, this looks ready to use.  Faurot doesn't give much detail on the offense, other than to be prepared for a tailback sweep in either direction, a screen to the tight end on the left, and a pass to the snapper, who is eligible because he's also the right end.  To contain this attack, he suggests a 4-4 defense in a Cover 3 zone, putting two linemen over the main offensive line, one over the snapper, and one in the massive right A gap.  The defensive ends (we would call them outside linebackers) cover down on the widest receivers and take the flats in the zone.

From a modern perspective this defense looks comically soft.  The pass defense isn't even close to properly covering down on the three receivers to the right, a defensive guard is covering the snapper despite Faurot's warning, and the numbers on the screen to the left tight end still look pretty great.  Even for an era before the option, before pass routes were timed with the QB's drop, before the wide receiver screen or even the wide receiver as we know it, it's hard to see how this would be sound defense.  If you lined up in this formation this Friday night, here's one guess at what you might see instead:



Now there are six defenders for the six man offensive line, the snapper has a legitimate undercoverage defender on him, and each receiver has a defender in position to discourage the hot throw or quick screen.  The defense is slightly outflanked to the tight end side by a fraction of a man, since the Will LBer doesn't want to align too far to either side of the snapper, but it's reasonably sound given the challenges it's facing.  There are plenty of other fronts and coverages possible here, (Cover 0 comes immediately to mind), but the threats to the defense remain obvious: five immediate receivers (including a quads set), the QB sweep, and those wide open A gaps.  For a team that already uses the shotgun, this formation could be an easy addition that would drive opposing teams nuts.  Here are a few possible plays that seem like natural fits:

  • QB sneak/sweep (blocked to both sides at once, QB chooses), with snapper pop pass
  • Quick screens to the WRs or the left TE
  • WR fly sweep, or read option, or inverted veer, to the left
  • Quads passing game, e.g. flood route plus whip
  • Slow screen to an ineligible lineman (must be a backward pass)
An hour or two of practice seems like enough to install this changeup formation with two or three plays, but the defense would likely spend several hours preparing for it, and opposing coaches many more hours thinking about it, instead of the base offense.  In a game, it could be deployed like any other changeup: right before halftime (to steal the other team's halftime adjustments) or right after (to steal their second half timeouts).  

If anyone dares to use this in a game, two reactions from fans and opponents are fairly predictable.  First, that anything so unusual is "not football" and that it is not sportsmanlike to create something new or use a tactic isn't already common on Saturdays or Sundays.  Second, that this is obviously just the next evolution of the modern game as it becomes ever more wide open and oriented toward skill and speed.  Either way, you can refer them to Don Faurot, creator of football's newest development of 1950.


1 Faurot, Don. Football: Secrets of the "Split T" Formation. Prentice-Hall, 1950.

Power blocking the read option

The redundantly named "read option" play caught on in the NFL last season with breakthrough performances from the likes of the 49ers' Colin Kaepernick and the Redskins' Robert Griffin III, after a decade of success in the major college ranks.  The play nearly always uses zone blocking, occasionally with a lead blocker, but there's no reason why it couldn't be run with power blocking instead.  In fact, for some teams power blocking the play may be a more efficient use of available talent.


Background

The term "read option" doesn't have a precise definition, but it normally refers to a family of related plays with two common characteristics:

  1. the quarterback executes an option in which he and the other potential ballcarrier run to opposite sides of the formation, and
  2. a defender to the side of the quarterback's run path is left unblocked as the option key.
Here is a basic diagram of the zone read play.  Extraneous receivers are left out of the diagrams for simplicity, but they would be stalk blocking or running their defenders deep depending on the defense.

In the basic zone read, the line executes an inside zone blocking scheme, here leading to combo blocks by the LG and C on the nose tackle, and by the RG and RT on the 3 technique.  These defenders are blocked horizontally along the line of scrimmage in the direction of their initial movement, with the offensive lineman away from this movement scraping off to the linebacker behind.  Q reads the unblocked DE and gives every time unless the DE chases down the line toward the F back.  F, if he gets the ball, runs to the heels of his linemen, reads their blocks, and cuts to the open lane.  

Hello, world

This is my blog.  There are many like it, but this one is mine.

I will post in this space my musings on American football strategy and tactics, for the coach or the fan.  Some of my posts may be applicable to the youth or major college level, but most of the time I will be thinking in the context of the High School varsity level.  What I write will have been researched to some varying degree, but it will be my opinion - you can keep the conversation going and let me know if I made a mistake or an omission in the comments.  I will read the comments, make corrections, and give credit where it's due.  Hopefully, this will be a space where everyone can learn, including me.  Thanks for reading.

Will