Tactics
Play action pass explained: why the fake still works
How a play action pass attacks a defender's assignment rather than his speed, the routes built behind it, and the disputed evidence that it needs no run game.
By CricketTaken EditorialPublished Tactics20 min read
A linebacker takes two steps forward. That is all. He does not fall over, he does not lose anybody, he is not beaten by any physical measure you could put a stopwatch to. Two steps, at the moment a receiver is running through the space he just left, and the completion goes for eighteen yards.
That is a play action pass explained in its entirety, and the reason it remains the most reliably productive call in American football is that it does not attack anybody's athleticism. It attacks their job description. A linebacker who steps up on a run fake is not being fooled about the world. He is doing the first thing his assignment tells him to do, in the correct order, and the offence has arranged for that correct first thing to be wrong.
Almost everything written about the play starts with the handoff fake and stops at "it freezes the linebackers". The interesting parts are all downstream of that: which defenders actually move, why the ones at the second level matter far more than the ones on the line, why the routes built behind it are shaped the way they are, and the genuinely disputed question of whether any of it requires the offence to run the ball well.
Play action pass explained as an attack on an assignment, not an athlete
Start with what a defender is actually doing on a normal snap.
Every defender who is not a deep coverage player has two responsibilities and a priority order between them. Against a run he has a gap to fill, a specific slot between two offensive linemen that he is accountable for, and if he is late to it the running back goes through it. Against a pass he has a zone to reach or a man to carry. He cannot do both at once, so he is taught to read a key, decide which play it is, and go.
The key is usually the offensive linemen in front of him. If they fire out low and come forward, it is a run. If they set back with their hands up, it is a pass. That read takes about a fifth of a second and it is the single most drilled habit in defensive football.
Play action puts a run key in front of a defender whose job then requires him to move towards the line of scrimmage. He is not guessing. He is executing. And two or three steps later, when the ball is still in the quarterback's hands and a receiver is running through the area he vacated, he is out of position in a way that no amount of recovery speed can fully repair, because a coverage defender's value comes from where he is standing rather than from how fast he can move.
That is the entire mechanism, and it explains three things that otherwise look odd. It explains why the play works on the first snap of a game before any run has been attempted. It explains why it works against fast defences as well as slow ones. And it explains why the defensive answer is never "run faster" but always "look somewhere else", which is a much harder thing to coach.
- Second-level defenders with a run fit4
- Rushing the passer regardless4
- Deep defenders outside the run fit3
A conventional four-rush, seven-cover defence. The split is structural rather than measured: it counts which defenders have a run responsibility that a fake can exploit.
Show the numbers
| Item | Value |
|---|---|
| Second-level defenders with a run fit | 4 |
| Rushing the passer regardless | 4 |
| Deep defenders outside the run fit | 3 |
The mechanics, and the part of the fake that matters most
The quarterback's role is the most visible and the least important. He opens his shoulders to the side of the run, extends the ball into the running back's belly, rides it a beat, and pulls it out. Good fakes hide the ball against the far hip on the way out, which is why quarterbacks are coached to carry it on the hip rather than back at the chest. It looks good on television and it is not what the second level is reading.
The running back's role matters more than the quarterback's. He has to run the track of the actual run play, at the actual speed, with his arms closed over the ball he does not have, and then continue through the hole and take on a defender or peel into a route. A back who slows down after the mesh point tells every linebacker the truth about half a second early, which is the entire margin the play is trying to buy.
The offensive line is the whole thing. This is the part almost every explanation skims, and it is the one that decides whether the fake works.
Linemen have to show run blocking. That means coming off low and hard, with the same first step and the same helmet placement as the run the offence actually runs, because the defensive linemen and linebackers are reading exactly that. Then they have to recover into pass protection against rushers who are now coming hard, from a body position that is much worse for pass blocking than a normal set. It is a technically miserable assignment and it is why some offensive lines are visibly better at play action than others.
The rulebook then puts a hard fence around it. An ineligible offensive player, which means the interior linemen, may not advance beyond the line of scrimmage before a forward pass is thrown, and in the professional code the allowance is one yard. In the college code it is three. That single difference is why run fakes and run-pass options are built differently in the two games: a college line can drive a defender genuinely downfield and still have the pass be legal, while a professional line has to sell a run with its pads and its feet while going almost nowhere. The pull, the down block and the second-level climb all have to be faked convincingly within a yard, and a lineman who gets carried away by his own performance draws a flag that costs more than the play was worth.
The receivers have a smaller job that is easy to see once you know to look. On a genuine run they would be stalk blocking, so on the fake they take a hard release and mime the first movement of a block before breaking into the route. A receiver who sprints vertically off the snap has told the corner it is a pass before the quarterback has finished turning.
- The formation presents a run lookUsually a tight end or an extra blocker, often a back offset to one side, and increasingly a quarterback under centre rather than in the shotgun, because a handoff from under centre is hidden and a shotgun handoff happens in plain view.
- The line fires out lowThe first step and the helmet placement match the offence's actual run play. This is the key the second level is reading, and it is the only part of the fake most linebackers ever see.
- The quarterback and back meshThe ball is extended, ridden and withdrawn. The back keeps running his track at full speed with his arms closed, because a back who decelerates gives the play away half a second early.
- The second level takes its run stepsLinebackers and any safety in the run fit move downhill to fill gaps. Two or three steps is all the play is asking for, and it is all the play needs.
- The line recovers into protectionRushers are now coming at linemen who are low, forward and out of position, which is the price the offence pays for the disguise and the reason play action sacks are uglier than ordinary ones.
- Routes arrive behind the vacated zoneThe crossing and intermediate routes are timed to reach the space at the moment the defenders who owned it are furthest from it, which is why they are deeper and slower-developing than drop-back concepts.
- The quarterback finds the throw or moves the pocketHe is either resetting his feet at the top of the fake or continuing away from it on a boot, and the throw is usually to a receiver moving across his body rather than standing still.
The sequence is fixed. What varies is which defenders read which key, and how long the offensive line can hold the disguise before it has to become pass protection.
The second level matters more than the first, and the reason is arithmetic
The defensive line is barely affected by a run fake, and this surprises people who assume the whole defence bites.
A defensive lineman's first responsibility on almost every snap is to get into the backfield or to hold a gap while doing it. He rushes on a run and he rushes on a pass. A run key changes how he plays with his hands and whether he squeezes down or works upfield, but it does not change his address: he was going towards the quarterback either way. Run blocking does slow the rush, because a lineman coming off low is harder to run around than one setting back, so the offence gets some extra time from the front. That is a real benefit and it is not deception.
The deception is spent almost entirely on the four or so defenders who sit between the line and the deep coverage. Their positioning is the product, and here the arithmetic is unforgiving.
Consider a hook defender who, on an ordinary drop-back pass, would retreat to about twelve yards and settle in the throwing window. Now give him a run key. He takes two hard steps forward, recognises pass, turns and runs, and finishes the play at six or seven yards instead of twelve. Every number here is invented, chosen to show the shape rather than to report anything.
Five yards of depth does not sound like much until you notice what lives at fifteen to eighteen yards on a passing concept. That is where dig routes break, where deep crossers arrive, and where a tight end running a seam route gets past the underneath coverage. A defender at twelve yards is in the throwing lane for those routes. A defender at seven is not in the lane at all, and he is facing the wrong way while the ball travels.
The second consequence is that the throw becomes easier as well as more open. A quarterback throwing over a defender at twelve yards has to put the ball on a specific trajectory. A quarterback throwing over a defender at seven has a window that is taller as well as wider, so the throw tolerates more error. Play action does not only create receivers who are open. It creates throws that are simpler, which is why quarterbacks of very different quality all show better numbers on it.
The routes built behind it, and why they are all crossing the field
Look at a collection of play action concepts and the same shapes keep appearing: deep crossing routes, dig routes, deep overs, post routes with a crosser underneath, and a back leaking out late. That is not fashion. Each of those shapes is a solution to a specific problem the fake creates.
Crossing routes exploit horizontal displacement. A defender who has stepped downhill has lost depth, but a defender who has stepped downhill and towards the side of the run has also lost width. A route that travels across the formation runs through the space he lost in both dimensions, and it does so with the receiver already at speed rather than breaking out of a stem.
They arrive late, which suits a play that develops late. A play action pass takes longer than a drop-back pass. The mesh point costs time, the quarterback's reset costs time, and the line's recovery into protection costs time. A quick route breaking at five yards is useless here, because it arrives while the defenders are still where they were. Deep crossers and digs arrive at the right moment by accident of their own geometry.
They are thrown across the quarterback's body into a moving target. A receiver crossing the field is easier to lead than a receiver breaking to a spot, and the throwing lane is wide open horizontally even if it is narrow vertically. This is the throw that play action manufactures more reliably than any other.
The vertical routes are there to hold the deep defenders. Almost every play action concept sends at least one receiver deep down a seam or a post, and most of the time he is not the read. His job is to keep a safety honest so that the safety does not fall off and cover the crosser underneath him. The concept is really two routes attacking one defender at two depths, which is the same bind logic that makes a Cover 2 corner miserable on a smash concept work.
The back leaking out is the cheapest completion in football. He has already sold the run, everybody has already forgotten about him, and he releases into a flat that a linebacker vacated by stepping into the run fit. It gains eight yards, it looks like nothing, and it is available roughly whenever the offence wants it.
There is one further shape that appears constantly and is worth naming because it is the purest expression of the idea: a deep over route, in which a receiver from one side runs across the formation at fifteen to twenty yards while the linebacker who should be under it has stepped up and the safety who should be over it has been pulled towards the vertical route on the other side. That route is not open because anybody was beaten. It is open because two defenders each did a correct thing that the concept had already anticipated.
It buys time as well as space, which is the half nobody counts
The yardage numbers get the attention. The protection benefit is at least as valuable and shows up in a different column.
Three separate things slow the rush on a run fake.
Run blocking is physically harder to beat quickly than pass blocking. A rusher expecting to run around a lineman who sets back finds instead a lineman coming at him low, and the first move of his rush plan is wrong. He is not blocked for longer because the blocker is better, he is blocked for longer because he attacked the wrong shape.
Edge rushers with contain responsibilities have to respect the run before they turn the corner. A defensive end who screams upfield on a run to his side has just let the back outside him, so he is taught to squeeze first and rush second. On a fake, that squeeze is a wasted step at exactly the moment his speed mattered.
And the quarterback's launch point often moves. On a straight drop-back the rush knows where he will be. On a play action pass he may be a step deeper, a step to the side, or twelve yards outside the pocket on a boot, and pass rushers who converge on an empty spot have effectively been blocked by geography.
There is a cost hiding in this. When play action fails, it fails badly. The line is in a worse body position, the quarterback has his back to the defence for a beat and cannot see an unblocked rusher arriving, and the routes take longer to come open. The negative plays produced by a broken play action snap are longer sacks and uglier ones. Any honest accounting of the play has to put that against its yardage advantage, and much of the reluctance to call it more often is really a reluctance to accept its worst outcomes.
The bootleg, the roll-out, and the price of cutting the field in half
The boot family takes the same fake and adds movement. The quarterback fakes the handoff to one side and then runs to the other, throwing on the move or continuing as a runner. The version with no blocker escorting him is usually called a naked bootleg, and the version with a puller leading him is a waggle.
What it gains is large. It removes the quarterback from the interior rush entirely, which matters most for offences whose interior protection is the weak point. It puts him outside the pocket where the rules on intentional grounding are more forgiving, so a play that breaks down can be thrown away safely. It forces defenders to run laterally, which is the slowest thing a defender does. And it presents a genuine two-way threat, since a quarterback rolling with the ball and an open field in front of him is a runner until somebody accounts for him.
What it costs is that half the field disappears. A quarterback rolling right cannot realistically throw back across his body to the far left hash, so the concept sends two or three receivers to the boot side at different depths, usually a deep route, a crossing route arriving at intermediate depth, and something shallow in the flat, and the read becomes a simple high-to-low progression on the run. It is one of the easier reads in football, which is why boot concepts are staples for young quarterbacks and for offences that want to simplify a passing game under pressure.
The unsolved problem is the backside defensive end. He is on the side the quarterback is running towards after the fake has pulled the line the other way, and if he is unblocked he is running at the quarterback with a clear view. Offences handle him three ways: leave him and out-leverage him with speed, chip him with a tight end who then releases into the flat, or pull a lineman to meet him and accept that the pull itself is a tell. None of the three is fully satisfactory, and the way blocking schemes distribute those responsibilities is the same problem that a zone blocking scheme solves on running plays, which is one reason boot concepts are usually married to a specific run family rather than bolted onto any offence.
The evidence that play action does not need a run game, and the honest caveat
Here is the finding that upended a century of coaching orthodoxy, and it is worth stating carefully because it is routinely overstated.
Published analyses of play-by-play data, most prominently the work done at Football Outsiders on rushing success and play action, found no meaningful relationship between how often a team ran the ball, how successfully it ran the ball, and how much more it gained per play on play action passes than on ordinary drop-backs. The advantage showed up for teams that ran constantly and teams that barely ran, for teams that ran well and teams that ran badly. Season after season, the great majority of teams gained more per play with a run fake than without one.
A second line of evidence went at the mechanism directly. Josh Hermsmeyer, working with player tracking data at FiveThirtyEight, examined whether defenders became less responsive to run fakes as a game went on, which is what you would expect if the fake worked by building a false belief that could be updated. He found that linebackers reacted to the first play action snap of a game about as much as they reacted to the eleventh. The response looks less like a belief and more like a reflex, which is exactly what you would predict from the assignment argument at the top of this article.
Both findings point the same way and both are real. Now the caveat, which the confident version of this argument usually leaves out.
Play action is not called at random. It is called disproportionately on first down and on second and short, in situations where the defence has to respect a run, often against heavier defensive personnel, and rarely on third and twelve. Ordinary drop-back passes include every obvious passing down in football, faced by nickel and dime defences with the pass rush teed up. Comparing the two therefore compares two different situations as much as it compares two different plays, and some part of the measured gap belongs to the down and distance rather than to the fake.
Analyses that control for situation generally find the advantage survives and shrinks. That is the honest position: play action helps, it helps without a good run game, and the raw yards-per-play gap overstates how much it helps. Anybody quoting the raw gap as the value of the fake is quoting a number that has a situation baked into it.
There is a second caveat with a different shape. The evidence tells us what happens at the rate teams currently call the play. It does not tell us what would happen if every team doubled that rate, and there are good reasons to think the advantage would erode, because the entire mechanism depends on defenders having a run responsibility they take seriously. An offence that never runs eventually faces a defence that stops honouring the fake, and the fact that the correlation is absent across the range of behaviour we have observed does not mean it would remain absent outside that range. This is the same category of reasoning problem that runs through the argument about going for it on fourth down, where the measured value of an aggressive choice is a marginal value at observed frequencies.
Why it is called less often than the numbers say it should be
If the play gains more per snap, the obvious question is why offences do not call it on every early down. Several answers, of descending respectability.
It cannot be run from every formation or tempo. A convincing run fake wants a back next to the quarterback, a defined run look, and usually a quarterback under centre. An offence in an empty formation with five receivers split out has nobody to fake to. A hurry-up offence sprinting to the line has no time to present a run look and no huddle in which to disguise it. The structural shift towards the shotgun over the last two decades cut deeply into the number of snaps from which a genuine fake is even available, and early-down passing from under centre has fallen by a very large multiple over that period. Some offences have begun quietly rebuilding it.
A shotgun fake is a weaker fake. With the quarterback under centre, the handoff happens behind the line with the ball hidden. From the shotgun, the mesh point happens in plain view, several yards deep, and a linebacker with any discipline can see the ball. Coaches disagree about how much this matters and the disagreement is genuine, with some analysts arguing that under-centre fakes are far more effective and others pointing to offences whose shotgun play action outperforms their under-centre version. The safest reading is that the fake is one input among several and that scheme quality can compensate for the weaker mesh.
It is slow, and slow is dangerous. The concepts behind it are deep concepts, and deep concepts need protection to hold. Against a defence that is winning up front, the play that takes the longest is the play that gets your quarterback hit.
It hurts your worst outcomes. Coaches are not maximising average yards. They are managing a distribution with a hard floor, and a sack on second and eight is worse than an incompletion by more than its yardage. Play action shifts both tails outwards.
And there is straightforward belief. A large number of coaches still hold that you must establish the run before the fake will work, which the evidence does not support, and that belief has real consequences for play calling because it makes the fake feel unearned early in games. The belief is not stupid, incidentally: it is an entirely reasonable inference from the mechanism, and it happens to be wrong about what defenders actually do.
- 17Depth of the crossing route, yards
- 5Yards the hook defender loses to the fake
- 0.4Extra time the fake buys the protection, seconds
- 20Yards of field the crosser travels laterally
Invented figures, chosen to show why the routes behind a fake are deep and slow rather than quick. Not a measurement of any play.
How defences train themselves to ignore the fake
The defensive answer is not to be faster. It is to move the eyes to a key that a fake cannot corrupt, and then to arrange the personnel so that fewer defenders have a run responsibility to abuse in the first place.
Read the linemen, not the backfield. A defender who watches the quarterback and the back is reading the part of the play that is designed to lie. A defender who reads the offensive linemen in front of him is reading the part that has to tell the truth eventually, because those linemen have to be in pass protection before the throw and cannot legally get more than a yard downfield. Coaching a linebacker to key a guard's hips rather than the mesh point is the single most effective anti-play-action measure available, and it is hard because it fights an instinct.
Look for the pull that is not a pull, and the set that is too early. Fine-grained keys exist. A guard who pulls flat rather than downhill, a tackle whose hands come up a beat early, a back whose track is a step wider than usual on the same look. Good defences accumulate these tells during a week of preparation, which is why play action often works better in the first quarter than the fourth against a well-prepared opponent.
Play two-high shells so fewer defenders are in the run fit. A defence with two deep safeties has already removed both of them from the run responsibilities that the fake exploits. The cost is a lighter box against the actual run, which is the trade every coverage shell makes, and it is why the rise of two-high defences and the rise of interest in the running game happened together.
Use lighter personnel. A nickel defender substituted for a linebacker is a defensive back whose instincts are coverage-first, so his response to a run key is smaller. Extra defensive backs now play the majority of snaps in the professional game, which has quietly reduced the number of defenders on the field who will bite hard on anything.
Split the field. Defences increasingly play one coverage on one side of the formation and a different one on the other, so that the side where the offence has loaded up gets the structure that best handles a deep crosser while the other side plays something cheaper. It complicates the quarterback's read on a play whose entire premise is that the read will be simple.
Make the fake impossible to trust. The run-pass option pushed this the other way, by giving the quarterback an actual choice after the snap rather than a predetermined one, so a defender who ignores the run key is punished by a real handoff. That arms race is the reason the two concepts are best understood together, and what a run-pass option asks a defender to decide is the mirror image of what a run fake asks him to forget.
None of these fully solves it. They reduce the magnitude of the reaction rather than removing it, which is the most honest thing that can be said about defending a play that exploits the correct execution of a defender's own assignment.
What to watch on the next run fake you see
The play is one of the easiest things in the sport to read once you know where to point your eyes, and almost impossible if you watch the ball.
Watch one linebacker, not the quarterback. Pick the one over the guard on the side of the fake and watch only him for the whole snap. The two steps are the play. Everything else is consequence.
Check where the quarterback is standing at the snap. Under centre means the ball disappears at the mesh point and the fake has its full effect. Shotgun means the handoff happens in daylight and the fake is doing less work, which tells you something about what the offence thinks it needs.
Count the seconds before the throw. Play action passes take noticeably longer than drop-backs. If the throw comes out quickly, the concept was a leak or a flat route rather than a deep crosser, and the offence was taking the cheap completion.
Look for who is uncovered at fifteen yards. The route that comes open is almost always a crosser or a dig arriving at the depth a second-level defender abandoned. Naming the vacated zone before the replay does is the habit that turns this from a highlight into a diagram.
Note the down and distance. A fake on first and ten is the offence using a situation that helps it. A fake on third and nine is the offence trusting the concept on its own, and those are the snaps that tell you what a coordinator really believes.
Then look at the offensive line's feet on the replay. Low and forward for a beat, then recovering with the hands up. If the line pass-set immediately, the fake was cosmetic and the second level never had a decision to make.
Do that for a few games and the play stops looking like a trick. It looks like what it is, which is an offence reading a defence's own coaching manual back to it, one assignment at a time. The rest of the schematic vocabulary, from coverage shells to blocking families, sits in the American football section, and most of it turns out to be the same argument about who is responsible for which piece of grass.
Common questions
What is a play action pass in football?
A play action pass is a passing play that begins by imitating a running play. The quarterback fakes a handoff to a back, the offensive line sells run blocking for a beat before dropping into protection, and receivers who appear to be blocking release into routes. The aim is to make defenders who have run responsibilities commit a step or two forward, opening the intermediate zones behind them.
Does play action require a good running game?
Published analyses of play-by-play and player tracking data have repeatedly failed to find a relationship between how often or how well a team runs and how much it gains from play action, and the effect shows up for teams that run rarely and badly as well as for teams that run constantly. The honest caveat is that play action is called in situations where passing is easier anyway, so part of the measured advantage belongs to the situation rather than to the fake. The evidence that some advantage survives is strong; the size of it is contested.
Which defenders does play action actually fool?
The second level, meaning linebackers and any safety with a run responsibility. Defensive linemen are largely unaffected because they rush on almost every snap regardless, and deep defenders in a two-high shell were never part of the run fit. The whole value of the play comes from a small number of defenders moving two or three yards in the wrong direction at the moment a route is arriving behind them.
What is a bootleg in football?
A bootleg is a play action pass on which the quarterback fakes the handoff and then runs to the opposite side from the fake, throwing on the move. It moves the launch point away from the interior rush, puts the quarterback outside the pocket where he can throw or run, and forces defenders to chase the ball horizontally. Its weakness is the backside defensive end, who is unblocked or lightly blocked and running straight at the quarterback's blind side.
Why do teams not run play action more often?
Partly belief, since many coaches still hold that a run fake needs a credible run game behind it, and partly structure. Play action takes longer to develop, is hard to run convincingly from empty formations or a hurry-up tempo, turns the quarterback's back to the defence for a beat, and produces worse outcomes when it fails. There is also a reasonable argument that the measured advantage is a marginal one at current usage rates and would shrink if every team called it twice as often.
Filed under American Football·nfl · offense · tactics · scheme · passing game