Switching defence in the NBA: the mismatch you invite

A centre catches the ball nine feet from the basket with a guard on his back. He has half a foot on the man behind him and a lot of weight, the passing angle is clean, and the nearest help defender is standing in the far corner watching it happen. Nothing broke down. Nobody was beaten off the dribble, nobody was late, no rotation failed.

The defence chose this. It chose it four seconds earlier, when two of its players traded assignments rather than fight over a screen, and it chose it knowing this was the likely consequence.

That is switching, stated honestly. Every other answer to a ball screen eventually asks a third defender to leave somebody, which turns the possession into a scramble and hands the offence an open man somewhere on the floor. A switch refuses to do that. It pays with a mismatch instead. American coaching staffs write it as switching defense, the rest of the English-speaking world writes defence, and either way the whole instruction is one word shouted twice.

Whether that is a good trade is not a question about the coverage. It is a question about the ten players on the floor, and the answer changes from possession to possession inside the same game.

What a switch actually buys, and it is not stopping the screen

The common description of switching is that it neutralises the screen. That is true and it is the least interesting thing about it.

Look instead at the three defenders who are not involved. In a drop, the big retreats and somebody has to tag the roller, which means an off-ball defender steps in and a shooter is briefly alone. In a blitz, two defenders go to the ball and three are left responsible for four attackers. In every version of help defence, the price of solving the two-man action is paid by a man standing thirty feet away who has done nothing wrong. A switch is the only coverage in the sport where the other three defenders do not move at all.

That is the purchase. Not the screen. Three shooters stay guarded, no closeout is required, and the corner never opens.

This is why switching arrived in force at the same time as five-out spacing, and the geometry behind that is worth reading separately in what the arc did to the distances every defender has to cover. Once the fifth attacker stands behind the line rather than under the basket, every rotation is longer, every closeout is later, and the cost of asking a third defender to leave somebody rises sharply. Switching is the answer that refuses to pay that rising price.

There is a second purchase that gets less attention and matters nearly as much. A switch kills choreography.

Most half-court offence is not one action, it is a sequence: a screen that sets up a hand-off that sets up a second screen for a shooter coming off the first. Each link in that chain depends on a defender being late from the previous one. Switching breaks the chain at the first link, because no defender is ever behind. The staggered screens produce nothing, the flare produces nothing, the back screen set on a retreating big produces nothing because no big is retreating. A team that has spent a week installing a set watches it collapse into a man dribbling at the top of the key with nine seconds left, which is not what anybody drew.

A third purchase is simplicity. Help defence is a set of conditional rules that five tired players have to apply correctly in half a second. Switching is one rule. It survives fatigue, it survives a new player arriving mid-season, it survives the fourth quarter of the second night of a back-to-back, and it does not require the low man and the top man to agree about anything before it works. There is real value in a coverage that cannot be miscommunicated, and it is a large part of why switching spread through the sport faster than any scheme in decades.

• 2Defenders who change assignment on a switch
• 0Off-ball defenders who have to move
• 3Defenders involved in a scram switch
• 3Seconds a defender may stand in the paint unattached

Structural counts and published rules, not measurements. The defensive three-second clause is what stops a defence simply parking a rim protector under the basket to cover for a mismatch.

The zero in that figure is the entire argument for the coverage. Every other row on a defensive coverage sheet has a number greater than zero in that position, and each of those bodies has to travel, arrive and contest.

The mismatch is the bill, and the offence decides when to collect

What a switch hands over is the right to choose the matchup, and that right is more valuable than it sounds because it is exercised at the offence's convenience rather than the defence's.

The offence does not have to take the first pairing it gets. It can screen again, and again, until the specific defender it wants is standing in front of the specific attacker it wants. This is what commentators call hunting, and the mechanics are unglamorous: a guard walks the ball up, calls for a screen from whichever teammate is guarded by the man he wants to attack, and if the defence trades assignments cleanly the first time, he simply asks for another one.

Then there are two attacks, and they are opposites.

A small attacker with a big man on him wants space. The other four clear out, the floor empties, and the possession becomes a race between a first step and a set of feet that were not built for it. The attacker does not need to beat his man cleanly. He needs to force him into a decision at speed, because a large defender who has to change direction twice is either late or fouling.

A big attacker with a small man on him wants the opposite. He wants no space at all. He seals the smaller defender behind him, gets as close to the basket as the rules allow, and asks for the ball to arrive before anybody can do anything about it. The dribble is nearly irrelevant here. The work is done with a forearm and a hip before the pass is thrown.

Both attacks cost the same thing, which is time.

17%17%25%42%

• First screen and the read4
• Re-screen to reach the wanted matchup4
• Dribbling the mismatch into position6
• Clock left to actually score10

A constructed example with round invented numbers, chosen to show the shape of the trade rather than to measure any team or possession. The professional American shot clock is 24 seconds.

Show the numbers

Worked example: how a hunted mismatch spends a 24-second clock

Item	Value
First screen and the read	4
Re-screen to reach the wanted matchup	4
Dribbling the mismatch into position	6
Clock left to actually score	10

Read those four rows the way the defence reads them. Fourteen of the twenty-four seconds have been spent obtaining an advantage, and the advantage that remains has to be converted by one man against one man with the clock visible over his shoulder. A four-on-three created by a blitz starts with eighteen seconds and four options. A hunted mismatch starts with ten seconds and one.

Switching converts an immediate advantage into a delayed one, and delayed advantages are defendable. That sentence is the whole case for the coverage.

There is a subtler benefit that good defensive staffs count on. A switch can talk a good offence into playing a worse one. A team whose actual strength is movement, screening and quick passing will, on seeing a mismatch, stop doing all of that and stand still while one player dribbles. If the isolation that results is worth less per possession than the offence it abandoned, the switch has done something better than neutralising a screen. It has persuaded five players to volunteer for the defence's preferred version of the game. Not every mismatch is worth attacking, and the discipline to ignore one is rarer than the skill to exploit it.

Switching defence in the NBA is a personnel question before it is a tactical one

Every coverage is public. There is nothing secret about switching, and no coaching staff has ever been surprised by it. What separates the teams that can run it from the teams that cannot is not knowledge, it is bodies.

Two tests decide it, and both are about what happens after the switch rather than during it.

Can the big man stay in front of a guard for four seconds without fouling? Not stop him, not block the shot, simply contain him long enough that the possession has to become something else. Perimeter defence for a large player is mostly about the first two steps and about hands. The rules give a defender a right to his own vertical space and very little beyond it, so a big who reaches, leans or rides an attacker with a forearm turns a mismatch into free throws, which is worse for the defence than the layup he was trying to prevent.

Can the smaller defender hold his ground for two seconds without being sealed? He does not have to stop a post-up. He has to make the entry pass difficult and buy enough time for the repair described in the next section. Strength and a low centre of gravity matter more here than height, which is why a compact, powerful wing often survives switching onto a centre better than a taller and lighter one does.

Those two tests are why positional size became a phrase in front offices. A team that switches is not asking each defender to guard one position. It is asking each defender to be plausible against any of the five, because the offence chooses which pairing gets tested and it will test the worst one every time. A roster of five players within a few inches of each other in height, with comparable strength and length, can switch everything. A roster with one small guard and one immobile centre can switch almost nothing, because both of those players are a standing invitation.

That is a roster decision made in the summer, not a tactical decision made in a timeout. It is the same purchase described in what a team gives up to make every defender interchangeable: rim protection is traded away in exchange for never having to help. Buying switchability means buying a certain kind of player, paying for him, and living with what he cannot do at the other end.

The reverse is also true and rarely said. A team with one genuinely elite rim protector should probably not switch, because switching is the coverage that removes him from the rim. Sending your best interior defender out to the arc to guard a ball handler is an expensive way to solve a problem that a well-executed retreat solves for nothing, which is the argument made at length in why the big man sits back and concedes the jump shot.

The scram switch repairs a mismatch without doubling anybody

Once a mismatch exists, a defence has three options: live with it, double it, or fix it. Doubling is expensive, because it reintroduces the rotation the switch was supposed to avoid. The fix is the scram switch, and it is the piece of the puzzle that made switching viable against teams with a real post threat.

The mechanics are simple. A third defender, one currently guarding an attacker the defence is not frightened of, comes to the mismatch and trades assignments with his outmatched teammate. The guard who was about to be posted takes over the harmless man on the perimeter. The bigger defender takes over the post threat. Two defenders have changed jobs, nobody has been abandoned, and the offence's four seconds of hunting have produced nothing.

Three details decide whether it works.

Timing. A scram has a deadline and the deadline is the seal. If the big man has already established deep position with the ball on its way, arriving late means the new defender inherits a lost battle rather than preventing one. The repair has to happen while the post-up is still being set up, which means the third defender has to read the switch, decide and move before the offence has finished doing what it is obviously about to do.

Who scrams. The correct scrammer is the defender guarding the least dangerous attacker, and if that man is standing in the corner, the trade leaves a guard closing out to the hardest spot on the floor. If he is standing above the arc, the closeout is shorter and the whole repair is cheaper. A defence therefore scrams from wherever it is least frightened, which is exactly the same logic that governs who tags a roller in the way a defence organises the act of leaving somebody.

Who calls it. The scram is called by the defender arriving, not by the one being rescued. This sounds like a detail and it is not. The man being posted is facing away from the floor with an arm across his chest and cannot see who is coming. If he lets go before he hears the call, he has created an uncontested layup rather than a repair.

The cost of a scram is that it never quite eliminates the mismatch, it relocates it. The guard who escaped the centre is now attached to somebody else, and the man he was originally guarding is now covered by a defender who arrived out of position. A patient offence answers a scram by immediately screening for the man the scrammer left, which starts the whole argument again from a fresh angle.

1. 1The offence screens for the guard at the topThe screener is the man being guarded by the opponent's largest defender. Nothing about the screen is unusual; its only purpose is to force the two defenders to trade assignments.
2. 2The defence switches cleanlyNo separation is created, no help is required, and the three off-ball defenders never take a step. To this point the coverage has worked exactly as designed.
3. 3The offence now has the matchup it wantedA guard is attached to a much larger attacker and a large defender is attached to the ball handler. Four seconds of clock have been spent buying that.
4. 4The big attacker walks to the block and sealsHe does not dribble. He puts his back into the smaller defender, gets an arm across, and turns the possession into a race between a passing angle and a repair.
5. 5The weak-side defender starts moving earlyHe is guarding a non-shooter above the arc, which makes him the cheapest man on the floor to move. He calls the scram loudly enough that the guard hears it before releasing.
6. 6The trade happens before the entry passThe larger defender takes the post, the guard sprints out to the perimeter. A mismatch that took four seconds to create has been undone in about one.
7. 7The offence re-screens instead of forcing itThe entry pass is no longer available, so the ball handler asks for another screen from a different teammate, and the exchange starts again with fifteen seconds gone.
8. 8The shot arrives late and contestedWhatever goes up now goes up against a set defence with all five men attached. The defence has conceded a mismatch three times and never once had to leave a shooter.

A constructed possession described step by step. Every decision in it is individually defensible, and none of the sequence reports a real play, team or player.

Nothing in that sequence was clever. It was a coverage doing what it promised, an offence collecting what it was owed, and a repair arriving on time. The reason it is worth walking through is that the whole possession contains no help defence at all, which is the thing switching is actually selling.

Pre-switching means the screen never gets to do its job

The scram fixes a mismatch after it exists. Pre-switching prevents one.

A pre-switch is a trade of assignments made before the screen arrives, usually while the ball is still being walked up or while the offence is setting up. The defence looks at what is obviously coming, decides which of the two pairings it would rather have, and swaps voluntarily. When the screen is then set, the switch that follows produces the matchup the defence chose rather than the one the offence was hunting.

That is a better outcome than it sounds, because it inverts the ownership of the choice. Hunting works precisely because the offence picks. Take the picking away and a screen is only a screen.

The cost is that pre-switching requires recognition. Somebody has to know what the set is before it happens, which is a scouting and preparation problem rather than an athletic one. It is one of the few defensive tools that gets sharply better as a playoff series goes on, because by the fourth game a defence has seen the opponent's entries often enough to trade assignments as the first cut is made.

Two relatives of the pre-switch belong here.

Top-locking denies the screen rather than the switch. The defender guarding the intended screener plays above him, on the high side, refusing to let him reach the spot where the screen was going to be set. The action never happens and the offence has to reorganise. It is aggressive and it is risky, because a top-locked player who cuts backdoor is running at an unguarded basket.

Switching only one part of a sequence is the more common tool and the more useful one. A defence can switch the first screen and chase the second, or switch anything above the free-throw line and stay attached to anything below it, or switch every off-ball screen while playing the on-ball screen in a drop. The offence's chain of actions depends on the defence answering all of them the same way. Answering them differently is disruptive out of all proportion to the effort involved, because the second and third actions of a set were designed against an assumption that has quietly stopped being true.

Nobody switches everything, and the exceptions are the real scheme

A team described as a switching team does not switch every screen. The list of when it does and does not is the actual coverage, and it is written along four axes.

By personnel. The most common line in professional basketball is switching one through four and treating the centre as an exception, either by having him drop when he is the screener's defender or by scramming immediately whenever he ends up on a guard. A defence with two players it does not want on the perimeter has fewer options and has to hide both, which is difficult, because the offence only needs to find one of them.

By location. A screen set at the top of the key can be switched safely, because a mismatch created there is thirty feet from the basket and there is time to repair it. A screen set on the wing puts the mismatch closer to the block and closer to the baseline, where the repair is harder and the entry pass is shorter. Some staffs switch above the break and refuse to switch below it for exactly that reason.

By clock. Almost everybody switches late in the shot clock, whatever their base coverage. With five seconds left, a mismatch is theoretical: there is no time to hunt, no time to clear out and no time to seal. Switching in that window costs nothing and prevents the last-second slip or curl that most late-clock actions are built around. If you want to know whether a team understands its own scheme, watch what it does with six seconds left rather than with eighteen.

By situation. Inbound plays are switched by nearly everyone, because a rotation from a stationary start is close to impossible and every inbound set in the sport is a screening sequence designed to produce one moment of separation. The same applies immediately after a timeout, when the offence has drawn something specific and the defence would rather answer with one rule than with five.

There is a hybrid that has become common enough to name. Rather than trading assignments on contact, the nearest defender takes the man attacking downhill and everybody behind him moves up one, which staffs variously call a veer or a next switch. It is a switch in the sense that assignments change hands, but the exchange is between a defender and the driver rather than between the two men involved in the screen, and it exists to let a defence switch actions where a straight trade would create the exact pairing it is trying to avoid. How this sits alongside the other families of answer is laid out in the full menu of coverages and what each one refuses to guard.

The rebounding bill is the part nobody schemes for

Here is the cost that never appears on the whiteboard, because the whiteboard is about what happens before the shot goes up.

A box-out is a matchup responsibility. It is assigned in the same breath as everything else, and it assumes the man responsible for an attacker is roughly capable of holding him off the glass. A switch reassigns that responsibility at the worst possible moment, with a shot already in the air and no time for anybody to renegotiate. The guard who traded onto the centre four seconds ago is now the man standing between a much larger player and the rim, and he has about a second to find him, get a hip into him and hold.

He can do the first two. The third is physics.

Centre boxes out the centre, offensive rebound conceded25%

Guard boxes out the same centre after a switch, conceded40%

An invented illustration with round numbers, not a measurement of any player, team or season. The rates are chosen only to show the shape of the cost, and the arithmetic that follows in the text uses them.

Show the numbers

Worked example: what the switched box-out is worth

Item	Value
Centre boxes out the centre, offensive rebound conceded	25%
Guard boxes out the same centre after a switch, conceded	40%

Take that constructed example and finish the sum, because the number that matters is not the rebounding rate. Over ten possessions ending with a shot from the same spot, the difference between those two invented rates is one and a half extra offensive rebounds. If a second-chance possession is worth a point and a bit, and the example here says one and two-tenths, those one and a half rebounds come to about one point eight points across ten possessions. That looks small until it is multiplied by every switch a team makes across a season, and it is paid silently, in a category nobody attributes to the coverage.

The professional American rulebook makes it worse in a specific way. An offensive rebound does not restart the shot clock at twenty-four, it resets it to fourteen, which is enough time to run something real. The second possession therefore begins with the defence still holding the switched matchups it never wanted, still with a guard on a centre, still with the big out on the perimeter, and with no dead ball to reset any of it. The mismatch that took four seconds to hunt has been handed back for free.

That is why offensive rebounding is the most under-rated counter to switching. It requires no skill on the perimeter, no isolation scorer and no clever set. It requires one large player who runs at the rim whenever a shot goes up and one opponent who has just traded a small defender onto him. Against a switching defence, crashing the glass is not a gamble on a missed shot. It is an attack on a structural weakness the defence created on purpose.

There is a quieter version of the same problem. The switch usually happens near the arc, which means the defence's biggest player is often the furthest man from the basket when the shot goes up. The two best rebounding positions on the floor, either side of the rim, end up covered by whoever happens to be nearby. A defence that switches everything is frequently rebounding out of position even on the possessions where nobody is mismatched at all.

The post-up counter, and why it is slower than it looks

The classic answer to a switch is to put the ball in the big man's hands with a small defender behind him. Everyone knows this. What is less obvious is how much of the battle happens before the pass, and how many rules are involved in it.

The seal is the whole thing. A post-up against a smaller defender is won by the attacker establishing position while the defender is still recovering from the switch, which is a window of about a second. The forearm goes across the chest, the hips turn, and the attacker's job from then on is to be a wall rather than to move. If the pass arrives while he holds that position, the shot that follows is close to the basket against a defender who cannot contest it without leaving the floor early and conceding the whistle.

If the pass does not arrive, the rules start working against him. The offensive three-second clause means an attacker cannot stand in the paint waiting, so the seal has an expiry: hold it too long and the whistle removes the possession without the defence doing anything at all. A post-up is a timed event, and the passer is on the clock as much as the scorer is.

The small defender's tools are all about denying the angle rather than winning the physical exchange, which he will not win.

Take the space early. The defender who gets his body into the spot before the attacker arrives has turned a seal into a wrestling match ten feet further from the rim, which is a different and much worse shot.

Front the post. Stand in front of the attacker entirely, so the only pass available is a lob over the top. This is the strongest option and the most dangerous, because a lob thrown over a fronting defender arrives with nobody between the ball and the basket. Fronting only works if somebody behind is watching for the lob, which means the defence has reintroduced help, which is the exact thing it switched to avoid.

Dig at the ball. A teammate steps down from the perimeter to swipe at the ball as the attacker turns. It is effective, it opens the man he left, and once again the promise of the coverage has been broken.

Wait for the scram. The best outcome of all, and the reason the timing of the repair matters so much.

Notice the pattern. Every serious answer to the post-up except the scram switch requires help. A switching defence that is forced to help has stopped being a switching defence, and it is now running a version of help defence it never drilled, from starting positions that make no sense, with the wrong players in the wrong places. That is the real damage a post-up does, and it is larger than the two points.

The counter to the counter belongs to the offence. If a defence scrams reliably, the offence stops trying to score out of the post and starts using it as bait: seal, force the scram, then throw the ball out to the shooter the scrammer has just abandoned. The post-up becomes a way of moving a specific defender rather than a way of scoring, which is exactly what a ball screen is. The two-man game and the post-up turn out to be the same argument conducted at different distances from the basket, as the mechanics in what a ball screen asks a defence to answer make clear.

Why switching defence spread through the NBA, and where it retreated

Switching became common for three connected reasons, none of them fashion.

The first is spacing. Once every attacker on the floor was a threat from behind the arc, the help defender who used to solve everything became the most expensive man in the building. Sending him anywhere meant leaving a shot worth more than the one he was preventing. Switching is the only coverage that never sends him.

The second is that the rules stopped protecting the alternatives. Contact on the perimeter was legislated out of the professional game, which made containing a quick ball handler one-on-one much harder and increased the number of screens an offence could profitably run. At the same time the defensive three-second rule prevented a defence from stationing a large man permanently under the basket to cover for whatever the perimeter defenders could not do. The rim protector had to be attached to somebody, and a switching defence is one where everybody is always attached to somebody.

The third is that the players changed. A generation arrived at professional level with the size of forwards and the movement skills of guards, which made the scheme available in the first place. You cannot switch one through five without five players who can defend one through five, and for a long stretch those players did not exist in sufficient numbers.

Then it retreated, and the retreat is more interesting than the arrival.

It retreated because offences learned to hunt properly. Switching everything is only free if the offence declines the invitation, and once teams organised their possessions around obtaining and attacking one specific pairing, the coverage started paying its bill every time down the floor rather than occasionally.

It retreated because of the glass, for the reasons set out above, and because the second-chance cost took several seasons to become obvious to anybody watching a box score.

It retreated because fouls accumulate. A large defender guarding quick players on the perimeter and a small defender wrestling large players in the post are the two situations in the sport most likely to produce a whistle. A defence can be structurally sound and still lose a game because its two best players are sitting down with foul trouble in the third quarter.

And it retreated because the alternative got better. Rim protection did not stop working, it stopped being automatic, and the teams that kept a genuine interior defender found that a well-executed retreat answered most ball screens perfectly well while keeping their biggest man in the place where he is worth most.

What survived is not switching as a philosophy but switching as a tool, applied by list. Switch these four players and not the fifth. Switch above the break. Switch late in the clock. Switch inbounds. Switch when the alternative is chasing a shooter through three screens. Almost every serious defence now switches something, and almost none of them switch everything, which is a more sophisticated position than either of the extremes it replaced.

What actually breaks a switching defence

Four things, in rough order of how hard they are to obtain.

A large attacker who can handle the ball facing the basket. This is the one that genuinely cannot be schemed against. A post-up can be scrammed, fronted or dug at. A big man who catches the ball twenty feet out with a guard on him and drives is attacking the smaller defender in the one place where help is furthest away, and any help sent to him leaves a shooter behind the arc.

Offensive rebounding. Structural, repeatable, and available on every possession that ends in a miss regardless of what the offence ran to produce it.

Slips and ambiguity. Switching requires two defenders to agree instantly on who has whom. Anything that makes the agreement ambiguous produces a moment where both take the same man or neither takes anybody: a screener who runs into position and peels away without making contact, two screens set at once, a screen arriving on a defender's blind side, a hand-off that might be a hand-off and might be a keep. None of these attack the coverage's logic. They attack the half second in which it is communicated.

Pace. A defence that has to talk is beaten by a defence that has not finished talking. Attacking early in the clock, before assignments have settled and while the last switch is still being assessed for repair, produces more mismatches for less effort than any amount of hunting in the half court.

Against all four, the defence's best answer is the same, and it is not tactical. It is having five players who can each guard most of the other team, so that no pairing the offence obtains is worth attacking. That is why the coverage is bought rather than taught.

Watch four things and you will know what a defence is doing

Naming the coverage takes one glance, and it is not at the ball.

The screener's defender's hips at the moment of contact. Square to the ball handler and sliding onto his outside hip is a switch. Turned towards the rim and retreating is a drop. Stepping up with hands high is a hedge or a blitz. That single observation identifies the answer before the ball handler has taken his second dribble.

How many screens the offence sets before it stops running offence. If a team screens twice and then stands still while one man dribbles, the switch has bought the delayed advantage it wanted. If a team screens once and immediately attacks, the pairing it obtained was worth attacking straight away, which tells you something about both rosters.

The weak-side defender guarding the least dangerous man. If he takes two steps towards the mismatch, a scram is coming and the post-up will not happen. If he stays glued to the corner, no repair is available and the defence has decided to live with whatever the pairing produces.

The box-out on the first shot after a switch. This is the tell nobody watches. Find the smallest defender on the floor and see who is next to him when the ball hits the rim. If it is the opponent's largest player, you are watching the coverage's real bill being presented, and it will be presented again on the next miss and the one after that.

Do that for a quarter and switching stops looking like a way of avoiding work. It is a decision to pay later rather than now, made by a coaching staff that has counted what its players can survive, and much of the tactical argument the professional game is currently having comes down to whether the bill arrives before the shot clock does.