NBA defensive schemes explained: choosing what to give up

A defence cannot take away the layup and the corner three on the same possession. That is not a coaching failure, it is arithmetic. Five defenders, a floor fifty feet wide, and an offence entitled to stand four men behind a line that sits twenty-two feet from the basket in the corners and nearly twenty-four at the top. Most accounts of NBA defensive schemes explained as a menu of names, drop and switch and blitz and zone, get the grammar wrong. These are not techniques a team owns. Each one is a sentence that begins with the word unless.

We will guard everything, unless you take this.

The useful question about any scheme is never how it works. It is what it has agreed to let happen, who signed off on that, and what it costs when the opponent notices. A coverage that concedes nothing has never existed and could only exist if the offence agreed to stand closer together than the rules oblige it to.

The two-man version of that problem, one defender responsible for two attackers, is worked through in the ball screen and the man it always leaves open. This is the five-man version. Everything below is about how a team decides, in advance and in writing, which shot it is prepared to see go up, and what happens to that decision when a good opponent tests it forty times in a row.

The rim or the three is the choice every scheme is built on

Two shots on a basketball court are worth taking. A shot at the rim converts at a higher rate than anything else, and it drags fouls and offensive rebounds along with it. A three-pointer converts at a much lower rate and pays half as much again for each one that drops. Multiply each rate by what the shot is worth and the two land close together, which is the entire economic foundation of modern basketball.

Everything else on the floor is worth less. The pull-up from eighteen feet is the most contested shot in the sport relative to its payoff, which is precisely why a defence is happy to see it. A scheme is not really trying to stop shots. It is trying to move attempts from the two profitable categories into the unprofitable one, and it can only do that by leaving something.

• 22ftCorner three, distance from the basket
• 23.75ftTop of the arc, distance from the basket
• 16ftWidth of the lane
• 4ftRadius of the restricted area

Court dimensions as published in the rulebook, not statistics. The corner three sits nearly two feet closer to the basket than the shot at the top of the arc, which is why the corners are the first place an offence goes and the last place a rotation can reach.

Now put numbers on the trade, with the warning that these are invented and chosen to be round.

Layup at the rim, made 65% of the time1.3pts

Corner three, made 39% of the time1.17pts

Above-the-break three, made 36% of the time1.08pts

Pull-up two off the dribble, made 42% of the time0.84pts

An invented illustration. Each figure is a made-shot rate multiplied by the value of the shot. None of these rates is reported from any real player, team or season; they are chosen to show why two categories of shot are defended and one is offered.

Show the numbers

Worked example: what each shot pays per attempt

Item	Value
Layup at the rim, made 65% of the time	1.3pts
Corner three, made 39% of the time	1.17pts
Above-the-break three, made 36% of the time	1.08pts
Pull-up two off the dribble, made 42% of the time	0.84pts

Read the gap between the bottom row and the rest. That gap is what every defensive scheme is trying to sell. The pitch to the offence is always the same: take this one instead, we will not contest it very hard, and we will keep our shape while you do.

The trade is not symmetrical, though, and the asymmetries are where coaching judgement actually lives.

Rim attempts bring free throws with them. A defence that walls off the paint physically, with bodies arriving from the weak side at speed, will foul more, and fouls are the one category of concession that also removes players from the game. Three-pointers bring variance. A team that gives up a large number of open threes has widened the distribution of outcomes, which is a disaster over eighty-two games and occasionally a reasonable gamble in a single one, particularly if the alternative is losing by twenty every night to a better team playing its preferred game.

And a shot conceded to one man is not the same shot conceded to another. This is the point where the abstract argument turns into a scouting report. A corner three surrendered to a career non-shooter is nearly free. The identical geometry, the identical rotation, the identical closeout, against a specialist standing in the same spot, is the worst thing that can happen on the possession. Schemes do not have a fixed price. The price is set by whoever is standing in the space you decided to leave.

NBA defensive schemes explained by what they refuse to guard

A scheme is three separate sets of instructions that people usually collapse into one word.

The first is the on-ball coverage: what the two defenders nearest a ball screen or a handoff are supposed to do. This is the part that gets a name and gets talked about.

The second is the off-ball rules: how far the other three stand from their men, whether they sit in the passing lane or a step towards the ball, whether they stunt at a driver and recover or hold their ground, and above all who is designated to leave his man when help is needed. This is the part that decides the possession.

The third is the transition rules: who gets back, who chases the ball, who takes the rim, and how the defence sorts itself out when nobody starts the possession matched up. Nearly every published explanation stops before this one exists.

Underneath the naming, the meaningful axes are only two. How many defenders is this scheme willing to send at the ball? And how attached to their own men are the defenders who are not near it? A defence that sends two at the ball and keeps three attached has volunteered for a four-on-three. A defence that sends one and lets the other four sag towards the paint has built a wall and has left the arc. Every named coverage is a coordinate on those two axes, which is why the man-to-man against zone distinction matters much less than people expect. A modern man-to-man with off-ball defenders sitting in the gaps looks like a zone with names attached. A modern zone that follows cutters and switches on the perimeter is man-to-man with a starting shape.

The reason a team commits to a scheme at all is not that its answer is better than the alternatives. It is that a pre-agreed answer removes the decision. Five players moving at once cannot negotiate. If the low man has to work out whether to help, he is already late, and the reason his teammate behind him did not rotate is that he was waiting to see what the low man would do. The value of a scheme is mostly that everybody is wrong in the same direction, which is recoverable, rather than wrong in five different directions, which is not. The internal machinery of that agreement, who leaves whom and in what order, is the subject of how help defence is organised as a system.

Drop keeps the rim and hands over the shot in front of it

Drop is the most-used answer in professional basketball because it is the only one that asks nothing of the other three defenders.

The screener's defender retreats towards the basket as the screen arrives, staying between the ball and the rim. The on-ball defender fights over the top of the screen and tries to arrive at the shooter's hip before the shot goes up. Nobody helps, nobody rotates, nobody leaves a shooter, and when the possession ends the defence is still standing where it started, which also means it is in position to rebound. That last point is undersold. A scheme that never scrambles is a scheme that always boxes out.

What it concedes is the shot in the space between the two defenders, which is to say the pull-up from around where the screen was set.

The depth of the retreat is a dial rather than a switch, and the settings run roughly from level with the screen to the edge of the restricted area. Higher up, the ball handler's shot is contested and the roller has more room behind. Deeper, the rim is sealed and the shot is free. A coaching staff sets that dial per opponent and sometimes per player, which is why the same team can appear to be running two different schemes in the same quarter. It isn't. It is running one scheme at two depths. The variants, and the reason the deepest setting became indefensible against certain guards, are pulled apart in what drop coverage is actually protecting.

The reason drop went from default to controversial is not that anybody discovered a flaw in it. The flaw was always there and it used to be cheap. When the shot in front of a dropping big man was a contested eighteen-footer worth two points, letting it go up was close to a win. When the same geometry produces a shot from two feet behind the arc, the concession is worth around a third more per attempt, and a scheme that was designed around giving up the worst shot on the floor is now giving up one of the best.

Most teams did not abandon drop. They bolted things onto it. The commonest addition is a low man or a nail defender who slides into the driving lane to shrink the space the ball handler is heading towards, which converts the coverage from two defenders into two and a half and starts to cost something behind the ball. The moment that half-defender exists, the offence knows exactly which shooter to look for, and the scheme has quietly become a different scheme with the same name.

Switching keeps the shape and hands over the matchup

Switching is the only answer that leaves the defence's shape completely intact. The two defenders trade assignments as the screen arrives, the screen accomplishes nothing, and the three defenders away from the ball never move an inch. Three shooters stay guarded. Nobody closes out from anywhere. No rotation can be botched because no rotation is required.

The bill arrives later, and it arrives in the offence's chosen currency.

After a switch the offence has been handed something it does not normally have, which is the right to select the matchup. It exercises that right by screening until it produces the pairing it wants. The team-level consequence is the one that matters here: a switching scheme is a promise made by all five defenders at once, and it is only as strong as the least equipped of them. One player who cannot hold a drive for two seconds turns the whole scheme into a delivery mechanism for isolations against that player. This is why switching is a roster decision made in the summer rather than a tactical decision made in a timeout.

Most teams therefore switch conditionally rather than universally. The common conditions are worth knowing because you can see them from the stands. Switch one through four but not the centre. Switch only when the screen is set above the arc. Switch only between players within a certain size band. Switch only late in the shot clock, when the offence no longer has time to hunt. Each condition is an attempt to keep the shape without handing over the matchup, and each one creates a seam, because the offence's counter is simply to set the screen at the angle or with the personnel that falls outside the condition.

There are two costs people forget. The first is the offensive glass, since a switch routinely leaves a guard responsible for boxing out a much larger man in the one moment nobody is thinking about rebounding. The second is that switching burns the defence's willingness to help: if the scheme's whole promise is that nobody leaves, then when somebody eventually does leave, nobody behind him is expecting it. The full mechanics, including which pairings a defence is prepared to accept and how it repairs the ones it isn't, are set out in what a switching defence actually survives.

Blitzing takes the ball away and hands over the numbers behind it

Sending two defenders at the ball is the most emotionally satisfying thing a defence can do and the most expensive. The best player gives the ball up, the crowd reacts, and behind the trap there are now three defenders responsible for four attackers.

The scheme is not the trap. Anybody can trap. The scheme is the arrangement of the three men left over, and it is a zone whether the coaching staff uses that word or not. The standard shape is a triangle: one defender takes the rim, the other two split the three remaining attackers between them by playing the passing lanes and reading the passer's shoulders. The rotations behind it are pre-assigned, and the sequence when the ball moves is an exchange in which the man who helps at the rim recovers to the nearest shooter while a perimeter defender sprints to the man he left. Done cleanly, it looks like teleportation. Done a beat slow, it is an uncontested corner three, because the corner is the longest journey on the floor and the last stop on every rotation.

Three costs come with a blitz, and only one of them is the open shot.

Fouls, because trapping is a collision sport and the second defender arrives at speed into a player holding the ball. The offensive glass, because a scrambled defence cannot box out anybody by name. And fatigue, since a defence that traps is asking two players to sprint twenty feet and two more to cover three men on every possession, and the effects show up in the fourth quarter of the second night of a back-to-back rather than in the play you were watching.

Against that, blitzing buys the two things nothing else buys. It removes the ball from the hands of the best decision-maker on the floor and hands the possession to whoever catches it, which is a bet on the opponent's third and fourth options. And it manufactures live-ball turnovers, which are worth much more than they look, because a steal at half court is the beginning of the highest-value possession in the sport for the team that gets it.

That is the honest way to think about it. A blitz is not aggression, it is a purchase. You are buying turnovers and discomfort with open shots and defensive rebounds, and the exchange rate is set by the opponent's ability to pass out of pressure.

Zone hands over the second pass, and the rulebook keeps it short

Zone defence has been legal in the NBA since the illegal defence rules were removed before the 2001-02 season. It remains rare, and the reason is a single clause introduced in the same rule package.

Defensive three seconds forbids a defender from remaining in the sixteen-foot lane for more than three consecutive seconds unless he is actively guarding an opponent, which the rulebook defines in terms of being within arm's length and in a guarding posture. The count is interrupted when a shot goes up, when the defender picks up an assignment, or when he clears the lane entirely. The penalty is a technical foul: one free throw, and the offence keeps the ball.

Read that clause next to the definition of a zone and the problem is obvious. The single greatest advantage a zone offers is the ability to station your largest player permanently in front of the basket, so that every drive ends at him and every layup is contested by a man who never had to rotate. The rule removes exactly that. A centre at the back of a 2-3 has to be attached to somebody, or he has to keep stepping out and back in, and an offence that understands the clause can drag him out simply by standing a player at the edge of the lane and moving him.

FIBA has no equivalent provision, which is why zone is a viable primary scheme in international basketball and a change of picture in the NBA. The same shape, played under two rulebooks, is two different tactics.

The other problems are less legislated but just as real. A zone is a set of areas, and areas have edges, and shooters live on edges. Against a group that can pass the ball around the perimeter faster than five defenders can slide, the second and third pass will find a gap, which is the concession the scheme is making. Rebounding responsibility becomes vague, since nobody is matched up with anybody, and the offensive glass is the most reliable way to attack a zone at any level.

So it survives as a specialist tool rather than a base. Teams use it to break an opponent's rhythm after a run, to protect a player in foul trouble by hiding him in a position with less responsibility, to change the picture for two possessions after a timeout so the opponent has to talk instead of playing, and occasionally to hide a defender who cannot be trusted one-on-one. Match-up zone sits between the two, starting in a shape and then converting to man-to-man principles once the ball moves. The full anatomy of the shapes and their seams is covered in where a zone's gaps actually are.

Three rule changes wrote the menu of NBA defensive schemes

Scheme variety in the NBA exists because of the rulebook, not in spite of it. Three changes did most of the work, and none of them was aimed at defence directly.

The abolition of illegal defence. Before 2001-02, an elaborate set of restrictions governed where defenders could stand relative to men they were not guarding, with the effect that helping away from the ball was largely legislated out. The result was a league in which the safest offence was to clear one side of the floor and let your best player work one-on-one, because the rules guaranteed nobody could pre-position to stop him. Removing those restrictions was aimed at breaking up isolation basketball. What it actually did was hand defensive coaches the right to build shapes, which is the precondition for a scheme existing at all.

Defensive three seconds. Introduced in the same package as the safety valve, and much the more consequential half. It forced every defensive structure to be a moving structure. Without it, the answer to almost everything would be a seven-footer standing under the rim for twenty-four seconds, and the entire vocabulary of coverages would collapse into one.

The hand-check ban. Enforcement was tightened decisively for the 2004-05 season, removing a defender's ability to place a hand or forearm on a perimeter ball handler and steer him. This is the change that made schemes necessary rather than optional. When a good defender could physically control the man he was guarding, containment was an individual skill and the other four could stay home. Once it became illegal, containment became a five-man problem, and a five-man problem needs a pre-agreed answer. Which is the definition of a scheme.

Add the restricted-area arc, which stopped a secondary defender from standing under the rim and collecting a charge, and the picture is complete: the rules removed every cheap individual solution to rim protection, and then forbade the cheap collective one. What is left is a set of expensive answers, all of which concede something, which is precisely why there are several of them rather than one correct one.

Any rule that made defending easier would shrink the menu, not expand it. Variety is what a sport looks like when nothing works well enough to dominate. The broader consequence, the way the floor stretched and the rotations lengthened once the arc became the main event, is traced in what the three-point line did to the shape of the game.

A scheme is only as good as the slowest man it exposes

Coaching staffs do not choose schemes they admire. They choose schemes their players can physically execute, and the menu available to any given five is much shorter than the list of options in a clinic.

Five abilities gate it, and each one closes a door when it is missing.

Containment at the point of attack. If a guard cannot stay attached over a screen, everything downstream has to compensate, permanently. The scheme is paying for one player's half second on every possession of the night.

Rim protection. A defence without it cannot drop, because dropping is a promise that the rim will be defended when the ball gets there.

Perimeter mobility from big men. A defence whose centre cannot survive four seconds against a guard cannot switch, which means it must help, which means somebody is open by design.

Closeout speed and length across the wings. Every scheme that helps is a scheme that has to recover, and recovery is a race between one man's legs and one pass.

Discipline under fatigue. This is the least glamorous and the one that decides most games. A scheme is a set of promises kept forty-eight times in a row. The fourth time a tired player guesses instead of executing, the scheme is no longer being run.

The offence's job is simply to find whichever of the five is thinnest and test it repeatedly, which is not a sophisticated plan, merely an available one. And because the test is repeatable, the weakness compounds. A defender who loses one possession in four does not cost his team one possession in four, he costs them every possession that gets aimed at him, and the aiming is free.

There is a roster consequence that follows and it is worth stating plainly, because it explains a lot of transactions. Most teams need two schemes, not one. The group with the starting centre on the floor can drop. The group without him cannot, so it has to switch, which requires bodies that can hold up on the perimeter. Coaching staffs that have not prepared a second answer end up running their first answer with personnel who cannot execute it, which is the most common way a good defensive plan is destroyed without anybody adjusting anything.

Foul trouble is the same problem arriving suddenly. A scheme calibrated around one rim protector becomes a different scheme in the eight minutes he is on the bench with four fouls, and an opponent who is paying attention will spend those minutes attacking the seam that has just opened.

Matchup hunting, and the scram switch that answers it

Hunting is the general-purpose attack on any scheme that concedes matchups, and it has become organised enough to have its own countermeasures.

The mechanics are simple. Identify the defender you want, then set screens involving the man he is guarding until the defence trades him onto somebody dangerous. If the defence refuses to switch, the screen itself has created an advantage. If it does switch, the hunt has succeeded. Either answer is acceptable to the offence, which is what makes the tactic so hard to refuse.

The refinements are where scouting shows up. Hunt in the last eight seconds of the shot clock, when the defence has no time to repair anything. Hunt immediately after a made basket, before the defence has finished matching up. Involve the target as the screener rather than as the ball handler's defender, since a screener's defender is the one the coverage forces to move. Put the target's man on the inbounder, which strands him behind the play. None of these requires a new action. They require a note on a whiteboard.

The defence has four counters and only one of them is a scheme change.

It can hide the vulnerable defender on whichever opponent is least involved, though a well-coached offence will simply involve that opponent. It can pre-switch, trading assignments before the screen arrives so that the screen produces the pairing the defence wanted. It can top-lock, denying the intended screener the route to set the screen at all. And it can scram switch, which is the most elegant of the four: once the mismatch exists, a third defender, usually the one guarding a non-shooter near the basket, sprints in to trade assignments with the exposed teammate before the ball ever gets there. The mismatch is repaired without anybody double teaming, and the possession restarts with no advantage on the floor.

The scram is why a team can switch everything and still hardly ever be posted up. It also has a cost, which the following possession makes visible.

1. 1The scheme is declared before the ball crosses half courtThis defence switches every screen between its four perimeter players and keeps the centre attached to the man he started on. Nobody will help, so nobody will be open. The concession has already been made in advance: the offence may choose its own matchup.
2. 2The offence sets a screen it does not care aboutA wing screens for the ball handler on the far side of the floor. The defence switches, as advertised. The purpose is not to score, it is to confirm that the switch is automatic rather than conditional.
3. 3The real screen involves the targetThe second screen is set by the man being guarded by the defence's weakest perimeter defender. The switch puts him onto the ball handler, and the offence has the pairing it came for with fourteen seconds still on the clock.
4. 4Three attackers empty the strong sideTwo go to the corners on the far side, one stands above the break. There is now no defender within fifteen feet of the drive who can help without leaving a shooter behind the arc.
5. 5The defence goes for the scramThe centre, who is guarding a non-shooter standing beside the basket, sprints out to trade assignments with the exposed defender. Completed in time, this repairs the mismatch and the scheme survives without a single rotation.
6. 6The offence attacks before the trade finishesThe ball handler drives on the first dribble rather than surveying the floor. A scram switch needs roughly a second and a half of communication and movement. The drive does not.
7. 7The centre is now between two menHe has left his own man to make the trade and has not yet arrived at the ball handler. For one beat he is guarding neither, which is the specific moment the offence was aiming at.
8. 8The lob goes to the man the centre leftThe non-shooter beside the basket has been unguarded since the scram began, and the pass is thrown above the rim where no recovering defender can contest it.
9. 9The scheme did not break, it paidThe switch was correct, the scram was correct, the sprint was as fast as a person can move. The defence conceded a matchup by design and then conceded the repair, which is the second bill on the same choice.

A constructed possession described step by step, not a report of any real play. Every defensive decision in it is individually correct.

Notice what that possession did not contain: a mistake. Every decision followed the scheme. The offence did not beat the defenders, it beat the agreement they were operating under, which is the difference between an execution failure and a design cost. Coaching staffs that cannot tell the two apart spend timeouts shouting about effort when the problem is arithmetic.

Transition defence is the scheme nobody bothers to name

Ask most people to describe a team's defence and they will describe what it does once everybody is standing in position. That is the smaller half of the job, and it is not where the cheap points come from.

Transition defence gets neglected for three structural reasons. It happens at the end of the offensive possession, so it belongs to nobody in the way a half-court coverage belongs to the defence. It is difficult to see, because the interesting part is a decision made twenty feet behind the ball by a player the camera has left behind. And it is largely decided before it starts, by two choices that are not defensive choices at all.

The first is shot selection. A long shot from a poor position produces a long rebound and a defence facing the wrong way, which is the most reliable way to hand over an easy basket without anybody being beaten.

The second is offensive rebounding policy. Every player sent to the glass is a player not getting back, and the trade is quantitative rather than moral. Send nobody and the possession ends at the first miss. Send everybody and the opponent's first outlet pass is a numbers advantage the half-court scheme never gets to answer. Most teams settle on sending one or two by rule and letting the rest retreat, and the rule is set by the opponent rather than by preference. Against a team that runs hard on every miss, the crashers stay home. Against a slow-starting team, the glass is nearly free. The offensive side of the same equation is worked out in how a team turns a defensive rebound into points.

The rules for the retreat itself are almost universal and almost never explained.

20%20%60%

• Stopping the ball1
• Taking the rim1
• Finding the other three attackers3

A structural allocation, not a measurement. Nearly every transition rule in the sport divides the five defenders this way before it says anything else about who guards whom.

Show the numbers

The three jobs the instant a shot goes up

Item	Value
Stopping the ball	1
Taking the rim	1
Finding the other three attackers	3

The first man back does not look for the player he was guarding. He takes the rim, because an uncontested layup is the worst outcome available and it is the one that arrives soonest. The second takes the ball, and his job is to slow it rather than steal it. The remaining three match up by proximity rather than by name, which is a deliberate instruction: find a body, take it, sort out the assignments once the ball has stopped.

That instruction produces cross-matches, which is the part that connects transition back to the scheme. A defence that gets back well but matches up badly ends the sequence with its centre on a guard and a guard on its opponent's best post player, and now the half-court coverage it wanted to run is unavailable. The cost of a bad transition possession is therefore not only the shot at the end of it. It is the next fifteen seconds, spent playing a scheme the team did not choose. Teams that fix this drill an explicit swap sequence for the first dead ball, and the good ones execute it without a word.

Live-ball turnovers are the same problem in its worst form, since the defence is not merely retreating, it is retreating from a standing start with nobody at the rim. Which is why a scheme that manufactures turnovers is not simply trading shots for steals. It is trading its own shots for the opponent's worst possessions, and the exchange rate is set by how quickly its own players recognise that the ball is gone.

How a coach changes a scheme inside a seven-game series

Regular-season defence is played against a schedule. Playoff defence is played against one opponent who is watching the same film you are, which changes what an adjustment is.

The most common misunderstanding is that adjustments are inventions. They almost never are. A staff has three or four coverages its players can execute in April, and no new one is being installed between games. What changes is the price list: which action gets which coverage, which player gets a second defender, which shooter is allowed to be left.

The ladder of adjustment usually runs in this order, because each rung costs more than the one before it.

Change the coverage on one action. The opponent's most-run set gets a different answer, and everything else stays as it was. This is cheap because it requires nothing new to be learnt, only a trigger to be remembered.

Change the treatment of one player. Blitz him specifically, or refuse to switch onto him, or take away his preferred direction and live with the other one. This costs more, because the rest of the defence has to rotate behind a rule that only applies sometimes.

Change who is on the floor. The most expensive and the most effective, because personnel is what the menu was limited by originally. A different centre makes a different set of coverages legal.

Change the base. Rare, late and usually a sign the first three did not work.

Two details separate staffs that adjust well from staffs that react. The first is timing: the right moment to change a coverage is before it visibly fails, while the opponent is still shooting badly against something that is about to stop working. Waiting for the evidence means changing while behind, which is when players trust the scheme least. The second is which player to aim at, and the counter-intuitive answer is often the second-best one. Taking away the best player's shots is expensive and usually fails; taking away the supporting cast's easy ones removes the possessions that made the star's difficulty tolerable.

Then there is the adjustment nobody sees, which is changing coverage inside a single possession. Start in drop on the first screen, blitz the second, switch the third. The point is not that any of the three is better. It is that the offence's read is made in half a second based on what it saw last time, so a defence that shows a different picture on the re-screen is attacking the read rather than the action. Doing this requires the very thing the personnel section listed last: five players who can execute three answers under fatigue without talking about it.

The counter-cycle is inevitable. Every adjustment creates a new concession, the opponent finds it by the following game, and the staff that wins the series is generally the one whose roster supports the most answers rather than the one whose whiteboard is cleverest.

Name a team's base coverage in one possession

The genuinely useful skill here is not memorising definitions. It is being able to watch a single possession, without commentary, and say what the defence has decided to give up. It takes four observations and only the last one involves the ball.

Watch the screener's defender's feet at the moment of contact. Retreating towards the rim is drop. Sliding across onto the ball handler's outside hip is a switch. Stepping above the screen with his hands up is a hedge or a blitz, and which one it is depends on whether he goes back. Sitting near the baseline while the ball is pushed down the sideline is ice. That single glance identifies the on-ball coverage more reliably than anything else available.

Watch the defender guarding the weak-side corner. How far off his man is he standing? Level with him means the scheme has promised not to help. Two steps towards the paint means it has already decided which shooter it is willing to leave, and that is the shot the possession is likely to produce.

Watch what happens on the re-screen. This is the observation that separates a base coverage from a one-off. If the same answer appears on the second and third screen of the possession, you are looking at the team's actual scheme. If the answer changes, you are looking at a staff that is attacking the ball handler's reads, which tells you something about both teams.

Then look at where the shot came from. A pull-up around the level of the screen means a drop did its job. A shot at the rim by the screener means nobody was assigned to stop him or the man assigned was late. A corner three means somebody helped and the rotation lost the race. A contested jump shot with four seconds left, taken by a player who was not involved in the original action, means the defence conceded the matchup and then made the offence spend the whole clock collecting it.

Do that for one half and the vocabulary stops mattering. What you are watching is a negotiation in which one side asks the same question forty times and the other side has written its answers down in advance, knowing that every answer leaves something. The best defences in the professional game are not the ones that concede least. They are the ones that concede the cheapest thing available against this particular opponent, and that keep doing it when they are tired.