You now know how the opening works and how a handful of pieces mate a lone king. This chapter bridges the two: how do you get from thirty-two pieces on the board to one of those clean endgames? By constantly "seeing" each other — and everything that follows in this chapter is built on that one idea. Every piece, on every turn, sees exactly three kinds of things from its square: its own teammates, empty squares, and enemy pieces. Get comfortable counting what a piece sees, and the rest falls into place.
Swap the queen for a rook or a bishop and you get a fun bit of arithmetic. A rook on d4 sees 14 squares — a whole rank plus a whole file — which is part of why a rook outranks a bishop or knight in material value. A bishop on the same square sees 13, spread across its two diagonals.
Fun fact: 14 + 13 = 27. Add up the rook's vision and the bishop's vision from the same square and you get exactly the queen's — which is exactly what you'd expect, since the queen moves like a rook and a bishop combined.
Nobody plays chess with one piece on an empty board, so let's put some traffic in the way. In the position below, White's queen still sits on d4, but now other pieces occupy some of those 27 squares — and can block the queen's path to others. Try to work out what the queen can and can't actually win before you press Show solution.
This is where "seeing" and "attacking" split apart, and the difference matters more than almost anything else in this chapter. The queen sees the pawn on f6 and the bishop on d7 alike — both are pieces it could legally capture. But only one of those is a genuine attack. Nothing defends the f6 pawn, so if White captures it, White simply keeps the extra pawn. The bishop on d7, on the other hand, is defended by Black's own king on c8: capture it, and the king recaptures the queen right back. Seeing a piece means you could physically move there. Attacking a piece means capturing it would actually win you something. Those are not the same claim, and beginners who confuse them give away their queen constantly.
Flip the board around and look at the same position through Black's eyes, and a much bigger problem jumps out. Try to spot White's best answer before revealing it.
Even though White's king could immediately recapture the knight if it took the queen, that trade is a disaster for White — a queen is worth roughly three knights. So it's White's move, and White must deal with this threat first. Conveniently, the move that gets the queen to safety also wins material: queen takes the pawn on f6, escaping the knight's attack and grabbing a free pawn in the same move. That's the whole game in miniature — every piece is simultaneously offering threats and dodging them, and your job is to spot both directions at once.
Once you can tell a real attack from a piece that's merely "seen," the next question is what to do when one of your pieces is the one under fire. There are exactly four options, and good players choose between them based on which one leaves them best off — not just safest.
The simplest defense. In the position below, White has no captures, no checks, and nothing obviously threatening — so instead, White looks to create a threat of its own. Try to find it before pressing Show solution.
Moving out of danger is always available, and it's the right choice whenever you can land on a square that also improves your position — capturing something else along the way, or aiming at a new target — rather than just a random safe square.
Suppose instead White had attacked the same queen from c5 rather than f4:
Capturing the attacking piece works whenever the resulting trade is even or good for you — trading equal material, or better yet, trading a lesser piece of yours for a more valuable enemy piece.
The remaining two defenses show up together in this next example. There's no capture or check on offer, but one plan creates a genuine attack on a Black minor piece — see if you can spot it, then step through the trunk for one way Black protects it, or switch to the Protect with a rook: ...Rac8 line to see the alternative.
Protecting the attacked piece works well when the piece you bring in to guard it also ends up better placed than before — you're not just surviving, you're improving.
The fourth and final defense is blocking: putting a piece between the attacker and the target so the line of attack no longer reaches. This shows up constantly with checks, since a king in check has no other legal way to survive but to move, capture the checker, or block. Here's an early position from the Sicilian Defense, only three moves into the game, where Black's king is in check from White's bishop on b5. Black can't move the king to safety, can't capture the bishop, and can't guard the king with another piece — see if you can find the one blocking piece that does more than just block.
Black actually had four different pieces that could interpose here — the queen to d7, or the knight to either c6 or d7, or the bishop to d7 as shown. Putting the queen there would have been a catastrophic blunder, since the bishop would simply capture it. The bishop move is the best of the bunch precisely because it's the most forcing: it blocks the check and counterattacks White's bishop, so White now has to spend a move dealing with that threat too.
Here's a second example of blocking, this time defending a queen rather than a king. White would like to create an attack on Black's queen. It's White's move — try to find how before revealing the answer.
Notice the pattern in both blocking examples: the strongest block is one that fights back, not merely one that stands in the way. But blocking has a dark side too — the piece you use to block can end up pinned to your own king or queen, unable to move again without disaster. Check what you're exposing before you commit to a block.
Before moving on to tactics, it's worth returning to the seeing-versus-attacking distinction, because there's a version of it that trips up nearly every beginner. Take this position:
Seeing a piece is not a threat by itself — you have to ask what happens after the capture. Now here's the trickier case: what happens when you attack an enemy piece of your own type? Levy Rozman writes that this is something his beginner students got wrong all the time — tempted by a move in the very same position above that looks like a strong threat on Black's rook. Same starting position, same board — but try to spot what's wrong with the natural-looking move before pressing Show solution.
You can never attack a piece of your own type — two rooks facing each other, two queens, two bishops on the same diagonal — because whichever one moves onto the other's line hands the opponent a free capture on their turn. Never make a move hoping your opponent "won't notice." Play only moves that work even against a perfect opponent.
With attacking and defending down, you're ready for tactics: forcing sequences that win material by combining threats faster than the opponent can answer them. Some tactics win a whole piece outright; others trade a cheaper piece of yours for a pricier one of the opponent's. Either way, tactics are what actually convert an opening into a winning endgame — without them, careful players trade down evenly and the game just fizzles into a draw.
The most fun tactic to learn is the fork: one piece attacking two or more enemy pieces at the same time. The opponent can only rescue one of them, so a genuine fork wins material by force.
Forks arrive with check especially often, since the king has to be dealt with before anything else — buying the forking piece a free turn to grab its other target. But knights are the deadliest forkers of all, check or no check, for a structural reason: nothing else on the board moves the way a knight does. When a knight attacks a queen, rook, or king, that piece can never attack the knight back, the way a rook attacks a rook or a bishop attacks a bishop on a shared line. The knight is simply untouchable from the squares it threatens. Here's White to move — try to find the devastating fork before you reveal it.
When a knight fork catches the king plus two or more other pieces, it earns the nickname family fork — the whole family under attack at once, and there's rarely anything the defender can do about it.
Not every fork wins material, though, and it's worth seeing why. Here White can play a move that attacks two Black pieces at once — can you find it, and then Black's way out?
The lesson here is simple but easy to forget mid-game: whenever you attack a piece, ask yourself where it's actually going to run — because a well-timed retreat can leave the opponent no worse off, or even improved. An even more extreme version of the same trap is thinking any multi-piece attack must be a fork at all:
Counting targets isn't the same as counting winnings. Before you get excited about a fork, ask of every piece it hits: is this one actually undefended, or would grabbing it just hand my piece away for free?
A pin immobilizes an enemy piece because moving it would expose a more valuable piece standing behind it on the same rank, file, or diagonal. There are two flavors. A relative pin lets the pinned piece move legally, but moving it loses material.
An absolute pin is the stricter version: moving the pinned piece is flat-out illegal, because the king itself sits behind it.
Both kinds of pin are dangerous for the same reason: the pinned piece can't move to defend itself, so it's an easy target to pile onto. That's the idea behind "pressuring the pinned piece" — bring a second attacker to bear on it, and the defender is stuck choosing between losing it outright or throwing good material after bad trying to save it. Here it's Black's move — see if you can spot the follow-up before revealing it.
That single move is devastating. White's other knight can't recapture on d4, because doing so would open the bishop's line straight onto the queen; meanwhile the new knight on d4 is directly attacking the queen too. White's queen has to run, and Black then simply captures the pinned knight on f3 with either minor piece — material won, purely from applying pressure to a piece that already couldn't move safely.
Pins cut both ways. If you block a check or an attack with one of your own pieces, double-check you haven't accidentally pinned that piece to your own king or queen — you may have solved one problem by creating a new one.
Bishops aren't the only pinners — rooks and queens do it just as well, usually against a king:
A skewer is a pin turned inside out: instead of a weak piece shielding a strong one, you attack the strong piece first, and when it's forced to move, you win whatever was standing behind it. Skewers hit kings especially often, since a king in check has to move no matter what it's leaving behind. Black has just left the king and rook lined up on the same diagonal — it's White's move, so try to find the check that punishes it.
Even when you're on the receiving end of a skewer, which square you retreat to still matters enormously. In the next position, Black is winning — a queen against a bare rook — but has carelessly left the king and queen on the same file. It's White's move again — find the skewer before revealing it.
That's the general shape of defending against a skewer: you usually can't save the piece behind the check, but you can often make sure you at least win something back for it.
The last tactic ties the whole chapter together, because it depends entirely on visualizing what your own pieces will suddenly see once something moves out of their way. A discovered attack happens when you move one of your pieces aside, revealing an attack from a different piece that had been blocked behind it — often while the moving piece creates a threat of its own at the very same time. It's White's move here — how would you attack Black's rook on f8?
Since a queen outranks a rook, Black should save the queen and let the bishop take the rook on f8 — this only works, though, because the rook giving the discovered attack was itself defended; otherwise Black's queen would simply take it for free.
When the revealed attack lands on the enemy king rather than a piece, you get a discovered check — arguably the single most dangerous tactic in the book, because the piece doing the moving is completely free to do anything it likes, anywhere on the board, while the opponent is still forced to deal with the check first.
Of the knight's five possible moves in that position, two of them — b5 or e6 — give discovered check while also hitting the queen; the other three would just be ordinary knight moves. That's the habit worth building: whenever you're about to move a piece, ask not only where it's going, but what it's uncovering behind it. As Levy Rozman puts it, always look for opportunities to set these up — and just as importantly, make sure your own king isn't sitting on a diagonal or file where your opponent could spring the same trick on you.
| Tactic | What it does | Watch for |
|---|---|---|
| Fork | One piece attacks two+ targets at once | Are the targets actually undefended? |
| Pin | A weaker piece can't move without exposing a stronger one behind it | Relative (legal, costly) vs. absolute (illegal, king behind it) |
| Skewer | A stronger piece is forced to move, exposing a weaker one behind it | Which square lets you recapture something back? |
| Discovered attack/check | Moving one piece reveals an attack from another | What does the moving piece also threaten? |
Every piece sees friendly pieces, enemy pieces, and empty squares, most of all from central squares — but seeing a piece isn't attacking it: an attack exists only if capturing would gain material once recaptures are counted, and you can never truly attack an enemy piece of your own type, since it sees you right back. When one of your pieces is genuinely attacked, you have four defenses: move it somewhere useful, capture the attacker if the trade favors you, protect it with a piece that also improves, or block with a piece that attacks back — watching for blocks and pins that end up working against you instead.
Tactics turn these ideas into forcing, material-winning combinations. Forks hit two or more pieces at once — knights are the natural specialists, since nothing can attack a knight back the way it attacks them, though a forked piece can sometimes still escape to a good square. Pins trap a lesser piece in front of a more valuable one, whether relative (costly but legal to break) or absolute (illegal, because the king is behind it), and pressure on a pinned piece snowballs fast. Skewers are pins in reverse: force the valuable piece to move first, then scoop up whatever it left exposed. Discovered attacks — including the especially vicious discovered check — spring from moving a piece to reveal a hidden threat, often creating two problems at once. Practice spotting all of these against an engine or a friend until it becomes automatic; that instinct is what separates beginners from everyone above them.
A piece is only genuinely attacked if capturing it would gain material once you account for recaptures. If the target is defended by something that would cost you more than you'd win, you can see it, but there's no real threat.
The two rooks see each other equally. If you move your rook onto that line, it becomes your opponent's move, and they simply capture your rook for nothing in return — you can't attack a piece of your own type this way.
Move it (best when you land somewhere that also creates a threat), capture the attacker (best when the trade is even or favorable), protect it with another piece (best when that piece also gets more active), or block the attack (strongest when your blocking piece attacks back).
Both protect the bishop on c5, but ...Rac8 does it with a rook instead of a pawn, without changing the pawn structure — and it activates a rook that was previously doing very little on a8.
Nothing else moves the way a knight does, so a piece it attacks — a queen, rook, or king — can never attack the knight back on that same move. Other pieces that attack each other (two rooks, two bishops on a diagonal) see each other in return; a knight's targets don't see it at all.
Because the rook can simply move away — to c8 — where it's no longer attacked and even ends up supporting its own pawn's advance. A fork only wins material if the opponent truly can't save both pieces; here, moving one to safety is enough.
In a relative pin, moving the pinned piece is legal but loses material, because a more valuable piece sits behind it. In an absolute pin, moving the pinned piece is illegal outright, because the piece behind it is the king.
Add more pressure to the pinned knight by bringing another attacker to bear on it — as in Board 20's ...Nd4 — forcing the defender to either lose the knight outright or spend more material trying to save it.
A pin attacks a lesser piece standing in front of a more valuable one, trapping the weaker piece in place. A skewer attacks the more valuable piece first; once it's forced to move away, you capture the lesser piece that was standing behind it.
Because those are the only squares from which the king can recapture White's rook after it takes the queen on e7. Moving anywhere else means Black loses the queen for absolutely nothing in return.
It's a discovered attack where the revealed line hits the enemy king. It's especially dangerous because the piece that moves to create it is completely free to do anything useful elsewhere — capture something, attack something else — since the opponent must respond to the check regardless.
Bc5 creates two threats on one move: the bishop itself attacks the rook on f8, while stepping off the e-file uncovers the other rook's attack on the queen. Black only gets one move, and it can only answer one of the two threats.