Chapter Seven introduced tactics as one- or two-move opportunities to win material or force checkmate — mostly things you could spot by looking one or two moves ahead. As an intermediate player facing sharper opponents, the goal shifts: tactics stop being something you calculate from scratch every time and start being something you recognize. Nobody is born able to spot a mate in two any more than a baby is born able to walk. Strong players aren't thinking harder in the moment than you are — they've simply seen thousands of similar positions before, so a winning shape jumps out instantly instead of requiring fresh analysis every time.
Rozman compares it to his own boxing training: he still flinches every time a punch comes near him, even in slow-motion drills, because he hasn't drilled the reaction enough times for his body to stop reacting to it consciously. The fix in boxing and in chess is the same — repetition until the response becomes automatic. You won't be born with tactical pattern recognition, and you won't get it by reading about it either; it comes from solving the same handful of recurring shapes over and over until your brain flags them without deliberate effort.
The only real way to build this pattern library is repetition: solving tactics puzzles against a real board or online, ideally by the thousands, so your brain starts flagging familiar shapes automatically instead of working them out fresh every time.
The single most useful habit for finding tactics — and for playing sound chess generally — is the CCA checklist: on every move, scan for checks, captures, and attacks, for both yourself and your opponent. These are the forcing moves. A forcing move narrows your opponent's replies from dozens down to one or two, which is exactly what makes a position calculable. A quiet move might allow thirty replies you can't hope to work through; a check might allow only one.
When several checks are available, don't just play the flashiest one — evaluate the check that leaves your opponent the fewest replies first, since it's usually the easiest to calculate to the end. Here's a composed position that rewards exactly this kind of scan. Try to find it before you press Show solution.
CCA also flags moves that look automatic but aren't the best choice. If one of your pieces comes under attack, your first instinct might be to save it — but check the CCA list before you do. Sometimes a check elsewhere is worth more than the piece under threat, an idea worth calling danger levels: ignore a small threat if you have a bigger one of your own. Try to find it before you press Show solution.
The same checklist scales up to longer forced sequences. When your opponent is one move from mating you, every one of your own moves has to be a check (or otherwise deal with the threat) until you either escape or turn the tables. Below, Black threatens mate on the very next move (...Qg2 or ...Qh1), so every White reply has to be a check. Try to find the forced mating sequence before you press Show solution.
"Mate in three" means there's a sequence where, no matter what the opponent does, mate follows in exactly three moves. Verify each of your opponent's replies really is forced — if they have a second legal option you haven't checked, your "forced mate" might not be forced at all.
Don't mistake any of this for an instruction to chase checks whenever you can. The point of drilling forcing sequences isn't to force mate on every move you get to play — it's to sharpen the same pattern recognition from Section 1, and that skill cuts both ways. The habit of scanning for checks, captures, and attacks is just as valuable defensively: it's what lets you notice the danger building against your own king a move or two before your opponent finds it.
Most positions won't end in immediate checkmate — CCA is just as useful for finding the move that wins material and leaves you winning later. Run the checklist in order: checks first (and reject the ones that don't work), then captures, then attacks — and don't stop calculating just because a trade looks fair. Try to find it before you press Show solution.
White's only check here, Bxf7+, just loses the bishop for a pawn, so it's rejected out of hand. Working down the CCA list to captures is what turns up Bxc6 and, a move later, the fork. None of this wins the game outright, but it wins a full piece for nothing — a decisive material edge to nurse into the endgame. This is exactly why CCA matters beyond mating attacks: it's a discipline for finding the best move in any position, not just sharp ones.
Removing the defender means attacking — or capturing — whatever piece is standing between you and a bigger prize, whether that's a mating square or a piece you want to win. If a checkmate or a winning capture is one defender away from working, look for a forcing move that eliminates that defender instead of abandoning the idea. Try to find it before you press Show solution.
Don't abandon a good idea just because one defender is in the way. Ask whether you can attack that defender first — with a check, a pawn push, or a second attacker — before giving up on the combination entirely.
A zwischenzug ("in-between move") is a forcing move you insert before playing the move that looked automatic. It has to demand a response — usually a check, or an attack on a queen or rook — so your opponent can't just ignore it and carry out their own plan.
Below, Black has just captured a center pawn with ...Nxd4, and it looks like White simply recaptures. Try to find White's best move before you press Show solution.
A zwischenzug doesn't have to be check — an attack on a queen or another major piece works just as well, since the opponent still can't stop to deal with anything else. Try to find it before you press Show solution.
A discovered attack happens when moving one of your pieces uncovers an attack from a piece behind it. The strongest versions create two threats simultaneously: the piece that moves makes its own threat, while the piece it unmasks makes a second, unrelated one — and the opponent usually can't answer both. Try to find it before you press Show solution.
The most dangerous discoveries combine a check with an attack on the opponent's queen, since the queen can't be defended by simply interposing on the checking line. This trap comes from the French Defense, Milner-Barry Gambit. Try to find it before you press Show solution.
A sacrifice is a deliberate investment of material for an attack, initiative, or a forced win later — think of it as a middlegame cousin of the opening gambits from Chapter Twelve. The pay-off doesn't have to be immediate, but it should be calculated, not hopeful. This position is from a Sicilian Defense: Black has just grabbed a bishop with ...Nxd3, and every recapture looks equally safe. Try to find White's best move before you press Show solution.
Rozman likes to plant exactly this kind of trap for readers: three different pieces can recapture on d3, and every one of them looks completely safe. If your hand was already reaching for one of them — he'd say he caught you. An even trade is exactly the moment CCA matters most, because "obviously fine" is precisely when nobody bothers to check for a check first. Run the checklist before every capture, not just the ones that already look complicated.
Deflection forces a defending piece away from a square or a piece it was protecting, usually with a bigger threat it can't decline. Try to find it before you press Show solution.
Attraction is the mirror image: instead of removing a defender, you lure — usually sacrifice — a piece onto a specific square where it becomes part of your mating net. Try to find it before you press Show solution.
CCA — checks, captures, attacks — is the discipline that turns tactical calculation from guesswork into a repeatable process: scan forcing moves first, for both sides, in that order, because forcing moves are what make a position calculable at all. This applies whether you're hunting a forced mate or simply looking for the best practical move in an equal, quiet-looking position — and the same scanning habit protects your own king just as often as it attacks your opponent's.
| Device | What it does |
|---|---|
| Removing the defender | Attacks or captures the piece guarding a square or another piece, clearing the way for the idea behind it. |
| Zwischenzug | Inserts a forcing move before the move that looked automatic, so the opponent can't get to their own plan first. |
| Discovered attack | Moving one piece uncovers an attack from a piece behind it — strongest when it's also check. |
| Sacrifice | Gives up material on purpose for an attack, initiative, or a calculated follow-up. |
| Deflection | Forces a defending piece away from what it was protecting. |
| Attraction | Lures (often sacrifices to force) an enemy piece onto a square where it becomes a liability. |
Sacrifices deserve one more comparison: think of them as the gambits of the middlegame. Chapter Twelve's opening gambits gave up a pawn for development and initiative before either side had castled; a middlegame sacrifice — a knight (three points) for a single pawn, say — makes the same kind of bet later in the game, cashing in material for a shredded enemy king and an attack that may or may not convert. That uncertainty, Rozman says, is part of the beauty of chess: a sound sacrifice is a calculated risk, not a guarantee.
None of this replaces calculation, and none of it substitutes for practice. Pattern recognition — the whole foundation of this chapter — only comes from solving puzzles, and Rozman is specific about the dose: not 10, not 50, not 100, but 1,000 or more before these shapes start feeling automatic at the board.
Checks, captures, attacks — the categories of forcing moves to scan for on every move, for both yourself and your opponent. The order is a practical habit: checks are the most restrictive (often leaving only one legal reply), so evaluating them first narrows the position fastest before you spend time on quieter options.
A forcing move — a check, a capture, or a serious attack on a valuable piece — sharply limits the opponent's legal or sensible replies, sometimes to just one. A quiet move might allow thirty different responses, far too many to calculate through to a clear conclusion.
It's attacking or capturing the piece that stands between you and a winning idea — a mating square or material you want to win. Find it by asking: if that one defender weren't there, what would I be threatening? Then look for a forcing move (often a pawn push or a capture) that removes or distracts it.
It has to be forcing — typically a check or an attack on a queen or rook — so the opponent can't simply ignore it and play the move they were planning. An in-between move that doesn't demand a response just wastes a tempo instead of gaining one.
It's when moving one piece reveals an attack from a piece that was behind it, so the moving piece and the revealed piece can threaten two different things at once. When the revealed threat is check, the opponent must deal with their king before doing anything else — including saving a queen that's also suddenly attacked, since a queen usually can't be defended just by blocking a check on a different line.
A sacrifice is a deliberate, calculated investment of material made because it buys something concretely worth more — a forced mate, a decisive attack, or a won position — not a hope that something might turn up. The difference is the calculation behind it, not the fact that material is given up.
Deflection forces a piece away from a square or piece it was defending, usually by threatening it with something bigger it can't ignore. Attraction does the opposite job with a similar method: it lures (often sacrifices material to force) an enemy piece onto a specific square, where it then becomes a liability in your mating net.
Every candidate move needs to either remove the mating threat directly or answer it with a check or bigger threat of your own — quiet developing or consolidating moves aren't safe options here. Check each of your available checks and captures in turn, since one of them may force a favorable sequence rather than just surviving passively.
Black is one move from mate, so every White move has to be a check or it loses on the spot. Rxa8 isn't even a check — it just hands away the rook for nothing while Black goes ahead and delivers mate. It fails the very first CCA filter, so there's no need to calculate it any further.
If White recaptures on d4 immediately, Black answers ...Bxb5 and the trade is even. Bxd7+ is check, so Black is forced to recapture the bishop before doing anything else — including saving the bishop on b5. Only once that in-between move has been played does cxd4 win the knight for free.
Any of the three recaptures accepts an even trade of a knight for a bishop and ends the tactic there. Running the CCA checklist first reveals Nxe6+, a check that also uncovers an attack on Black's queen — sacrificing a knight for a pawn, but winning the queen outright a move later. Snap-recapturing costs White the entire point of the position.