A chess clock is two timers in one housing, wired so that only one of them can run at a time. When it is your turn your side counts down; you make your move on the board, then press the button on your side with the same hand you moved with, which stops your timer and starts your opponent's. If your time runs out before your opponent's does, you normally lose — but not always, and the exceptions are worth knowing before you sit down.

This guide covers analog and digital clock mechanics, the handling rules an arbiter will enforce, what happens when a flag falls, Fischer increment versus Bronstein delay, the FIDE time-control boundaries, and how to spend your time once you understand all of it.

What the Clock Does

A game with a clock has two independent budgets, and every second you spend deciding comes out of yours alone. The clock does not care whether you are winning, which is why time is treated as a resource on equal footing with material. At the appointed start of the session White's clock is started — even if White has not arrived yet, because White moves first.

The clock starts here, on White's side. Nobody has moved yet, and White's time is already running. Play this position →

Where the clock sits is the arbiter's decision in a tournament; the usual convention otherwise is Black's right — which is White's left — so both players reach it naturally.

Analog Clocks and the Literal Flag

The vocabulary of chess time comes from mechanical clocks. An analog clock has two spring-driven faces, and just before the minute hand reaches 12 it lifts a small hinged lever called the flag; when the hand passes 12, the flag drops. That physical fall is what "flagging" means and why players still say "your flag is down."

Because the flag falls at the top of the hour, you do not set an analog clock to the time control directly. For a 5-minute game set both faces to 5:55 so the flags fall at 6:00; for a 90-minute session set them to 10:30. Get this wrong and you will discover it at the worst possible moment.

Analog clocks cannot do increment or delay. They drift, they can be knocked out of true by a hard press, and reading how much time is really left when the hand is near the flag is guesswork. They are fine for a casual game, but every serious event now uses digital.

Digital Clocks

A digital clock shows remaining time as a number, counts in seconds (and often tenths under a minute), and can be programmed with increment, delay, multiple periods, and a move counter. The move counter is convenient and untrustworthy: it advances on clock presses, so a single missed or doubled press puts it out of sync permanently. If your control changes after move 40, the scoresheets are the record that counts — one of several practical reasons to keep clean chess notation.

A digital clock also adds the next period automatically, so the extra 30 minutes for reaching move 40 can appear before anyone has verified that 40 moves were really played. Treat a sudden jump in your opponent's time as something to check, not to trust.

The Handling Rules

None of these are difficult; all of them are things beginners get quietly wrong in their first tournament.

Press with the hand that moved

You must press the clock with the same hand you used to move the piece. This is not fussiness: if you could move with your right hand and hover over the button with your left, you would gain a fraction of a second every move and could effectively play and press simultaneously in a scramble. One hand does both jobs, all game.

After 1.e4 — the pawn is placed with one hand, and that same hand presses the clock. Only then is the move complete and Black's time running. Play this position →

The move is not complete until you press

Once you release the piece, the move stands — but the move is not completed until you press, and until you do, your own time keeps draining. Your opponent has no obligation to remind you. Forgetting to press after a move you were pleased with is one of the most common ways to lose a rapid game.

One hand, and the king first when castling

Every move is made with one hand only, castling included: touch the king first, move it two squares, then move the rook, then press. Touching the rook first can commit you to a rook move under touch-move — the full procedure is in the guide to castling. Promotion is the same: place the new piece with the same hand, then press. Pressing while the pawn still sits on the eighth rank waiting for a queen is not a completed move.

Do not hover, do not slap

Keeping a finger poised over the button during your opponent's turn is forbidden, as is punching the clock, picking it up, or knocking it over. A firm, flat press is all that is needed. You are entitled to stop both clocks to summon the arbiter — that is what the extra button or lever is for. There is more on the unwritten side of this in the chess etiquette guide.

Flag-Fall: What Actually Happens

A flag-fall is not automatic. Under the FIDE Laws the flag is considered to have fallen when the arbiter observes it or when either player makes a valid claim; if neither of you notices, the game carries on. If you are winning on time, look at the clock — nobody will do it for you.

Two things override a flag-fall entirely:

  • Checkmate ends the game instantly, even if your flag falls a second later.
  • A drawn position ends it too. Stalemate, or any position where no series of legal moves can produce mate, ends the game the moment it arises. Knowing the boundaries of stalemate is worth more in a scramble than any opening line.

The insufficient-material exception

Here is the rule that decides a great many blitz games: if your flag falls, you lose — unless your opponent cannot checkmate you by any possible series of legal moves. Note the wording. It is not "cannot force mate," it is "cannot mate at all, even if you cooperate."

Two knights cannot force mate against a bare king, but a two-knight mating position does exist, so if you flag here you lose:

Black to move — and Black's flag falls. Black loses. Two knights cannot force mate, but a mating position exists, so the Laws count it as sufficient material. Play this position →

A single bishop is different: no arrangement of king and bishop against a bare king is checkmate, cooperative or otherwise, so a flag-fall here is a draw:

Same scenario, one piece different: Black's flag falls and the game is drawn. King and bishop can never checkmate a lone king. Play this position →

Now add a pawn on the flagging side and everything changes again. King and knight cannot mate a bare king — but they can mate a king that is boxed in by its own pawn:

Checkmate. The h7 pawn is Black's, and it takes away the only flight square the knight does not cover. Because this position is reachable, flagging against a lone knight is a loss whenever you still have a pawn. Play this position →

The practical lesson is blunt: when you are about to flag, your own extra material can be what convicts you — and when you are the one claiming, glance at what is left on the board before you announce a win.

Both flags down

If both flags are down and nobody can establish which fell first, the Laws distinguish between periods: in the final period — the one in which all remaining moves must be completed — the game is drawn; in an earlier period it simply continues. In practice this rarely survives contact with a digital clock, which stops the moment the first side hits zero.

Increment and Delay

Fixed-time-only chess ("sudden death") produces a specific ugliness: completely won positions decided by who can shuffle pieces fastest with four seconds left. Increment and delay are the two engineering answers, and they are not the same thing.

Fischer increment

A fixed number of seconds is added to your clock after you complete each move. Bobby Fischer patented a digital clock built around this idea in the late 1980s, and it was used in his 1992 rematch with Boris Spassky. The key property is that increment accumulates: with a 30-second increment, moving in 10 leaves you 20 seconds richer than you were, so your total time can climb back up. That is why a 30-second increment makes technical endgames playable — you can never be reduced to a permanent scramble as long as you keep moving.

In notation, 3+2 means 3 minutes plus 2 seconds per move; 90+30 means 90 minutes plus 30 seconds per move from move one.

King and rook against a bare king. On a 30-second increment this is trivially winnable with seconds on your main clock. In sudden death with eight seconds left, it is a coin flip. Play this position →

If you cannot yet play that position quickly and without thinking, learn it — the king and rook checkmate and the king and queen checkmate are the two techniques the clock will test most often.

Bronstein delay

Named for David Bronstein, who argued for the idea. The clock adds back the time you actually used, up to the set maximum: move in 2 seconds with a 5-second Bronstein delay and you get 2 seconds back; take 20 and you get 5 back. The crucial difference from Fischer increment is that your main time can never rise above where it started. You get a free window each move, but you cannot bank the unused part.

Simple (US) delay

The same guarantee, displayed differently: your clock does not begin counting down until the delay has elapsed, so with a 5-second delay the first five seconds of every move are free and the display simply sits still. Nothing is added afterwards. US Chess events commonly write it as G/5 d3 — game in 5 minutes with a 3-second delay. In time remaining, simple and Bronstein delay are equivalent; only the numbers on the screen behave differently.

Time Controls, Format by Format

FIDE draws its boundaries using the total time available for a whole game, calculated as the base time plus 60 times the increment — that is, the time you would have if the game lasted 60 moves.

| Category | Base + 60 × increment | Typical example | | --- | --- | --- | | Standard (classical) | 60 minutes or more | 90+30, 120 min for 40 moves | | Rapid | more than 10, less than 60 minutes | 15+10, 25+10 | | Blitz | 10 minutes or less | 3+2, 5+0 |

Bullet is not a FIDE category at all. It is an online convention for anything below roughly 3 minutes, and it exists because a mouse and a premove button make speeds possible that physical pieces do not.

Classical

The championship control is 90 minutes for the first 40 moves, then 30 minutes for the rest, with a 30-second increment from move one. Other classical events use a single long period, sometimes with a second control at move 60.

Classical rewards deep calculation, long-term plans and real preparation, and it has its own discipline: a two-period control creates a hard cliff at move 40. Reaching move 40 with two minutes left is survivable; reaching move 38 with two minutes left is usually fatal. Note too that once you are under five minutes in a period without an increment of at least 30 seconds, you are not required to keep score — you complete the scoresheet afterwards.

Rapid

Roughly 10 to 60 minutes per side, usually with a small increment; the FIDE World Rapid Championship uses 15+10. Rapid is the best format for improvement: enough time to form a plan, check it, and notice the tactic that refutes it, but not enough to drift. Games finish inside an hour, so you can play several and still have the energy to review them.

Skills it trains: planning, calculation under mild pressure, converting advantages.

Blitz

Three to five minutes per side, most commonly 3+2 or 5+0; the FIDE World Blitz Championship uses 3+2. Nothing is calculated deeply here. You run on pattern recognition, openings you already know, and a feel for which positions are easy to play quickly. Blitz is genuinely useful for sharpening tactical vision and stress-testing a repertoire — and harmful if it is all you play, since it rewards speed over accuracy and hides habits a longer control would punish.

Skills it trains: intuition, opening familiarity, practical decision-making.

Bullet

One or two minutes each, typically 1+0 or 2+1, and online only. Bullet is a different game wearing chess clothing: premoves, mouse speed, and the deliberate creation of positions that are awkward to move quickly in. Top speed specialists play it at an extraordinary level — see Hikaru Nakamura's speed chess — but it teaches almost nothing transferable to a slower game.

Which should you play?

To improve, put most of your playing time into rapid, use blitz in moderate doses to test openings and sharpen recognition, and play some classical so you find out what your chess looks like when time is not the excuse. Rating pools are usually separate per format, which is worth knowing before you panic about a number — see how chess ratings work.

Spending the Clock Well

Time management is not "move faster." It is spending your budget where it changes the result.

Budget roughly. In a 90-minute, 40-move control you average about two minutes a move. You will not spend it evenly, but you should know when you have overdrawn. A useful checkpoint: at move 20, have you used more than half?

Spend on the branch points. The moves worth long thought are the ones that change the position permanently: pawn breaks, structural commitments, whether to trade queens, whether to enter a forcing line. Quiet developing moves in a position you understand deserve seconds, not minutes.

Do not re-solve solved problems. If you decided last move that the plan is ...c5 and nothing has changed, play ...c5. A lot of clock time disappears into re-deciding things.

Do not move instantly out of fear either. The most expensive habit in time trouble is playing the first legal-looking move to stay ahead of the clock. That is how won positions become lost ones — most cheaply through the back rank:

Checkmate in a scramble. The pawns never moved, the rook came to the eighth, and the game ended in the time it takes to press a clock. Play this position →

One free tempo spent on h3 or ...h6 earlier removes an entire class of time-trouble disaster. The full anatomy is in the back-rank mate guide, and the wider category in avoiding common blunders.

When you are short, simplify toward technique. With ten seconds and no increment, a clean extra rook in a simple endgame beats a bigger advantage in a sharp one, because you can play it without thinking. Trade pieces, remove counterplay, and steer toward positions your hands already know — a Philidor or a basic rook ending is far easier at speed than a queen middlegame.

With increment, treat the increment as a floor. On 3+2, dropping to five seconds is recoverable if you can move in under two; on 5+0 it is terminal. Know which kind of game you are in.

Notice your own pattern. Some players are chronically slow in the opening; others spend everything on one move and play the rest on fumes. Look at where the clock actually went — pre-game preparation and the deeper treatment in time management in chess both help, and if it is the clock rather than the position that makes you tense, chess anxiety is solvable in its own right.

Practise With a Real Clock

All of this becomes automatic with repetition and stays theoretical without it. Play timed games until the press stops being something you have to remember.

You can do that on the free board at LocalChess — set your own time control, add an increment or leave it as sudden death, and play the engine or a friend on the same device. Play one opening at 15+10 and then at 3+2 and watch how differently you think. Then take a diagram above, hit Play this position, and see whether you can convert it with ten seconds on the clock. That is the drill that turns clock rules into clock skill.