Board vision · Part 2 of the fundamentals series Ask a strong player what colour c6 is and the answer comes back in under a second ...
Board vision · Part 2 of the fundamentals series
Ask a strong player what colour c6 is and the answer comes back in under a second — no counting, no arithmetic. That is not talent. It is a map they built on purpose, and you can build the same one in two weeks.
Most beginners "know" the board the way you know a stranger's phone number — only while looking at it. The moment the board leaves your eyes, the map goes fuzzy. That fuzziness is what makes you hang a bishop to a knight you never saw, lose your place halfway through a calculation, and write Nf3 on the scoresheet when you played Nf6.
This post is about fixing that. Not by grinding all 64 square names into your head with flashcards — that method is slow and, as you will see below, it is not what strong players actually did. There is a much better route, and it takes about ten minutes a day.
What "memorizing the board" actually means
It is four separate skills, and they stack. Each one is easier once the one before it is automatic, which is why people who try to jump straight to blindfold chess get stuck.
| Layer | What you learn | You know it when… | Time |
|---|---|---|---|
| 1. Names | Every square's coordinate, from both sides of the board | You can point to f5 without counting up the file | 2–3 days |
| 2. Colours | Whether any named square is light or dark | You answer "g4?" in under a second | 3–5 days |
| 3. Lines | Ranks, files and especially diagonals as whole objects | You feel that c1 and h6 are connected | 1–2 weeks |
| 4. Knight geometry | Where a knight lands, and how far it really is | Knight forks jump out instead of being calculated | 2 weeks of drills |
White on the right. The near-right corner square must be light. Sitting behind White, that corner is h1, and h1 is light. Set the board up the other way and every square name you learn is shifted onto the wrong grid. It is also the classic tell that a film's prop department never asked a chess player — watch for it.
Layer 1: the names
Files run a to h left to right from White's seat. Ranks run 1 to 8 away from White. Every square is file letter first, rank number second: e4, b7, h1. If any of that is new, start with the full walkthrough in our chess notation guide and come back.
The trap here is one-sided practice. Almost everyone learns the board from White's side only, then sits down with the black pieces at a tournament and finds the whole map has flipped. Train both orientations from day one — the drill below has a toggle for exactly this.
Drill 1 · Find the square
Tap the square you are asked for · 10 questions
Press Start.
Layer 2: the colours, in two seconds
This is the layer almost everyone skips, and it is the one that pays off fastest. Knowing colours instantly tells you which squares your bishop can never defend, which squares a knight will land on next, and whether an endgame is drawn before you calculate a single move.
You do not have to memorize 64 answers. There is one rule that generates all of them.
Turn the letter into its position in the alphabet: a=1, b=2, c=3, d=4, e=5, f=6, g=7, h=8.
Now compare it with the rank number. Both odd, or both even → dark. One odd, one even → light.
So c3 is 3 and 3, both odd, dark. e4 is 5 and 4, mixed, light. g5 is 7 and 5, both odd, dark.
Some people find it easier as a light switch. Two switches wired to one bulb: if both are flipped the same way the light is off, if they differ the light is on. Same-parity means dark, mixed means light. Others prefer addition — add the two numbers, an even total is dark, an odd total is light. All three are the same rule wearing different clothes. Pick whichever one sticks and use only that one.
The rule is scaffolding, not the destination. You use it for a week or two, and then you stop needing it — the answers start arriving before the arithmetic does. If you are still calculating after a month of daily practice, you are not doing enough repetitions per session.
The map
Here is the whole board with every square named. Switch it to The maths to see the rule applied to all 64 squares at once, then to Colours only and try to name squares from memory.
The colour map
Tap any square for its name, colour and working
32 light squares 32 dark squares
Drill 2 · Light or dark
Keyboard: press L or D · 30-second round
Press Start, then answer as fast as you can.
The shortcut nobody tells beginners
You already know more colours than you think. Every opening move you have ever played is a square whose colour you can recall through the move rather than through arithmetic. This is the fastest on-ramp there is, especially for kids, because it hangs new facts on things they already care about.
| You know this | So you know | Colour |
|---|---|---|
| Bishop to c4 in the Italian | It is the light-squared bishop | c4 light |
| Bishop to b5 in the Ruy Lopez | Same bishop, same colour | b5 light |
| Bishop to g5 pinning a knight | That is the dark-squared bishop | g5 dark |
| Fianchetto Bg2 and Bb2 | Two different bishops, two different colours | g2 light, b2 dark |
| Black fianchettos Bg7 and Bb7 | Mirror of White's, so colours swap | g7 dark, b7 light |
| Scholar's Mate lands on f7 | The queen came down a light diagonal | f7 light, f2 dark |
| The four centre squares | Two of each, diagonally paired | e4, d5 light · d4, e5 dark |
Notice what just happened. You did not memorize seven isolated facts — you attached them to patterns already in your head. That is not a study hack, it is the mechanism behind every strong player's board memory, and there is sixty years of research explaining why it works.
Why rote memorization fails (and what works instead)
In the 1940s the Dutch psychologist Adriaan de Groot showed chess positions to players for a few seconds and asked them to rebuild them. Masters were startlingly good. In 1973 William Chase and Herbert Simon added the experiment that made the result famous: they repeated it with randomly scattered pieces.
The masters' advantage nearly vanished. Faced with a position that could not occur in a real game, a grandmaster remembered barely more than a beginner. Later work refined this — a small skill effect on random positions does survive — but the headline held: in game positions masters recalled roughly three times as many pieces as club players, and on random positions only about one and a half times as many.
The explanation is chunking. Masters do not store 32 separate pieces. They store familiar clusters — a castled king with its pawn shield, a knight outpost with its supporting pawn, a standard rook-and-pawn ending — each recalled as one unit. Estimates of how many such chunks a master holds run from tens of thousands upward.
Memory in chess is a by-product of meaning. Drilling a1 through h8 as 64 disconnected labels is exactly the random-position condition, and it is exactly as hard as it sounds. Learn the board through rules (the parity trick), relationships (diagonals, knight jumps) and positions you already know (openings, mating patterns) and the same 64 facts cost a fraction of the effort.
Layer 3: stop seeing squares, start seeing lines
The board is 8 ranks, 8 files, and 26 diagonals of two squares or more — 13 slanting each way. Ranks and files are easy; the eye reads rows and columns naturally. Diagonals are not. Nearly every "I never saw that bishop" blunder is a slanted line that never registered.
Two of those diagonals matter more than the rest. a1–h8 and h1–a8 are the only eight-square diagonals. They cross all four centre squares and connect corner to corner, which is why the fianchetto exists as an opening idea at all. And here the two layers lock together: a1–h8 is entirely dark, h1–a8 is entirely light. A diagonal never changes colour, which is why a bishop is stuck on one colour for the whole game — and why "colour complex" and "diagonal strategy" are the same subject.
The tracing drill
This one needs a real board, and the physical part is not optional — touching the squares is what welds the pattern in.
- Put a bishop on c1. Say out loud every square it reaches on both diagonals, touching each one.
- Now do it with your eyes closed, then open them and check.
- Repeat from f1, c8, f8, then from central squares like d4 and e5.
- Do the same with a rook, sweeping the full rank and file and naming each square.
Three minutes a day for a week. The goal is that c1 and h6 stop being two square names and become two ends of one object.
Layer 4: the knight, where everyone leaks rating points
Bishops and rooks move along lines your eye already follows. The knight does not, and that is why most blunders below club level involve one — either a fork you missed, or a square you put a piece on without noticing a knight covered it.
Here is the fact that turns knight geometry from calculation into knowledge:
A knight always changes square colour with every single move. Light, dark, light, dark — without exception, forever.
So a knight on a dark square reaches dark squares only in an even number of moves, and light squares only in an odd number.
That one law does a surprising amount of work. Take a knight on d4 and ask how many moves it needs to reach d5 — the square directly in front of it. One move is impossible, you can see that. Two is impossible too, and you can now prove it without looking: d4 is dark, d5 is light, so the answer must be odd. It takes three, for example d4–f5–e7–d5.
The nastier one: a knight on a1 needs four moves to reach b2, the square diagonally touching it. Same colour, so the answer is even; two moves does not work; four does. Knights are slowest exactly where they look closest, and that costs games in endgames.
Knight Sight
This drill comes from Michael de la Maza's micro-drill programme, popularised on Chess.com, and it is the single most efficient board-vision exercise there is. On a real board: put a knight on a1 and physically touch the two squares it can reach. Move it to a2, repeat. Work through every square on the board, rank by rank. Ten minutes. Do it daily for a week and the knight's destinations stop being calculated and start being seen.
Use the trainer below when you do not have a board in front of you.
Drill 3 · Knight sight
Tap every square the knight can reach in one move
Press New position.
Once destinations pop out instantly, switch to shortest paths. Knight on d5: minimum moves to e4? To c6? To h8? Work outward in rings around the starting square, and use the colour-flip law to rule out impossible answers before you calculate. (Answers: two, two and three. Notice that the law alone tells you the first two must be even and the last must be odd.)
The 14-day plan
Ten minutes a day, every day. Short and frequent beats long and occasional here, because this is pure pattern learning — the kind of memory that responds to spacing rather than grinding. One 70-minute Sunday session will do less for you than seven ten-minute ones.
- Day 1–2Names, both sidesDrill 1 until you are above 90% correct from the White view, then switch to the Black view and rebuild it there. Expect the second one to feel like starting over. It is supposed to.
- Day 3–5Colours with the ruleStudy the colour map, then Drill 2 twice a day. Use the parity rule deliberately — say the working out loud. Target 20 correct in a 30-second round.
- Day 6–7Colours without the ruleSame drill, but answer on instinct and only fall back on the arithmetic when you stall. Add the opening-anchor table. Target 25 in 30 seconds.
- Day 8–10LinesThe bishop and rook tracing drill on a real board, three minutes. Then name both long diagonals end to end from memory, in both directions.
- Day 11–14Knight sightKnight Sight across the whole board on a real set, plus Drill 3. Finish each session with three shortest-path questions.
- ThenMaintainTwo short sessions a week, and one refresher the morning of a tournament. Board vision fades like any skill, but it comes back in minutes rather than days.
Six mistakes that waste weeks
| Mistake | Why it hurts | Do this |
|---|---|---|
| Learning only from White's side | Half your games are played from the other end | Alternate orientations in every session |
| One long session a week | Pattern memory needs spacing, not volume | 10 minutes daily, no exceptions |
| Rote-drilling 64 names cold | This is the random-position condition — maximum effort, minimum retention | Use the parity rule and opening anchors |
| Skipping colours to rush at blindfold play | Your mental board collapses two or three moves in | Finish layers 1–2 before visualizing anything |
| Drilling knights on screen only | Touching the squares is what cements them | Do Knight Sight on a real board with your finger |
| Practising with pieces on the board | You end up reading landmarks instead of the grid | Drill on an empty board and both labelled and unlabelled |
So can you actually play blindfolded?
Yes — and it is much less mystical than it looks. Blindfold play is built on exactly the four layers above, running fast enough that you can update the map faster than you forget it. Strong players do not hold a photograph of the board in their heads. They hold relationships: which pieces defend which, which lines are open, where the king is safe. That is why the drills on this page are the real preparation, not a separate hobby.
The record book gives a sense of the ceiling. George Koltanowski played 34 boards blindfold at Edinburgh in 1937 over thirteen hours — the version Guinness recognised for decades. Miguel Najdorf reached 45 boards in São Paulo in 1947, in an exhibition he staged partly to signal to his family in Poland that he was still alive. Marc Lang made 46 in 2011. The current mark belongs to Timur Gareyev, who played 48 games simultaneously at the University of Nevada in December 2016 — nineteen hours, wearing a blindfold, on an exercise bike, scoring 35 wins, 7 draws and 6 losses.
Nobody is asking you to do that. But the first rung of that ladder is answering "what colour is c6?" without thinking, and that rung is genuinely two weeks away.
Drill the coordinates properly
Our ten-level Chess Notation Master takes square recognition from first steps to full-speed recall, in the browser and free.
Open Chess Notation MasterQuestions people ask
How long does it really take to memorize the chessboard?
Square names take two to three days of ten-minute sessions. Colours become automatic in about a week if you use the parity rule rather than raw memorization. Knight geometry takes the longest — two weeks of daily drilling before destinations genuinely pop out. Total: around two weeks to a usable internal map, and it keeps sharpening for months after that.
What is the fastest trick for square colours?
Convert the file letter to a number (a=1 through h=8) and compare it with the rank. Both odd or both even means dark; one of each means light. So c3 (3,3) is dark and e4 (5,4) is light. Treat it as scaffolding — use it for a week or two until recall overtakes the arithmetic.
Do I need to know square colours to be good at chess?
You do not need it to enjoy chess, but it removes a surprising amount of friction. It tells you instantly which squares your bishop can never touch, lets you evaluate opposite-coloured bishop endings at a glance, and makes knight calculation faster through the colour-flip law. It also makes reading and following annotated games far less tiring.
What age can a child start this?
Around six or seven, once they can play a full game. Keep sessions to five minutes and make them physical — pointing at squares, calling colours out loud, racing a timer. The opening-anchor table works especially well with children because it attaches new facts to moves they already enjoy playing.
Should I practise from Black's side too?
Yes, and most players badly neglect it. If you only ever train from White's view, sitting down with the black pieces feels like a different board. Alternate orientations in every session from the start — the toggle in Drill 1 is there for this.
Why does a knight need three moves to go one square forward?
Because a knight changes square colour on every move. d4 is dark and d5 is light, so any route between them must use an odd number of moves. One is impossible, so the answer is three — for example d4–f5–e7–d5. The same logic explains why a1 to b2, two touching squares of the same colour, takes four.
Is board vision the same as visualization?
They are related but not identical. Board vision is knowing the fixed map — names, colours, lines, knight jumps — and it never changes from game to game. Visualization is tracking a position as it moves through that map. Board vision is the foundation; visualization is what you build on it, which is why the layers on this page come in the order they do.
Will this stop me from blundering?
It stops one particular family of blunders: the ones where you simply did not see a square was covered. That is a large share of losses below club level, especially the knight ones. It will not fix rushed moves or bad plans — but it clears out the errors that have nothing to do with how well you understand chess.
Sources and further reading de Groot (1946) and Chase & Simon, "Perception in Chess" (1973), on expert recall and chunking; Gobet & Simon (1998) on the revised picture for random positions; Michael de la Maza's micro-drills, as published on Chess.com, for Concentric Squares and Knight Sight; Chess.com and ChessBase reporting on the blindfold simultaneous records of Koltanowski, Najdorf, Lang and Gareyev.
Next in this series: how to calculate three moves ahead without losing the position. If a particular drill on this page gave you trouble, say which one in the comments — it tells us what to write.
No comments