Mental math techniques worth learning, and how to practise them

Arithmetic is a skill, not a talent

People who calculate quickly are almost never doing what you are doing, only faster. They are doing something different: working in a direction that suits the head rather than the page, rounding to a convenient number and correcting afterwards, and recognising shapes they have met before. None of that is innate. It is a small set of techniques plus enough practice that choosing between them stops being a decision. The techniques below are the ones that repay learning first.

Work left to right

On paper you add and multiply from the right, because that is what carrying requires. In your head this is the wrong way round, because the digits you need first for an estimate are the ones on the left, and because holding a half-finished number from the right is harder. To add 347 and 285, take 300 and 200 to get 500, then 40 and 80 to get 620, then 7 and 5 to get 632. You always have a usable approximation, and the load stays small.

Compensate to a round number

Round to something easy, then correct. For 68 plus 27, make it 70 plus 27 and subtract 2. For 495 plus 380, make it 500 plus 380 and subtract 5. For 8 times 49, do 8 times 50 and subtract 8. The correction is almost always simpler than the original problem, and this single habit removes most of the difficulty from awkward-looking numbers.

Double and halve

Multiplication is unchanged if you double one factor and halve the other. For 16 times 25, halve and double to get 8 times 50, then again to get 4 times 100, which is 400. This works particularly well when one number is even and the other is close to a round number, and it turns many two-digit multiplications into something you can do without writing anything down.

Times five, times fifty

Multiplying by 5 is halving then multiplying by 10. For 84 times 5, halve to get 42, then add a zero for 420. Multiplying by 50 is the same with two zeros. This is faster and less error-prone than the standard method, and it is the shortcut most people adopt permanently once they try it.

Times nine and times eleven

To multiply by 9, multiply by 10 and subtract the original number: 9 times 47 is 470 minus 47, which is 423. To multiply a two-digit number by 11, add the digits and place the sum between them: 11 times 43 is 4, then 4 plus 3, then 3, giving 473. When the digit sum exceeds nine, carry into the first digit: 11 times 68 gives 6, then 14, then 8, so 748.

Times twenty-five

Multiplying by 25 is multiplying by 100 and dividing by 4. For 36 times 25, take 3600 and divide by 4 to get 900. Dividing by 4 is two halvings, so nothing here needs long division. The same idea covers 75, which is three quarters of 100, and 125, which is an eighth of 1000.

Squares and near-squares

Squares ending in 5 have a rule: take the leading digits, multiply by one more than themselves, and append 25. For 35 squared, 3 times 4 is 12, so the answer is 1225. For numbers near a square you know, use the difference of two squares: 49 times 51 is 50 squared minus 1, which is 2499. Learning the squares up to about 25 pays for itself repeatedly.

Estimate first, calculate second

Before any calculation worth doing, produce a rough answer by rounding hard. It costs a second and it catches the errors that matter, which are not small slips but answers wrong by a factor of ten. This habit is more valuable in daily life than any individual trick, because most real arithmetic exists to support a decision, and a decision usually needs the right order of magnitude more than it needs the last digit.

Subtract by adding up

Subtraction is often easier run forwards. To take 168 from 400, count up: 168 to 200 is 32, and 200 to 400 is 200, so the answer is 232. This is how change is given in a shop and it avoids borrowing entirely, which is where most mental subtraction errors come from. It also keeps a running partial answer, so an interruption does not cost you the whole calculation.

Break numbers apart, not down

Any awkward multiplication can be split into pieces you already know. For 23 times 7, do 20 times 7 to get 140, then 3 times 7 to get 21, then add for 161. This is the distributive property and it is the single most reusable idea in mental arithmetic, because it turns anything unfamiliar into a sum of things that are familiar. Most of the named tricks above are special cases of it.

Check with the last digit

A quick verification costs almost nothing: the last digit of a product depends only on the last digits of the factors. If 23 times 7 gave you 160, the last digit should be 1, since 3 times 7 ends in 1, so you know immediately that something slipped. Combined with an order-of-magnitude estimate, this catches the large majority of errors without redoing the calculation.

Learn the pieces that repay themselves

A small stock of memorised facts removes far more work than any technique. Times tables to twelve, squares to twenty-five, the doublings up to a thousand, and the common fraction and percentage equivalents. These are not glamorous and they are what separates someone who calculates comfortably from someone who reconstructs everything from scratch each time. Learn them in short spaced sessions rather than in one sitting.

Divide by chunking

Division is easier when you subtract convenient multiples rather than following the written algorithm. To divide 462 by 14, note that 14 times 30 is 420, leaving 42, and 14 times 3 is 42, so the answer is 33. You are choosing chunks you already know instead of producing digits one at a time, which fits how the head works far better.

Fractions and percentages

Ten per cent is a decimal shift; five per cent is half of that; one per cent is two shifts. Most percentage questions become easy by building them out of those three. Fifteen per cent of 240 is 24 plus 12, which is 36. Percentages are also reversible: 8 per cent of 50 is the same as 50 per cent of 8, which is 4, and choosing the easier direction is often the whole trick.

Practise short and often

The most reliable finding about practice schedules is that spacing beats massing. A 2006 review by Cepeda and colleagues examined distributed practice across a large body of studies and found that spreading the same amount of practice over time produces better retention than concentrating it. In practical terms, five minutes daily beats forty minutes on Sunday, and it is also the schedule people actually keep.

Correct every error immediately

An error you do not examine is a repetition waiting to happen. When you get something wrong, the useful step is not to try again but to see which stage broke: the method chosen, the arithmetic inside it, or a slip in holding a number. Naming the stage takes seconds and stops the same mistake recurring, which is why practice with immediate correction beats practice with a score at the end.

Measure honestly

Track two things: accuracy at a fixed difficulty, and the time taken. Speed alone rewards guessing and accuracy alone rewards caution, and only the pair together shows improvement. Compare weeks rather than sessions, since attention and tiredness move a single session more than skill does. Numvolt is built around this: a calm mode with no timer where every wrong answer shows the method behind the correction, and timed sprints when you want to measure the pair together.

Further reading

• Cepeda et al. (2006), Distributed practice in verbal recall tasks: a review and quantitative synthesis, Psychological Bulletin