Arithmetic — A Complete Reference

1. Why Mental Arithmetic Still Matters

In an age when every phone carries a calculator and every spreadsheet auto-totals, it is fair to ask whether mental arithmetic is still worth practicing. The honest answer is that fluency with numbers is not really about replacing the calculator — it is about catching the moments when the calculator is wrong, the inputs are misread, or a quick decision must be made before any device is even unlocked. A shopper who instantly senses that a “30% off” sticker on a 47-dollar item should land near 33 dollars, not 17, has a protective layer of intuition that no app provides.

The deeper reason is cognitive transfer. Arithmetic is the substrate on which algebra, statistics, and programming are built. A student who must consciously calculate 7×8 has fewer mental cycles left for the algebra problem on top of it; an engineer who cannot sanity-check a unit conversion is one decimal point away from a serious mistake. Studies of working memory consistently show that people who have automatized basic facts solve higher-level problems faster and with fewer errors, because their attention is free to focus on structure rather than mechanics.

Mental arithmetic also produces a uniquely portable form of estimation. Engineers call this “back-of-the-envelope” reasoning, physicists call it Fermi estimation, and traders simply call it pricing in your head. Whatever the label, it is the same skill: reach an approximate answer in seconds, decide whether the precise calculation is even worth the effort, and notice immediately when a printed result is off by an order of magnitude. Finally, regular practice strengthens the neural pathways linked to numerical reasoning — even ten focused minutes a day measurably improves calculation speed within a month.

암산은 계산기를 대체하는 능력이 아니라 잘못된 입력·잘못된 결과를 즉시 알아채는 직관입니다. 수 감각은 대수·통계·프로그래밍의 토대이며, 추정(Fermi estimation)을 통해 의사결정을 가속화합니다.

2. Number Sense & Place Value

Every arithmetic algorithm taught in school rests on one quiet idea: positional notation in base 10. The digit 4 means four when it sits in the ones place, forty in the tens place, four hundred in the hundreds place, and four thousand in the thousands place — and so on indefinitely. The number 4 273 is therefore not four symbols but a compact polynomial: 4·10³ + 2·10² + 7·10¹ + 3·10⁰. Carrying, borrowing, long multiplication, and long division are all just bookkeeping on this polynomial.

The reason this matters is that place value awareness is what separates fluent calculators from people who memorize procedures. When a learner understands that 47 + 38 is really “(40 + 30) + (7 + 8),” every addition strategy in the next section becomes natural rather than magical. When they understand that multiplying by 10 shifts every digit one place to the left, the rule “add a zero” stops being a mysterious trick and becomes a consequence of how digits are arranged.

Base 10 is the dominant convention because humans grew up with ten fingers, but the same positional idea works in any base. Computers use binary (base 2), where the only digits are 0 and 1 and each place is a power of two: 1011₂ means 1·8 + 0·4 + 1·2 + 1·1 = 11. Programmers also use hexadecimal (base 16, digits 0–9 and A–F) because each hex digit packs exactly four binary digits, making it a compact way to write byte values. A color like #FF8800 is just three hex bytes for red, green, and blue. The same place-value reasoning a child uses to add 27 + 35 generalizes directly to adding 0x2A + 0x17 in a debugger.

10진법 자릿값은 모든 산술 알고리즘의 토대입니다. 같은 원리는 컴퓨터의 2진법·16진법에도 그대로 적용되어, 자릿값을 이해하면 모든 진법에서의 계산이 자연스러워집니다.

3. Addition Strategies

The school algorithm for addition processes columns right-to-left, carrying when a sum exceeds nine. It works, but for mental arithmetic it is rarely the fastest path. Several left-to-right and friendly-number strategies are much easier to hold in working memory because they let you announce useful partial results immediately rather than waiting for the rightmost column.

The bridge-to-ten strategy splits a number into two parts so that the first part completes the running total to a multiple of ten. To compute 47 + 38, split 38 into 3 + 35: 47 + 3 = 50, then 50 + 35 = 85. Once the running total hits a round number, the remaining addition involves no further carries. Friendly-number strategies push this further: replace 38 with 40 and subtract 2 at the end. So 47 + 38 becomes 47 + 40 − 2 = 87 − 2 = 85. This is called compensation: deliberately overshooting by a small amount that is easy to correct later.

A third option is genuine left-to-right addition. Add the largest place values first and refine: 47 + 38 = (40 + 30) + (7 + 8) = 70 + 15 = 85. The advantage is that the first partial sum (70) is already a good approximation, which is why estimation and exact calculation feel like the same process when done left-to-right. The disadvantage is that you must remember to carry in the last step. With a little practice, all three methods become interchangeable, and skilled mental calculators switch between them based on which numbers happen to be friendly. The same logic extends to subtraction’s mirror operations, covered next.

덧셈은 자리 올림 알고리즘보다 친근한 수 만들기·보정(47 + 38 = 47 + 40 − 2 = 85)·왼쪽부터 더하기가 암산에서 훨씬 빠릅니다. 세 가지를 상황에 맞게 전환하는 것이 핵심입니다.

4. Subtraction Strategies

Subtraction breaks more brains than addition because the school borrowing algorithm is harder to track in your head: scratching out digits, marking a small one above the next column, and remembering whether the current digit is the original or the modified one. Three alternative strategies sidestep most of the pain.

Counting up treats subtraction as the missing addend in an addition problem. For 503 − 178, ask “what plus 178 makes 503?” Step up to round numbers: 178 + 2 = 180, then 180 + 20 = 200, then 200 + 300 = 500, then 500 + 3 = 503. The total stepped up is 2 + 20 + 300 + 3 = 325, so 503 − 178 = 325. This is exactly how cashiers used to make change before electronic registers — and it is robust because every step lands on an easy round number.

The equal-addition method (sometimes called the Austrian method) avoids borrowing by adding the same amount to both numbers, which preserves their difference. For 503 − 178, add 22 to both: 525 − 200 = 325. Choosing the right “boost” is the only skill — pick the amount that turns the subtrahend (the number being subtracted) into a clean multiple of ten or one hundred.

Complementary subtraction is most useful when crossing zero, where school borrowing is at its worst. Compute 1000 − 287 by noticing that 1000 − 287 = (999 − 287) + 1. Subtracting from a string of nines never requires borrowing — just digit-by-digit: 9 − 2, 9 − 8, 9 − 7 = 7, 1, 2 — then add the 1 back: 712 + 1 = 713. Each of these methods rewards you with a calmer subtraction than borrowing ever can, and they reinforce the relationship between addition and subtraction that beginners often miss.

뺄셈은 자리 내림 대신 더해가기(missing addend), 같은 수 더하기(오스트리아식), 보수법(특히 0을 가로지를 때)을 쓰면 훨씬 깔끔합니다. 503 − 178은 더해가기로 325, 1000 − 287은 보수법으로 713이 됩니다.

5. Multiplication Strategies

Multiplication is where mental arithmetic gets genuinely fun, because there are so many shortcuts that the choice of method is part of the puzzle. The single most useful idea is the distributive law: a·(b + c) = a·b + a·c. Every “trick” below is a special case of distributing.

To compute 27 × 8 by hand, split 27 into 20 + 7: (20 × 8) + (7 × 8) = 160 + 56 = 216. Splitting by place value works for any size factor — and is exactly how long multiplication on paper is justified. Doubling and halving is another distributive trick: 16 × 25 is the same as 8 × 50 (double one factor, halve the other), which is the same as 4 × 100 = 400. Whenever one factor is even and the other ends in 5 or 25, doubling-and-halving can collapse the problem to a multiplication by 100 or 1000 in two steps.

Multiplication by powers of ten is the cleanest case: appending a zero multiplies by 10, two zeros multiply by 100, and so on, because base-10 place value does the work for you. Combine this with distribution and almost any small problem becomes tractable: 36 × 11 = 36 × 10 + 36 × 1 = 360 + 36 = 396. The lattice or grid method, finally, organizes a multi-digit product into a small table — one row per place value of the first factor, one column per place value of the second — and is essentially long multiplication with the carries deferred until the end.

Worked example. Compute 27 × 64 mentally. Distribute the smaller factor: 27 × 64 = 27 × (60 + 4) = 27 × 60 + 27 × 4. Each piece is easy: 27 × 6 = 162, so 27 × 60 = 1620; and 27 × 4 = 108. Add: 1620 + 108 = 1728. Cross-check with a doubling argument: 27 × 64 = 27 × 2⁶ = 54 × 32 = 108 × 16 = 216 × 8 = 432 × 4 = 864 × 2 = 1728. Two independent paths to the same answer make a mental-math result trustworthy.

곱셈은 본질적으로 분배법칙입니다. 27 × 8 = 20×8 + 7×8 = 216, 16 × 25 = 8 × 50 = 400, 10의 거듭제곱 곱셈은 자릿수 이동으로 처리됩니다. 두 가지 경로로 같은 답을 얻어 검산하는 습관이 중요합니다.

6. Division Strategies

Division is the operation people fear the most, partly because it has two distinct interpretations that schools rarely separate. In the partitive view, 12 ÷ 4 asks “if I share 12 cookies equally among 4 children, how many does each get?” In the quotative view, the same expression asks “if I hand out cookies 4 at a time, how many groups can I make?” Both yield 3, but real-world problems usually fit one interpretation more naturally — and once a student notices the distinction, word problems become easier to set up.

For mental computation, the most reliable approach is estimate and refine. To compute 432 ÷ 18, ask first “roughly how many 18s fit in 432?” Twenty 18s would be 360, and thirty would be 540, so the answer lives between 20 and 30. Try 24: 24 × 18 = 24 × 20 − 24 × 2 = 480 − 48 = 432. The estimate landed on the exact answer in one refinement. This loop — estimate, multiply back, adjust — is exactly the loop the long division algorithm formalizes, one digit at a time.

Dividing by fractions confuses learners because the rule “invert and multiply” feels arbitrary. The intuition is simpler: dividing by ½ asks “how many halves are in this number?” — and there are obviously twice as many halves as wholes. So dividing by ½ is the same as multiplying by 2, dividing by ⅓ is the same as multiplying by 3, and dividing by ¾ is the same as multiplying by 4/3. Formally, a ÷ (b/c) = a · (c/b) because multiplying numerator and denominator of the original fraction by c/b clears the denominator. Once “invert and multiply” is grounded in the question it answers, fraction division stops being a magic incantation and becomes a special case of multiplication.

나눗셈은 등분(partitive)과 묶음(quotative) 두 해석이 있고, 추정 후 보정하는 방식이 신뢰성이 높습니다. 분수 나눗셈의 "뒤집어 곱하기"는 "이 수에 ½이 몇 개 들어있나?" 같은 질문을 정량화한 결과입니다.

7. Fractions, Decimals & Percents

Fractions, decimals, and percents are three notations for the same idea — a part of a whole — so fluency means being able to convert between them in either direction without pausing. The benchmark equivalences that should feel automatic are: ½ = 0.5 = 50%, ¼ = 0.25 = 25%, ¾ = 0.75 = 75%, ⅕ = 0.2 = 20%, ⅒ = 0.1 = 10%, ⅓ ≈ 0.333 ≈ 33.3%, and ⅛ = 0.125 = 12.5%. Memorizing this short table covers most percentage problems that occur in daily life.

Conversions follow simple rules. A fraction becomes a decimal by long-dividing the numerator by the denominator: 3 ÷ 8 = 0.375. A decimal becomes a percent by multiplying by 100 — that is, by shifting the decimal point two places to the right: 0.375 → 37.5%. A percent becomes a decimal by reversing that shift: 12.5% → 0.125. And a decimal becomes a fraction by writing it over the appropriate power of ten and simplifying: 0.24 = 24/100 = 6/25.

For mental work, three shortcuts handle the majority of cases. Ten percent of any number is obtained by sliding the decimal one place left: 10% of 247 is 24.7. One percent is two places left: 1% of 247 is 2.47. Fifty percent is half. From these three building blocks, any reasonable percentage can be assembled: 35% of 80 is (30% of 80) + (5% of 80) = 24 + 4 = 28, where 30% = 3 × 10% and 5% = ½ × 10%.

Worked example. A 45-dollar shirt is 30% off — what is the sale price? Compute 30% of 45 by first finding 10%: 4.50. Triple it: 13.50. The discount is 13.50, so the sale price is 45 − 13.50 = 31.50. Cross-check: 70% of 45 (the price you actually pay) equals 7 × 4.50 = 31.50. Same answer through two paths — a reassuring habit, especially when a store sign is ambiguous about whether a percentage is the discount or the price.

분수·소수·백분율은 같은 양의 세 가지 표기일 뿐이며, ½ = 0.5 = 50% 같은 기준 등가관계를 외워두면 변환이 즉시 됩니다. 10%(소수점 한 칸 이동), 1%(두 칸), 50%(반)로 거의 모든 퍼센트 계산을 조립할 수 있습니다.

8. Order of Operations (PEMDAS / BODMAS)

When an expression mixes several operations, the order of operations convention specifies which to perform first. The English-speaking world remembers it as PEMDAS (Parentheses, Exponents, Multiplication, Division, Addition, Subtraction); much of the Commonwealth uses BODMAS (Brackets, Orders, Division, Multiplication, Addition, Subtraction). Both encode the same rule: parentheses outrank exponents, exponents outrank multiplicative operations, and multiplicative outranks additive. Within the same tier — for example, multiplication and division — operations are performed left to right in the order they appear.

The convention exists because, without it, the same string of symbols would have ambiguous meaning. The expression 2 + 3 × 4 must equal either 14 or 20, and pre-electronic mathematicians agreed it should equal 14 (multiplication first) so that polynomials like 3x² + 2x − 1 could be written without parentheses around every term. Convention is not the same as natural law; it is a choice that allowed mathematical writing to compress.

The viral debates over expressions like 6 ÷ 2(1 + 2) arise because two slightly different conventions disagree. Strictly applied left-to-right, division and multiplication have equal precedence: 6 ÷ 2 × 3 = 3 × 3 = 9. Some older or implicit-multiplication conventions treat 2(1 + 2) as a tightly bound product that should be evaluated before the division, yielding 6 ÷ 6 = 1. Both readings are defensible under their respective traditions, which is precisely why such expressions are bad mathematical writing: a careful author uses an extra pair of parentheses or a fraction bar so that no one has to guess. In modern programming languages, the standard is unambiguous — / and * are equal precedence and left-associative — so 6 / 2 * 3 always evaluates to 9.

PEMDAS/BODMAS는 같은 식이 같은 값을 갖도록 약속한 관습입니다. 같은 우선순위의 ×, ÷는 왼쪽부터 처리하며, 6÷2(1+2) 같은 모호한 식은 괄호를 더 써서 의도를 명확히 해야 합니다.

9. Common Mistakes & Pitfalls

• Sign errors with negatives. Subtracting a negative becomes addition (5 − (−3) = 8) and multiplying two negatives produces a positive (−2 × −3 = 6). The dangerous case is mixing both: −3² is usually read as −(3²) = −9 in standard precedence, while (−3)² = 9. When in doubt, write the parentheses explicitly rather than trusting reader-side telepathy.
• Decimal point misplacement. Multiplying 0.4 by 0.3 produces 0.12, not 0.012 — count total decimal places in both factors (1 + 1 = 2) and place the point that many positions from the right of the integer product (12 → 0.12). A common variant: forgetting that 12.5% is 0.125, not 0.0125, when converting to a decimal multiplier.
• Ignoring parentheses. Dropping a pair of parentheses is one of the most frequent algebra bugs that surfaces in arithmetic: −(3 − 5) is +2, not −3 − 5 = −8. If a minus sign sits in front of a parenthesized expression, distributing it changes the sign of every term inside.
• Calculator habits causing precedence bugs. Spreadsheets and programming languages follow strict PEMDAS, but cheap pocket calculators often evaluate keystrokes left-to-right with no precedence (“9 − 1 = 8, × 2 = 16” instead of 7). Always verify which mode your tool is in, especially when entering long expressions in one chain.
• Rounding too early in chained calculations. Rounding 1/3 to 0.33 before multiplying by 300 gives 99, but the true value is 100. Whenever a result feeds into the next step, keep at least two or three more digits than you ultimately need and round only once at the very end.
• Confusing percent of and percentage points. A rate moving from 10% to 12% has risen by 2 percentage points — but by 20 percent of its original value. Newspapers conflate the two constantly; precise reasoning depends on choosing the right one.

가장 흔한 실수는 음수 부호, 소수점 위치, 괄호 분배 누락, 그리고 중간 단계에서의 조기 반올림입니다. "퍼센트포인트"와 "퍼센트"의 차이도 통계 해석에서 중요한 함정입니다.

10. Applications & Related Concepts

Strong arithmetic shows up in every domain where numbers need to be trusted:

• Daily life. Tipping 15–20% in restaurants is just 10% (decimal shift) plus half of itself or a doubled tenth. Splitting a bill among four people is two halvings in a row. Unit-price comparisons at a grocery store — say, $3.20 for 400 g versus $4.50 for 600 g — reduce to mentally dividing price by quantity and comparing per-100 g rates (80¢ vs 75¢).
• Finance. Compound interest can be estimated by the rule of 72: dividing 72 by the annual percentage rate gives the rough doubling time. At 6% per year, money doubles in about 12 years; at 9%, in about 8. Investors who can carry this mental rule run sanity checks on projections that brokers and spreadsheets sometimes get spectacularly wrong.
• Science and engineering. Fermi estimation asks “how many piano tuners in Chicago” or “how many heartbeats in a lifetime” — multi-step approximations where each factor is known only to one significant figure but the product is still within a factor of two of reality. Engineers use the same habit to sanity-check finite-element simulations: if the printed answer is far from the back-of-envelope estimate, one of the two is wrong, and finding which is half the job.
• Education. Arithmetic is the foundation on which algebra, statistics, and calculus rest. A student fluent in mental math arrives at algebra free to focus on the unknown letters rather than the numerical mechanics surrounding them; a student fluent in algebra arrives at calculus with the same advantage.

Cross-links: Arithmetic is the foundation for statistics and probability; for arithmetic-meets-geometry (areas, lengths) see /docs/geo/. The matrix and vector guides show how the same arithmetic operations generalize to higher dimensions.

Practice → C:Arith

암산은 팁 계산·청구서 분할·단가 비교 같은 일상부터 72의 법칙(복리), Fermi 추정, 공학적 자릿수 점검까지 모든 정량적 판단의 기반입니다. 실전 연습은 C:Arith에서 가능합니다.

11. Explore Each Operation in Depth

Each of the four basic operations has its own focused, worked-example guide. Use these as deep dives whenever you need the standard algorithm, a step-by-step example, and the key properties for one specific operation:

• Addition — the column algorithm, carrying, and a worked example (376 + 248 = 624).
• Subtraction — borrowing across zeros and a worked example (503 − 178 = 325).
• Multiplication — long multiplication, the area model, and 27 × 34 = 918.
• Division — long division, remainders, and 950 ÷ 4 = 237 r 2.

사칙연산마다 표준 알고리즘·단계별 예제·주요 성질을 담은 심화 가이드가 있습니다: 덧셈, 뺄셈, 곱셈, 나눗셈.