Significant Figures Calculator
Count the significant figures in any number, round to any number of sig figs, or run a whole calculation and see which input limited the answer.
- Significant
- Not significant
- Not declared
- Significant digits
4560- First significant digit
4- Last significant digit
0- Last significant place
- millionths (10⁻⁶)
- Scientific notation
4.560 × 10⁻³- E notation
4.560e-3- Notation read
- decimal
- Exact value
0.00456
Why
- Every non-zero digit is significant, starting at the leading 4.
- The 3 zeros before the first non-zero digit only place the decimal point, so they are not significant.
- The 1 trailing zero is significant because the notation declares it — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision.
- The last significant digit sits in the millionths place (10⁻⁶).
Kept: 123 · First discarded digit: 4 · Decision: down
The first discarded digit is 4 followed by 5. The last kept digit stays as it is.
- Significant
- Not significant
- Not declared
- Plain decimal
12300ambiguous — trailing zeros cannot declare precision- Scientific notation
1.23 × 10⁴- E notation
1.23e+4- Rounding method
- Half away from zero (standard classroom rounding)
- Input precision
- 5 significant figures
Step by step
- Read the precision of the input"12345" carries 5 significant figures.
- Locate the cutKeeping 3 significant figures means keeping the digits 123 and discarding 45.
- Apply the rounding policyThe first discarded digit is 4 followed by 5. The last kept digit stays as it is.
- Write the answer so it declares its precisionPlain decimal form would hide the precision, so the answer is written as 1.23 × 10⁴.
- AMBIGUOUS_DECIMAL_FORM Written plainly the answer is "12300", whose trailing zeros cannot declare precision. Write 1.23 × 10⁴ to state 3 significant figures unambiguously.
What limited the answer?
1.4 — the measured operand with the fewest significant figures (2).
Rule: The result carries as many significant figures as the measured operand with the fewest significant figures. Exact numbers do not limit the result.
Operands
4.561.4limits the answer
- Raw (unrounded) result
6.384- Reported result
6.4- Scientific notation
6.4 × 10⁰- Rule applied
- Multiplication & division — fewest significant figures
- Rounding
- Half away from zero (standard classroom rounding) — kept 63, first discarded digit 8, rounded up
- Guard digits
- 40 — Every intermediate value is carried unrounded; rounding is applied once, to the reported answer.
Step by step
- 4.56 × 1.4Unrounded value 6.384. Fewest significant figures among the measured operands: 2, so the reported value is 6.4.Multiplication & division — fewest significant figures
How to Use the Significant Figures Calculator
Count
Type a number and the calculator reads it as written before it reads it as a value. It reports how many significant figures the notation declares, marks each digit as significant, a leading zero or an undeclared trailing zero, names the place of the last significant digit, and restates the value in scientific notation. If the number is a count or a definition rather than a measurement, tick exact and it will carry unlimited precision.
Round
Enter a number and a target number of significant figures. The result shows where the number is cut, what the first discarded digit is, whether the discarded part was an exact tie, and which rounding policy decided the last kept digit. If the plain decimal answer would hide the precision — as 1000 does for three significant figures — the answer is reported in scientific notation instead.
Calculate
Type a whole expression. Each number becomes an operand card you can switch between Measured and Exact, and any operand whose notation is ambiguous gets a control for declaring what you meant. The answer card names the limiting operand (for × and ÷) or the limiting decimal place (for + and −), shows the unrounded result alongside the reported one, and walks through the steps.
What Are Significant Figures?
The significant figures of a measurement are the digits that carry real information about how precisely it was measured. They are also called significant digits, or just sig figs. In 4.56 g there are three; in 0.00456 g there are still three, because the zeros only locate the decimal point.
The idea exists because a written number makes a claim. Reporting a mass as 4.5 g says you know it to about a tenth of a gram; reporting it as 4.500 g says you know it to about a thousandth. Writing more digits than your instrument supports is a false claim of precision, and dropping digits your instrument earned throws away information. Significant figures are the bookkeeping convention that keeps the written number honest — a fast, approximate stand-in for a full uncertainty statement, not a replacement for one.
Significant Figures Rules
Six rules cover every case. The calculator applies exactly these, and its explanation panel names the one it used for each digit.
| Rule | Example | Sig figs | Why |
|---|---|---|---|
| Non-zero digits are always significant | 123.45 | 5 | Every non-zero digit was measured; nothing else could have put it there. |
| Leading zeros are never significant | 0.0045 | 2 | They only place the decimal point. Writing the value as 4.5 × 10⁻³ makes them disappear without changing anything. |
| Captive zeros are always significant | 1002 | 4 | A zero trapped between significant digits cannot be a placeholder — it is a measured value of zero. |
| Trailing zeros after a decimal point are significant | 4.500 | 4 | Nobody writes those zeros unless they mean them; they exist only to declare precision. |
| Trailing zeros in an integer with no decimal point are ambiguous | 1200 | 2, 3 or 4 | The notation cannot tell you whether the zeros are measured or placeholders. See the next section. |
| In scientific notation, every mantissa digit is significant | 1.200 × 10³ | 4 | The exponent carries the magnitude, so the mantissa carries nothing but precision. |
A sign is not a digit: −0.0320 and 0.0320 both have three significant figures. An exact number — a count, a defined conversion factor or a mathematical constant — is outside the system entirely and has unlimited significant figures.
Why Are 100, 1000 and 1200 Ambiguous?
Because plain integer notation has no way to say where the measurement stopped and the placeholders began. If you measure a distance as “about a hundred metres” and write 100 m, you mean one significant figure. If you measure it carefully as 100.0 m and then drop the decimal, the same three characters now understate what you know. Nothing in “100” distinguishes the two.
| Written as | Value | Declares | Verdict |
|---|---|---|---|
1200 | 1200 | nothing — 2, 3 or 4 significant figures | Ambiguous. Read as 2 significant figures by the common classroom convention. |
1200. | 1200 | 4 significant figures | Unambiguous, but the trailing point is easy to miss (and easy to lose in a spreadsheet). |
1.2 × 10³ | 1200 | 2 significant figures | Unambiguous. |
1.20 × 10³ | 1200 | 3 significant figures | Unambiguous. |
1.200 × 10³ | 1200 | 4 significant figures | Unambiguous. |
This calculator refuses to pretend the ambiguity is not there. It reports the common classroom interpretation — trailing zeros in a bare integer are not counted, so 100 reads as one significant figure, 1000 as one and 1200 as two — but it labels that reading as a convention, keeps the value flagged as ambiguous, lists every unambiguous way to write it, and lets you declare which precision you actually meant before the receipt becomes final.
Significant Figures in Calculations
There is no single rule. Multiplication and division count digits; addition and subtraction count decimal places. Using the wrong one is the most common significant-figure error there is.
Multiplication & Division
The result carries as many significant figures as the measured operand with the fewest significant figures. For 4.56 × 1.4 the exact product is 6.384, but 1.4 has only two significant figures, so the answer is reported as 6.4 and 1.4 is the limiting operand. Powers and roots follow from this: raising a measurement to a whole-number power is repeated multiplication by itself, so 2.0³ = 8.0 keeps two significant figures.
Addition & Subtraction
The result is rounded to the leftmost least-significant decimal place among the measured operands — the place where the least precise number ran out. For 18.0 + 1.013 the exact sum is 19.013, but 18.0 is only known to the tenths, so the answer is 19.0. Note what did not happen: 1.013 has four significant figures and 18.0 has three, yet the answer has three — not because of any digit count, but because the tenths place is where the precision stopped.
This is also why subtraction can destroy precision. 100.0 − 99.9 = 0.1: two four-significant-figure measurements produce a one-significant-figure answer, because almost every significant digit cancelled. The calculator flags this rather than quietly reporting a confident-looking number.
Mixed Operations
When the two rules meet in one expression, apply each at its own step and never round in between. For (12.11 + 0.3) × 4.20 the bracket is limited by the tenths place, so it reports as 12.4 — but 12.41 is what enters the multiplication. The product 52.122 is then limited by the three significant figures of the bracket, giving 52.1. Round the bracket first and you inject an error the multiplication then amplifies; the step panel on this page shows the reported intermediate as a view while the unrounded value carries forward.
Exact Numbers vs Measured Numbers
A measured number came from an instrument and carries a limited precision. An exact number did not, and has unlimited significant figures. Three kinds show up constantly:
- Counted quantities. 12 students, 3 trials, 7 samples. You did not measure 12 — you counted it, and there is no such thing as 12.0000001 students.
- Defined conversion factors. 1 inch = 2.54 cm exactly, 1 foot = 12 inches, 1 km = 1000 m. These are definitions, not measurements, so they never limit a converted result.
- Mathematical constants. π and e have infinitely many known digits; the only limit is how many you choose to use. In this calculator they are exact, so πr² with r = 2.50 cm gives 19.6 cm² — three significant figures, set by the radius alone.
Why it matters: 4.56 × 2 gives 9 if the 2 is a measurement (one significant figure) and 9.12 if the 2 is a count. Same arithmetic, different reported precision. In Calculate mode each operand has a Measured/Exact control, so this is a decision you make explicitly instead of one the tool makes for you.
Rounding to Significant Figures
The algorithm is the same whatever the magnitude:
- Find the first significant digit — the leftmost non-zero digit.
- Count the target number of significant figures from there. That last digit is the one you will keep.
- Look at the very next digit. Above 5, round up; below 5, leave it; exactly 5 with nothing after it is a tie, resolved by your rounding policy.
- Replace every discarded digit before the decimal point with a zero placeholder, and simply drop the discarded digits after it.
- Check that the written answer still declares the precision you claimed. If it cannot, switch to scientific notation.
Worked at three targets: 0.0045678 becomes 0.005 at 1 significant figure, 0.0046 at 2 and 0.00457 at 3. Going the other way, 12345 becomes 1.2 × 10⁴ at 2 significant figures — written as 12000 it would read as two, three, four or five, which is exactly the ambiguity the previous section describes.
The same algorithm at a range of targets — every row below is produced by the calculator above:
| Number | Target sig figs | Rounded | What decided the last digit |
|---|---|---|---|
45.5147 | 4 | 45.51 | First discarded digit 4 → round down. |
2.009 | 2 | 2.0 | First discarded digit 0 → round down. |
0.03659 | 3 | 0.0366 | First discarded digit 9 → round up. |
2500 | 3 | 2.50 × 10³ | Zeros added — the input declares less precision than this. Plain decimal form would be ambiguous, so scientific notation is reported. |
12345 | 3 | 1.23 × 10⁴ | First discarded digit 4 → round down. Plain decimal form would be ambiguous, so scientific notation is reported. |
999.5 | 3 | 1.00 × 10³ | First discarded digit 5 (exact tie) → round up. Plain decimal form would be ambiguous, so scientific notation is reported. |
0.0045678 | 3 | 0.00457 | First discarded digit 7 → round up. |
1.2 | 3 | 1.20 | Zeros added — the input declares less precision than this. |
Step 5 is the one most calculators skip. Rounding 999.5 to three significant figures carries through every nine and gives 1000, which plainly written declares one significant figure, not three. The honest answer is 1.00 × 10³, and that is what this page reports.
Half away from zero vs half to even
The two policies differ only on an exact tie. Half away from zero — standard classroom rounding — always pushes a tie away from zero, so 12.345 becomes 12.35 at four significant figures and −2.5 becomes −3 at one. Half to even, also called banker's rounding, sends a tie to the nearest even digit, so 12.345 becomes 12.34 and 2.5 becomes 2; it is the IEEE 754 default because always rounding ties up introduces a small but systematic upward bias across many values. Both are available here, and the receipt always names which one decided the digit.
Significant Figures and Scientific Notation
Scientific notation writes a value as a mantissa between 1 and 10 times a power of ten, and that split is exactly what makes it unambiguous: the exponent carries all the magnitude, so every digit left in the mantissa is there to declare precision. 1.2 × 10³, 1.20 × 10³ and 1.200 × 10³ are the same value with two, three and four significant figures — a distinction plain 1200 simply cannot express.
Two consequences worth remembering. First, converting to scientific notation never changes how many significant figures a number has; it only makes the count visible. Second, an exponent is not a measurement: the 3 in 10³ is exact and contributes nothing to the count.
To convert between the two forms in either direction, use the Scientific Notation Calculator; to round to decimal places rather than significant figures, use the Rounding Calculator.
Logarithms, Powers and Exact Constants
Calculate mode implements four further conventions, and each one is named in the receipt when it fires.
- Powers with a whole-number exponent keep the significant figures of the base, because they are repeated multiplication: 2.50² = 6.25 has three.
- Square roots keep the significant figures of the radicand: √2.0 = 1.4.
- Logarithms carry precision only after the decimal point, so a logarithm is written with as many decimal places as its argument has significant figures: log(1.0 × 10³) = 3.00, because the argument has two.
- Antilogarithms invert that rule: 10x and ex keep as many significant figures as x has decimal places.
π and e are treated as exact and never reduce a result's precision. Where no convention is safe — a measurement raised to a measured power, for instance — the calculator reports the unrounded value and issues a precision notice instead of asserting a significant-figure count it cannot justify. Logarithms, antilogarithms and non-integer powers are evaluated in double precision (about 15 digits) and labelled as such; everything else is exact decimal arithmetic to 40 guard digits.
Quick Answers: How Many Significant Figures?
Every row below is produced by running the calculator above, so the table cannot disagree with the tool.
| Number | Significant figures | Unambiguous form | Why |
|---|---|---|---|
0.06900 | 4 SF | 6.900 × 10⁻² | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
2.025 | 4 SF | 2.025 × 10⁰ | 1 zero sits between significant digits, so it is significant. |
2.00 | 3 SF | 2.00 × 10⁰ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
5.0 | 2 SF | 5.0 × 10⁰ | The 1 trailing zero is significant because the notation declares it — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
5.00 | 3 SF | 5.00 × 10⁰ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
100.0 | 4 SF | 1.000 × 10² | The 3 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
0.01 | 1 SF | 1 × 10⁻² | The 2 zeros before the first non-zero digit only place the decimal point, so they are not significant. |
20.0 | 3 SF | 2.00 × 10¹ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
1.00 | 3 SF | 1.00 × 10⁰ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
0.50 | 2 SF | 5.0 × 10⁻¹ | The 1 trailing zero is significant because the notation declares it — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
0.01030 | 4 SF | 1.030 × 10⁻² | The 1 trailing zero is significant because the notation declares it — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
0.004560 | 4 SF | 4.560 × 10⁻³ | The 1 trailing zero is significant because the notation declares it — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
1002 | 4 SF | 1.002 × 10³ | 2 zeros sit between significant digits, so they are significant. |
1.200e3 | 4 SF | 1.200 × 10³ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
1200. | 4 SF | 1.200 × 10³ | The 2 trailing zeros are significant because the notation declares them — a written decimal point (or the mantissa of scientific notation) is a claim about measured precision. |
100 | ambiguous (1–3); common classroom interpretation 1 SF | 1 × 10² | Trailing zeros with no decimal point declare nothing. |
1000 | ambiguous (1–4); common classroom interpretation 1 SF | 1 × 10³ | Trailing zeros with no decimal point declare nothing. |
1200 | ambiguous (2–4); common classroom interpretation 2 SF | 1.2 × 10³ | Trailing zeros with no decimal point declare nothing. |
Worked Examples
18 examples covering counting, rounding, all four operations, ambiguity, scientific notation, exact numbers and mixed expressions. Each answer is generated by the engine on this page.
-
1. Counting with leading and trailing zeros
0.004560→ 4 significant figuresThe three leading zeros place the decimal point; the final zero is a deliberate claim of precision.
-
2. Captive zeros inside a number
1002→ 4 significant figuresA zero trapped between significant digits cannot be a placeholder, so it always counts.
-
3. Every kind of zero at once
0.01030→ 4 significant figuresLeading zeros out, captive zero in, trailing zero in — the answer is four.
-
4. An ambiguous integer
1200→ 2 significant figuresWithout a decimal point the notation simply does not say how precise 1200 is.
-
5. Scientific notation settles it
1.200e3→ 4 significant figuresThe same value, written so the precision is impossible to misread.
-
6. Rounding a large integer
12345 → 3 SF→ 1.23 × 10⁴The plain answer would end in zeros that cannot declare precision, so it is reported in scientific notation.
-
7. A carry that changes every digit
999.5 → 3 SF→ 1.00 × 10³Rounding up a run of nines lifts the exponent — and makes the plain form ambiguous.
-
8. Rounding a small decimal
0.0045678 → 3 SF→ 0.00457Leading zeros never count towards the target, they only locate the decimal point.
-
9. Asking for more precision than you have
1.2 → 3 SF→ 1.20The zeros can be written, but they assert precision the measurement does not support.
-
10. Multiplication
4.56 * 1.4→ 6.4Raw result
6.384· Multiplication & division — fewest significant figuresLimited by 1.4
The operand with the fewest significant figures decides the answer.
-
11. Division
12.5 / 3.0→ 4.2Raw result
4.16666666666666666666666666666666666666666· Multiplication & division — fewest significant figuresLimited by 3.0
Division uses the same rule as multiplication.
-
12. Addition
18.0 + 1.013→ 19.0Raw result
19.013· Addition & subtraction — least precise decimal placeLimited by 18.0
Place value, not significant-figure count, limits a sum.
-
13. Subtraction and cancellation
100.0 - 99.9→ 0.1Raw result
0.1· Addition & subtraction — least precise decimal placeLimited by 100.0 and 99.9
Four significant figures in, one out — subtraction can destroy precision.
-
14. A measured “2”
4.56 * 2→ 9Raw result
9.12· Multiplication & division — fewest significant figuresLimited by 2
Read as a measurement, a bare 2 has a single significant figure.
-
15. The same “2”, marked exact
4.56 * 2→ 9.12Raw result
9.12· Multiplication & division — fewest significant figuresLimited by 4.56
A counted or defined 2 imposes no limit, so the measurement keeps its precision.
-
16. Mixed operations with guard digits
(12.11 + 0.3) * 4.20→ 52.1Raw result
52.122· Multiplication & division — fewest significant figuresLimited by 0.3 and 4.20
The bracket is reported to three significant figures but 12.41 — unrounded — is what gets multiplied.
-
17. An exact constant
pi * 2.50^2→ 19.6Raw result
19.634954084936206875· Multiplication & division — fewest significant figuresLimited by 2.50
π carries unlimited precision, so the radius alone sets the answer.
-
18. Scientific notation in a calculation
6.022e23 * 2.0→ 1.2 × 10²⁴Raw result
1204400000000000000000000· Multiplication & division — fewest significant figuresLimited by 2.0
Very large and very small measurements obey exactly the same rules.
Frequently asked questions
How many significant figures are in 0.01030?
Four. The two leading zeros only place the decimal point and are not significant. The 1 and the 3 are non-zero, so they count. The zero between them is a captive zero and always counts, and the final zero is significant because it sits after the decimal point, where it can only be there to declare precision. The significant digits are 1, 0, 3 and 0.
Are trailing zeros significant?
Only when the notation declares them. After a decimal point they always are: 4.500 has four significant figures and 0.50 has two. In an integer written without a decimal point they are ambiguous — 1200 could carry two, three or four. Writing 1200., or switching to 1.20 × 10³, removes the ambiguity.
How many significant figures are in 100?
It is genuinely ambiguous: one, two or three. Nothing in the notation says whether the zeros are measured or are just placeholders. The common classroom reading is one significant figure, and this calculator reports that reading as a labelled convention while still flagging the value as ambiguous. Write 1.00 × 10² for three, 1.0 × 10² for two, or 100. for three with an explicit decimal point.
How many significant figures are in 1000?
One, two, three or four — the notation does not say. By the common classroom convention it is read as one significant figure. To be unambiguous write 1 × 10³, 1.0 × 10³, 1.00 × 10³ or 1.000 × 10³, or add an explicit decimal point (1000.) to declare all four.
Does zero count as a significant figure?
It depends entirely on where the zero sits. A zero between significant digits always counts (1002 has four). A zero before the first non-zero digit never counts (0.0045 has two). A zero after the last non-zero digit counts only if a decimal point or a scientific-notation mantissa declares it (0.0400 has three; 400 is ambiguous).
Are exact numbers limited by significant figures?
No. Counted quantities (12 students), defined conversion factors (1 inch = 2.54 cm exactly, 1 foot = 12 inches) and mathematical constants such as π and e carry unlimited significant figures and never limit a result. In Calculate mode, mark an operand Exact and it drops out of the precision decision: 4.56 × 2 gives 9 if the 2 is a measurement, and 9.12 if it is a count.
Do I use sig figs or decimal places for addition?
Decimal places. Addition and subtraction are limited by the leftmost least-significant place among the measured operands, not by significant-figure counts. In 18.0 + 1.013 the sum is 19.013, but 18.0 is only known to the tenths, so the answer is 19.0 — even though 1.013 has more significant figures than 18.0.
When should I round intermediate calculations?
Never. Round once, at the end. Rounding an intermediate result injects an error that the rest of the calculation then amplifies. This calculator keeps every intermediate value at full precision (forty guard digits) and rounds only the reported answer; the step panel shows an intermediate to the correct number of significant figures as a view, while the unrounded value is what actually carries forward.
How are logarithms handled?
Only the part after the decimal point in a logarithm carries precision information, so a logarithm is written with as many decimal places as its argument has significant figures: log(1.0 × 10³) is reported as 3.00, because the argument has two significant figures. Antilogarithms invert the rule — 10^x keeps as many significant figures as x has decimal places. Both are computed in double precision and labelled as such.
Is scientific notation better for showing significant figures?
Yes, and it is the only notation that is never ambiguous. In a × 10ⁿ every digit written in the mantissa is significant, so 1.2 × 10³, 1.20 × 10³ and 1.200 × 10³ state two, three and four significant figures for the same value of 1200. Whenever a rounded answer cannot be written unambiguously in plain decimal form, this calculator reports it in scientific notation instead.
What does 45.5147 rounded to 4 significant figures give?
45.51. The first four significant digits are 4, 5, 5 and 1; the first discarded digit is 4, which is below 5, so the last kept digit stays as it is. The answer's last significant digit sits in the hundredths place.
Why does rounding 999.5 to 3 significant figures give 1.00 × 10³?
Rounding up the tie carries through every 9, so 999.5 becomes 1000 — but written plainly, 1000 reads as one significant figure, not three. Scientific notation is the only way to say the answer is precise to three significant figures, so it is reported as 1.00 × 10³.
What is the difference between half up and half to even rounding?
They differ only on an exact tie, where the discarded part is exactly half. Half away from zero (standard classroom rounding) always rounds a tie away from zero, so 12.345 becomes 12.35 at four significant figures. Half to even (banker's rounding) sends a tie to the nearest even digit, so the same number becomes 12.34; it avoids the small upward bias that always-round-up introduces across many values.
Methodology
Every number you type is read lexically before it is read numerically: the calculator records the raw token, the normalized token, which characters are significant digits, the place value of the last significant digit, the notation used and whether the notation actually declares a precision. Only then is the value built, as an exact decimal (a BigInt coefficient and a power of ten) rather than a floating-point double, so 0.1 + 0.2 is exactly 0.3 in every receipt. Counting follows the standard rules — non-zero digits always count, captive zeros count, leading zeros never count, trailing zeros count only when a decimal point or a scientific-notation mantissa declares them. Trailing zeros in an integer written without a decimal point are reported as a structured ambiguity with an explicit minimum and maximum, alongside a clearly labelled common classroom interpretation that is never presented as the declared precision. Multiplication and division take the fewest significant figures among the measured operands; addition and subtraction are rounded to the leftmost least-significant decimal place among the measured operands; operands marked exact impose no limit at all, and pi and e are exact constants. Intermediate values are carried unrounded to forty guard digits and rounded exactly once, on the reported answer, so display precision can never feed back into the arithmetic. Powers with whole-number exponents keep the base's significant figures, roots keep the radicand's, logarithms take as many decimal places as the argument has significant figures, and antilogarithms take as many significant figures as the exponent has decimal places; a measurement raised to a measured power has no accepted convention, so the calculator reports the unrounded value and declines to assert a precision. Expressions are parsed by a dedicated recursive-descent parser — never eval — and every failure returns a specific error code. The result is a serializable receipt containing the raw result, the final result, the rule applied, the limiting operand or place, the rounding decision and its policy, warnings and steps; the same receipt drives the page, the tests and an independent second implementation that re-derives every counting and rounding decision by different arithmetic.
Assumptions
- A number typed without an explicit 'exact' mark is treated as a measurement, so its written digits declare its precision.
- Trailing zeros in an integer with no decimal point are treated as ambiguous, and the common classroom reading (they are not significant) is reported as a convention, not as the declared precision.
- A whole-number exponent in a power is an exact count of repeated multiplications, and a literal 10 used as the base of a power is the exact base of the decimal system.
- Chained operators at the same precedence level evaluate left to right; exponentiation is right-associative and binds tighter than a leading minus sign.
Limitations
- Expressions are limited to 500 characters, 32 numeric operands and 16 levels of nesting; exceeding a limit returns a structured error rather than a partial answer.
- Significant figures are a bookkeeping convention, not an uncertainty calculation. For real error propagation use explicit uncertainties; this page says so wherever the convention is only an approximation.
- Logarithms, antilogarithms and non-integer powers are evaluated in double precision (about 15 digits) and are labelled as such; addition, subtraction, multiplication, division, integer powers and square roots are exact to 40 guard digits.
- A measurement raised to a measured power has no generally accepted significant-figure rule, so the calculator returns a precision notice instead of a significant-figure count.
- Uncertainty intervals, unit tracking and dimensional analysis are out of scope.
No independent third-party mathematical review yet. The author is responsible for the methodology; correctness is enforced by golden vectors, invariant tests and an independent second implementation that re-derives counting and rounding by different arithmetic and must agree on every published case.
How this calculator is tested
Engine: significant-figures v1.0.0 · Receipt: v1.0.0 · Last updated:
Every build runs 103 golden vectors against this engine — 34 counting cases, 20 rounding cases, 22 arithmetic-propagation cases, 4 ambiguity cases and 23 structured-error cases — plus 6 families of invariant (property) tests covering leading-zero invariance, sign invariance, scientific-notation equivalence, exact-number monotonicity, guard-digit invariance and displayed-result/receipt consistency.
A second, independent implementation re-derives every counting and rounding decision by different arithmetic — a right-to-left digit scanner and exact rational scaling, rather than the engine's left-to-right scan and digit-string carry — and the two are compared on every published case: 307 of 307 cross-checks agree, with 0 disagreements. A single disagreement fails the build.
The 18 rows of the quick-answer table, the 10 example chips and the 18 worked examples on this page are generated by running the same engine, so a published answer cannot drift from the calculator that produced it. Every calculation runs locally in your browser: nothing you type is transmitted.
Your recent calculations
Calculations you save are listed here, stored only in this browser — never sent to a server. At most 20 are kept.
Sources
References for the reporting conventions, the rounding rules and the measurement context (these document the method — they are not input datasets):
- NIST Special Publication 811 — Guide for the Use of the International System of Units (SI), including its guidance on rounding numerical values and on how many digits to report
- NIST Technical Note 1297 — Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results
- NIST Reference on Constants, Units and Uncertainty — Uncertainty of Measurement Results
- JCGM 100:2008 (BIPM) — Evaluation of measurement data: Guide to the Expression of Uncertainty in Measurement (GUM), the primary reference for why significant figures are a reporting convention rather than an uncertainty calculation
- ISO 80000-1:2009 — Quantities and units, Part 1: General, which sets out the rules for rounding numerical values
- IEEE 754-2019 — Standard for Floating-Point Arithmetic, whose default roundTiesToEven rule is the “half to even” policy offered here
- OpenStax, Chemistry 2e §1.5 — Measurement Uncertainty, Accuracy, and Precision (the classroom statement of the counting and arithmetic rules)