Roman Numeral Converter

Your inputs

Convert a numeral

Direction

Switching carries the current answer into the field below.

Number to convert

Any whole number from 1 through 3,999 — the range the seven letters cover.

Your inputs are calculated locally and are not stored.

2,026 in Roman numeralsMMXXVI

MM (2,000) + XX (20) + V (5) + I (1) = 2,026.

Roman numeral: MMXXVI
Number: 2,026
Symbol groups: 4
Subtractive pairs: None — every symbol here is added

How MMXXVI breaks down, largest first
Group	Value	Form
MM	2,000	Additive — repeated symbols
XX	20	Additive — repeated symbols
V	5	Additive — repeated symbols
I	1	Additive — repeated symbols

Formula & methodology

One table going out, one round trip coming back

Writing a number as a numeral has exactly one right answer, and one short algorithm finds it. Take a table of values in descending order — with the six subtractive pairs slotted in at the value they stand for — and repeatedly subtract the largest value that still fits, writing its symbol each time.

M 1000  • CM 900  • D 500  • CD 400  • C 100  • XC 90  • L 50  • XL 40  • X 10  • IX 9  • V 5  • IV 4  • I 1

The seven letters: I 1, V 5, X 10, L 50, C 100, D 500, M 1000
The six pairs: CM, CD, XC, XL, IX, IV — only I, X and C ever subtract
Repetition limit: Three of any symbol in a row; the pair replaces the fourth

Reading a numeral is where converters disagree with each other. The usual parser scans the symbols and subtracts any that is smaller than the one after it. That reader is happy to tell you IIII is 4, VV is 10 and IC is 99 — and none of those are Roman numerals. You can bolt on rules to catch them: no symbol four times, only I, X and C subtract, and only from the next two values up. Do that and you have five conditions and a sixth case nobody thought of, such as XIIX for 18.

This converter uses one rule instead. It reads the numeral to a number, writes that number back out in standard form, and requires the two strings to match exactly. Because the writer has precisely one way to spell each number, a non-standard numeral cannot survive the trip. IIII reads as 4 and comes back IV. VV reads as 10 and comes back X. IC reads as 99 and comes back XCIX. All three fail the same comparison, and the rejection message can name the correct spelling because it has just computed it.

Worked example

2026 written out, and MCMXCIV read back

Start with 2026 and walk down the table. M is 1,000 and fits twice, taking 2,000 and leaving 26 — write MM. CM, D, CD and C are all larger than 26, so skip them. XC and L are too, and so is XL. X is 10 and fits twice, taking 20 and leaving 6 — write XX. IX is 9, too large. V is 5 and fits once, leaving 1 — write V. IV is too large. I fits once — write I. The answer is MMXXVI, four groups totalling 2,000 + 20 + 5 + 1.

Now the other direction, with MCMXCIV. Group it first: M stands alone at 1,000. C is followed by a larger M, so CM is a pair worth 900. X is followed by a larger C, so XC is a pair worth 90. I is followed by a larger V, so IV is a pair worth 4. Total 1,994 — the year that appears at the end of a great many film credits.

And a failure, because that is the interesting case. Type IC and this tool reads it as 99, exactly as a naive parser would. Then it writes 99 back out: XC takes the 90, IX takes the 9, giving XCIX. XCIX is not IC, so IC is rejected — with the correct spelling in the message. The same happens to MMMM, which reads cleanly as 4,000 and then has nowhere to go, because 4,000 has no spelling in the seven letters at all.

The 3,999 ceiling

Why the range stops where it does

M is the largest of the seven letters, worth 1,000, and no symbol may be repeated more than three times. That caps the thousands at MMM, and MMMCMXCIX — 3,999 — is the largest number the ordinary alphabet can spell. There is no letter for 5,000 to play the role V plays for 5 and L plays for 50.

Two devices exist for going higher, and neither survives being typed into a text box. The vinculum is an overbar meaning multiply by a thousand, so V with a bar is 5,000 and X with a bar is 10,000; in digital text that bar is a combining mark that renders inconsistently across fonts and is routinely lost when text is copied, pasted or indexed. The apostrophus forms are older still, and Unicode does encode them — ↁ at U+2181 is five thousand, ↂ at U+2182 is ten thousand — but they are absent from most fonts and from every keyboard. A converter that emitted either would be handing back strings its own users could not retype, so this one stops at 3,999 and says why.

The famous exception

Why clock faces so often show IIII

Look at an analogue clock with Roman numerals and there is a good chance the 4 is IIII rather than IV. This is not an error and it is not modern: it is a horological convention centuries old, and it is worth knowing about precisely because it contradicts the rule this tool enforces.

The explanation most watchmakers give is about the look of the dial. Four o’clock sits directly opposite eight o’clock, and VIII is a wide, heavy numeral; a lone IV facing it looks thin and the dial reads as lopsided, whereas IIII carries similar visual weight. It also divides the face tidily: the first four numerals are all built from I, the next four introduce V, and the last four introduce X.

Two other explanations circulate widely and are much weaker. One says a set of numeral blanks for a dial casts out neatly as twenty I, four V and four X only if 4 is IIII — a nice arithmetic coincidence, but no evidence that any workshop chose the convention for it. The other says early makers avoided IV because it opens the name IVPITER; the story is late, and Roman inscriptions themselves are inconsistent about IV and IIII long before any clock existed. What is certain is that the convention is not universal — the Great Clock at Westminster, the one whose bell is Big Ben, uses IV.

This converter returns IV, because a converter that returned two spellings of the same number would be useless for checking anything. If you are laying out a clock dial, use IIII and be in excellent company.

Assumptions

What this converter assumes

Standard modern notation, as taught and as used on monuments, book prefaces, film copyright lines and monarch numbering. This is a stricter system than the Romans themselves used — surviving inscriptions contain plenty of IIII, XXXX and other forms that are rejected here.
The range is 1 to 3,999. There is no zero and there are no negatives, because the notation has neither.
Whole numbers only. The Roman twelfths system for fractions is a separate notation with its own symbols and is not offered.
Lower case input is accepted and normalised — mcmxciv is read as MCMXCIV. Output is always upper case, using the ordinary ASCII letters rather than the dedicated Unicode numeral characters at U+2160 and above, so the result can be copied anywhere.
Anything that is not one of the seven letters is rejected outright, including spaces inside the numeral and the digit 1 typed in place of I.

Common questions

Roman numeral FAQ

What is 2026 in Roman numerals?

MMXXVI. Two thousand is MM, twenty is XX, five is V and one is I, and Roman notation simply writes the groups in descending order: MM + XX + V + I. The year before it, 2025, is MMXXV; the year after, 2027, is MMXXVII.

Why is 4 written IV and not IIII?

Because standard notation allows a symbol to repeat at most three times, and gives you a shorthand instead: put a smaller symbol before a larger one and it is subtracted. IV is 5 − 1. Only six such pairs exist — CM (900), CD (400), XC (90), XL (40), IX (9) and IV (4) — because only I, X and C may subtract, and each may subtract only from the next two values above it. That rule is what makes IC wrong for 99: I can only stand before V or X, so 99 is XCIX, which is (100 − 10) + (10 − 1).

Then why do clock faces show IIII?

Because clock dials follow a horological convention rather than the notation. IIII is a real and very old exception, and the explanation most watchmakers give is visual: IIII sits opposite VIII on a dial, and four heavy strokes balance the eight-stroke numeral across the face, while a lone IV looks thin next to it. It also makes the left half of the dial all-I numerals and the right half introduce V and X, which reads as a tidier progression. Two other stories circulate — that a set of numeral castings comes out neatly as twenty I, four V and four X only if 4 is IIII, and that early makers avoided IV as the opening letters of IVPITER — and neither is well evidenced. The convention is not universal: the Great Clock at Westminster, the one behind Big Ben, uses IV.

Why does this converter stop at 3,999?

Because 3,999 — MMMCMXCIX — is the largest number the seven ordinary letters can spell. M is the biggest of them at 1,000, and no symbol may repeat more than three times, so MMM caps the thousands at 3,000. Anything larger needs the vinculum, an overbar meaning × 1,000, or the archaic apostrophus characters Unicode encodes at U+2181 and U+2182. Neither has a dependable plain-text form: overbars are combining marks that render inconsistently and disappear when text is copied, and the apostrophus characters are missing from most fonts.

How do I read a long numeral like MCMXCIV?

Left to right, in groups: M is 1,000, CM is 900, XC is 90, IV is 4 — total 1,994. The trick is to spot the subtractive pairs first. Whenever a symbol is followed by a larger one, the two belong together and the first is subtracted; otherwise each symbol simply adds. Once you have grouped them, the values only ever descend.

Is there a Roman numeral for zero or for a fraction?

Not in the notation. Roman arithmetic had no zero symbol at all; medieval scribes writing in Latin used the word nulla when they needed one, which is a word rather than a numeral. Fractions were written in a separate twelfths system using S for semis (a half) and dots for each additional twelfth, with its own symbols and rules. Neither is offered here.

Primary sources

Sources and review notes

Nothing on this page can go stale: the seven letters and the six subtractive pairs have been fixed for centuries, and the arithmetic is exact integer work. The one genuinely contested item is why clock dials use IIII, which is treated above as the open question it is.