The Roman Numeral Converter converts in both directions: whole numbers from 1 to 3,999 into standard Roman numerals, and valid Roman numerals back into numbers, showing the step-by-step build either way so the result can be checked, learned from, or pasted into the document, tattoo, book chapter, or film credit that needs it.
How It Works
Thirteen symbols carry the whole system: the seven base values (I 1, V 5, X 10, L 50, C 100, D 500, M 1,000) and the six subtractive pairs (IV 4, IX 9, XL 40, XC 90, CD 400, CM 900). Number-to-Roman works greedily from the largest symbol down, taking each as many times as it fits. Roman-to-number first validates the input against the canonical pattern (at most three repeats, subtractive pairs only where standard allows), then reads left to right, adding each symbol's value, except a symbol followed by a larger one, which is subtracted as a pair.
Parameters
| Input | Meaning | |---|---| | Mode | Number to Roman, or Roman to number | | Value | A whole number 1 to 3,999, or a valid Roman numeral (any case) |
Worked examples
1,994 builds as M (1,000) + CM (900) + XC (90) + IV (4) = MCMXCIV, and the converter's step list shows each subtraction as it happens. 2026 is MMXXVI. 3,888 is the marathon case, MMMDCCCLXXXVIII, the longest numeral in range at fifteen characters, a good demonstration of why positional notation won. Backwards, XLII parses as XL (50 - 10 = 40) + I + I = 42, and CDXLIV reads CD (400) + XL (40) + IV (4) = 444, three subtractive pairs in one numeral, which is precisely the pattern the validator exists to enforce: forms like IIII, VX, or IL look plausible and are not canonical, and the converter rejects them with an explanation rather than guessing.
The subtractive rule, and its one famous exception
Standard notation never repeats a symbol more than three times; instead, one-before forms handle 4, 9, 40, 90, 400, and 900. Only I, X, and C may subtract, and each may precede only the next two larger symbols (I before V and X; X before L and C; C before D and M), which is why 49 is XLIX and never IL. The famous exception is on clock faces, where IIII traditionally appears instead of IV, for visual symmetry against VIII and by watchmaking custom: it is a deliberate archaism, and the converter follows the canonical rule rather than the clockmaker's, so IIII is rejected here. Lowercase input is accepted and normalized, matching the convention of front-matter page numbers (xiv, xxiii) in books.
Range limits and what lies beyond them
The classical system tops out naturally at 3,999 (MMMCMXCIX): there is no symbol above M, and canonical repetition caps at three. Medieval and later scribes extended it with the vinculum, a line over a numeral multiplying it by 1,000, so 10,000 could be written with an overlined X; the converter does not implement overlines, in line with most modern usage. There is likewise no zero (the concept postdates the system; medieval writers used the word nulla or the later N), no negatives, and no fractions in this tool, though Roman commerce used duodecimal fractions with their own symbols. Dates are the commonest modern use: copyright years, building cornerstones, regnal numbers, Super Bowls, and outline numbering all live comfortably inside 1 to 3,999, which is why the range rarely binds in practice.
Frequently Asked Questions
Why does the converter reject IIII if clocks use it?
Because it validates the canonical form taught and used in modern standard notation, where 4 is IV. Clock IIII is a preserved traditional variant, not the general rule, and accepting every historical variant would make validation meaningless (IIII, VIIII, and IL would all have to pass as spellings of numbers that already have one standard form).
What is the Roman numeral for 2026?
MMXXVI: MM for 2,000, XX for 20, VI for 6. Zero needs no symbol because Roman notation has no place-holding zero; empty positions are simply absent, which is exactly why arithmetic in this system was so laborious.
Can I convert numbers above 3,999 another way?
With the vinculum convention (overline multiplies by 1,000) any number can be written, but support in fonts and tools is inconsistent, which is why this converter, like most, stops at the canonical 3,999 and says so rather than producing output other systems cannot read back.