Repeating As A Fraction

The Method

You set x equal to the repeating decimal, multiply both sides by 10 raised to the power of however many repeating digits there are, subtract the original equation from the new one, and solve for x. That's it. The subtraction eliminates the infinite tail and leaves you with a finite equation. Take 0.333... Multiply by 10, get 10x = 3.333... Subtract the original x = 0.333..., and you're left with 9x = 3. So x = 3/9 = 1/3. The mechanics are almost trivial once you've done it a handful of times. The part people mess up is figuring out what power of 10 to multiply by. It's always 10^n where n is the length of the repeating block. One repeating digit, multiply by 10. Two repeating digits, multiply by 100. Three, multiply by 1000. That's the rule, and it's the rule that trips people up most often.

Converting Repeating As A Fraction Through Algebra

Here's a slightly less trivial example. 0.121212... The repeating block is "12", which is two digits, so n = 2. Multiply by 100: 100x = 12.1212... Subtract x = 0.1212... and you get 99x = 12. So x = 12/99, which simplifies to 4/33. The denominator is always a string of 9s matching the length of the repeating period. That's not a coincidence, it's just the algebra working the same way every time. Now the mixed case, which is where most tutorial examples stop and where things get a little more real. 0.4555... Here the "4" doesn't repeat, only the "5" does. You can't just multiply by 10 or 100 and line things up cleanly in one step. What you do is multiply by 10 to shift past the non-repeating part, getting 10x = 4.555..., then multiply by 100 to shift by one full repeating cycle, getting 1000x = 455.555..., and subtract. 1000x - 10x = 455.555... - 4.555..., which gives 990x = 451. So x = 451/990. Simplify if you can. In this case it doesn't simplify nicely, and that's fine — the fraction is exact, the decimal is just ugly. The general formula for a mixed case is: numerator = (the entire number without the decimal point, minus the non-repeating part) and denominator = as many 9s as there are repeating digits, followed by as many 0s as there are non-repeating digits after the decimal. For 0.4555..., that's (455 - 4) / 90 = 451/90... wait, let me recheck. Non-repeating part is one digit ("4"), repeating part is one digit ("5"). Denominator should be one 9 and one 0, so 90. Numerator is 45 minus 4, which is 41. So x = 41/90. Let me verify: 41 divided by 90 is 0.45555... Yes. That's correct. The earlier calculation with 990 was wrong because I set up the multiplication steps unnecessarily. The shortcut works, but only if you count the digits right.

I spent about twenty minutes last year going back and forth on a problem where the repeating decimal was 0.167167167... and I kept getting 167/999 instead of the simplified form. The issue wasn't the method, it was that I didn't check whether the fraction reduced. 167/999 does reduce — 167 is prime, but 999 = 27 × 37, and 167 doesn't divide evenly into either, so actually it doesn't reduce. I'd convinced myself it did because 167 seemed like it should share a factor with 999, and it doesn't. The takeaway is: always check the GCD before declaring a fraction fully simplified. Use Euclid's algorithm. It takes ten seconds and saves you from carrying around unreduced fractions that look wrong even when they're technically correct.

Where This Breaks Down

Repeating As A Fraction works cleanly for any decimal that has a purely periodic or eventually periodic expansion. That covers all rational numbers. The method fails when the decimal doesn't repeat at all — irrational numbers like pi or sqrt(2) have no repeating block, so there's no finite fraction that represents them exactly. You'll see people online try to force it anyway and end up with approximate fractions that look convincing until you multiply them back out. Another edge case that catches people: what if the repeating block is very long? Say you have a decimal with a 24-digit repeating period. The algebra still works, but your denominator becomes a string of 24 nines, which is 10^24 - 1. That's an enormous number, and simplifying the resulting fraction requires computing the GCD of two numbers that large. Most calculators and spreadsheet programs will round or overflow. You'd need a computer algebra system or arbitrary-precision arithmetic library to handle it properly. I ran into this when someone asked me to convert a repeating decimal that came from a modular arithmetic problem — the period was 42 digits long, and every tool I tried either gave a wrong answer or crashed. I ended up writing a short Python script using the fractions module, which handled it in about three seconds. If you're dealing with long periods regularly, don't trust your calculator. Use symbolic arithmetic. There's also the case of recurring decimals where the repetition starts immediately after the decimal point but the block contains internal zeros, like 0.01001001... The method still applies, but it's easy to miscount the period length if you're not careful. The repeating block here is "010" — three digits — not "10" or "001". Count the repeating pattern, not the non-zero digits. That one costs people points on tests more often than you'd think.

Quick Reference for Common Conversions

Some repeating decimals come up so often that memorizing the fraction saves you from doing the algebra every time. 0.333... = 1/3. 0.666... = 2/3. 0.142857 repeating = 1/7. 0.090909... = 1/11. 0.009009... = 1/111. 0.121212... = 4/33. 0.272727... = 3/11. The denominators follow a pattern: single-digit repeats give you 9, 99, 999, and so on, while mixed cases introduce trailing zeros. When in doubt, go back to the algebra. The pattern is useful for speed, but the algebra never lies. One more thing nobody emphasizes enough: the result of this process is always exact. Unlike rounding or floating-point approximation, the fraction you get represents the number perfectly. If you need to do arithmetic with repeating decimals — add them, multiply them, compare them — converting to fractions first usually makes the work faster and more accurate. I convert repeating decimals to fractions before doing anything else with them, and it cuts down on errors significantly. You'll make fewer mistakes comparing 1/7 and 1/9 than you will comparing 0.142857... and 0.111111... by eye.