Exact Fractions vs Rounded Decimals in Multi-Step Calculations

Fractions preserve rational relationships

One third is exactly 1/3. Decimal 0.333 is slightly smaller. Adding three exact thirds gives one; adding three copies of 0.333 gives 0.999. This difference is caused by early rounding, not by an inconsistency in fraction arithmetic.

Some decimals terminate exactly

A reduced fraction has a terminating base-ten decimal only when its denominator’s prime factors are two and five. One eighth is 0.125 and one twentieth is 0.05. One sixth still contains factor three in its denominator and repeats.

Reduce after the exact operation

For 1/6 + 1/4, use denominator 12: 2/12 + 3/12 = 5/12. The result is approximately 0.416666…. Adding 0.17 and 0.25 instead gives 0.42, already rounded differently. Cross-canceling factors before multiplication keeps large intermediate integers smaller without changing the value.

Separate mathematical rounding from binary storage

Even a decimal that terminates in base ten may not be represented exactly by binary floating point. A calculator can avoid this for integer fractions by using exact integer numerators and denominators. A decimal display remains an approximation unless it fully represents the rational value.

Round at a declared boundary

Currency, physical measuring and reporting may require finite decimals. State the rounding rule and precision, and apply them at the relevant boundary. Keep exact intermediate values when the later calculation needs them. Do not report more digits than the underlying measurement supports merely because an exact fraction is available.

Quantify the error rather than just noticing it

Rounding one third to 0.33 creates an error of 1/3 − 33/100 = 1/300, about 0.003333…. Multiplying that rounded value by three creates total error 1/100 = 0.01. Keeping the exact fraction until after multiplication returns one exactly, which can then be displayed as 1.00.

When a displayed decimal is used as a fresh input, its rounding becomes part of the new mathematical record. A later calculator cannot reconstruct the omitted repeating digits automatically. Keep the numerator and denominator, or sufficient unrounded data, when transferring between stages.

Cross-cancel before multiplying large fractions

For 14/15 multiplied by 25/21, cancel fourteen and twenty-one by seven to get two and three, then cancel twenty-five and fifteen by five to get five and three. The product is 10/9. Direct multiplication gives 350/315, which reduces to the same value, but the intermediate integers are larger.

Exact arithmetic avoids floating-point integer loss, yet reducing intermediates still makes calculations easier to audit and can reduce work. Cross-cancellation is valid for multiplication; cancelling unrelated terms across a sum such as (14 + 25)/(15 + 21) would be a different and incorrect operation.

Currency adds a rounding boundary, not a new fraction identity

An exact share of $10 divided by six is 5/3 dollars, or 166.666… cents. Physical payments need integer cents. Four people can pay $1.67 and two $1.66, totaling $10 exactly. Rounding all six shares to $1.67 would collect $10.02.

Fractions describe the exact share before allocation. The quotient/remainder cent split decides how to meet a fixed payable total under indivisible cents. It is not evidence that 5/3 suddenly equals either 1.66 or 1.67; those are the practical allocated amounts.

Exact representation does not imply exact measurement

A measured length recorded as 2.35 meters can be written exactly as 47/20 in the arithmetic, but the measuring instrument may have uncertainty. Rational notation preserves the recorded number; it does not remove that uncertainty. Choose exact fractions for symbolic proportions and exact integer counts, keep decimals at stated reporting boundaries, and describe physical results with precision supported by the source measurements. This separates arithmetic error, display rounding and measurement uncertainty.