Overview of Cross-Division
Cross-division is an analogue of cross-multiplication that can be used to calculate an arbitrary number of digits of a quotient, even when the divisor is too large to work with directly.
We illustrate the method with:
12,345,678 ÷ 4,321
Why This Works
Cross-division is essentially cross-multiplication run in reverse.
It relies on the same distributive principles, but instead of building a product digit-by-digit, we remove the effect of the divisor digit-by-digit.
1. Set Up the Numbers
For cross-division, it does not matter whether the problem is written vertically or horizontally:
12,345,678
÷ 4,321
or
12,345,678 ÷ 4,321
Each digit of the dividend interacts with each digit of the divisor, but instead of column-by-column subtraction, we work along place-value diagonals, just as in cross-multiplication.
2. Split the Divisor into Two Components
Take the leftmost digits of the divisor that you are comfortable multiplying by a one-digit number.
Most people start with two digits, though starting with a single digit is also common.
In this example, the divisor 4321 can be split as:
4 | 321 (one-digit start), or
43 | 21 (two-digit start)
Unlike cross-multiplication, using smaller starting groups increases the likelihood of needing a correction later, but makes the early steps easier.
3. Calculating the First Digit (Using Two Digits)
The dividend has 8 digits, the divisor has 4 digits, and the leftmost digits of the dividend are smaller than those of the divisor.
Therefore, the integer component of the quotient will have 4 digits.
Take the leftmost digits of the dividend: 123
Now estimate the largest digit y such that:
43 × y < 123
The largest such value is y = 2.
So the first digit of the quotient is 2.
Subtract:
123 − (2 × 43) = 37
4. Calculating Subsequent Digits
From here on, we repeat a cycle:
Bring down the next digit of the dividend
Apply cross-multiplication corrections
Estimate the next digit of the quotient
Second Digit
Bring down the next digit (4) → 374
Apply the cross-multiplication correction:
Multiply the first quotient digit (2) by the first digit of the second component (2) → 4
Subtract: 374 − 4 = 370
Now estimate:
43 × y < 370 → y = 8
Subtract:
370 − (8 × 43) = 26
Multiply by 10 and bring down the next digit → 265
Apply cross-multiplication corrections:
8 × 2 = 16
2 × 1 = 2
Subtract:
265 − 16 − 2 = 247
Third Digit
Estimate:
43 × y < 247 → y = 5
Subtract:
247 − (5 × 43) = 32
Multiply by 10 and bring down the next digit → 326
Apply corrections:
5 × 2 = 10
8 × 1 = 8
326 − 10 − 8 = 308
Fourth Digit
Estimate:
43 × y < 308 → y = 7
Subtract:
308 − (7 × 43) = 7
At this point, we have the integer part of the quotient:
2857
Decimal Digits
Multiply by 10 and continue:
7 → 77
Apply corrections:
7 × 2 = 14
5 × 1 = 5
77 − 14 − 5 = 58
Since 43 × 1 < 58, the first digit after the decimal point is 1.
The quotient is now:
2857.1
This process can continue indefinitely if the decimal expansion repeats.
5. What If the Initial Estimate Is Incorrect?
Estimation errors (or overflows) will occur more frequently when starting with a single digit. Fortunately, they are easy to fix.
Suppose we had started with 4 | 321 instead of 43 | 21.
Correcting an Overflow
Start: 4 × y < 12 → y = 2
12 − (2 × 4) = 4
Bring down next digit → 43
Apply correction: 3 × 2 = 6
43 − 6 = 37
Continue:
4 × y < 37 → y = 8
37 − (8 × 4) = 5
Bring down next digit → 54
Apply corrections:
8 × 3 = 24
2 × 2 = 4
54 − 24 − 4 = 26
At this point, one might incorrectly try y = 6, but the correct digit is 5.
Proceeding with 6 produces a negative result, immediately signaling an overflow.
To correct:
Reduce the estimate from 6 to 5
Adjust using the divisor components
Resume from the last correct remainder
This correction is quick and local — no prior work is lost.

