√(2π·10)·(10/e)¹⁰ prints 3.598695618741⋅10+6, and 10! itself prints 3.6288⋅10+6 — 3,628,800. The estimate is short by 0.8296 %, and the formula's own error term, 1/(12n), predicts 0.8333 % for n = 10: a rule that tells you how wrong the rule is, to within four thousandths of a percentage point.
The first factor alone, √(2π·10), prints 7.926654595212, and it is worth separating because it is the part that came from Stirling's and de Moivre's real question — not the size of n! but the constant in front of the growth. Multiplying the estimate by e^(1/120), about 1.0083681, lands within eleven of the exact 3,628,800. For the argument to be useful at all it has to be tested on a factorial small enough to compute: at n = 10 the exact value is one keystroke, and at n = 100 it is 158 digits nobody can write down.
Answer
3.598695618741⋅10+6
How to read it
√(2π·10)·(10/e)¹⁰ on the workbench. The second link is the exact 10!, and the third is the √(2πn) factor by itself.
Source
Stirling's formula n! ~ √(2πn)·(n/e)ⁿ, from Abraham de Moivre's 1730 asymptotics and published by James Stirling in 1730; the relative error is about 1/(12n), and the full expansion is n! ~ √(2πn)(n/e)ⁿ·e^(1/(12n) − 1/(360n³) + …)
Work it in the calculator
These links open the right calculator with this problem's numbers already typed in. The line above is what the screen should show — check it yourself.