Evaluate
for a non-negative integer .
Split off one sine and integrate by parts. Write and take
The boundary term vanishes at both ends — at the top and at the bottom — which is exactly why these limits are chosen.
Convert back to . What remains is
so
The original integral reappears on the right, which is the signal to solve for it algebraically rather than integrate again.
Solve for to get the reduction formula. Adding to both sides gives , hence
Each application lowers the index by two, so the recursion terminates at or depending on the parity of .
Compute the two base cases.
These are the sole source of the : it survives only in the even chain.
Unroll the recursion into the Wallis formulas.
For example and .
Note the single formula that covers both parities. Using the Gamma function,
Numerical integration for through agrees with this to ten decimal places in every case ✓, and it extends the result to non-integer .
Need to solve a different problem like this? Open the solver →