The java.lang.Math package ships with several methods related to Euler’s number (e ≈ 2.71828) that are easy to overlook. Two of them, Math.exp(x) and Math.expm1(x), come up all the time in statistics, machine-learning loss functions, interest calculations, and scientific code. In 2026, most hot-path numeric code still ends up calling these static intrinsics because the JIT compiles them down to single CPU instructions on modern JDKs.
public class MathMethods {
public static void main(String[] args) {
new MathMethods().proceed();
}
private void proceed() {
// Math.exp returns e^x
System.out.println(Math.exp(5));
// Math.expm1 returns e^x - 1, more accurate for small x
System.out.println(Math.expm1(5));
}
}Expected output:
148.4131591025766
147.4131591025766When to prefer expm1 over exp
For very small values of x, Math.exp(x) - 1 loses precision because you are subtracting two nearly-equal doubles. Math.expm1(x) computes e^x – 1 directly and keeps the significant digits. The same trick applies to Math.log1p(x), which is the partner function for computing log(1 + x) accurately near zero. Reach for the “1p” and “m1” variants in any financial or probabilistic code where small-x behavior matters.
Charlie has over a decade of experience in website administration and technology management. As the site admin, he oversees all technical aspects of running a high-traffic online platform, ensuring optimal performance, security, and user experience.
