Average Error: 58.5 → 0.7
Time: 4.7s
Precision: binary64
\[\]
↓
\[\]
↓
double code(double x) {
	return ((double) (((double) (1.0 / 2.0)) * ((double) log(((double) (((double) (1.0 + x)) / ((double) (1.0 - x))))))));
}

↓
double code(double x) {
	return ((double) (((double) (1.0 / 2.0)) * ((double) (((double) (x * ((double) (2.0 * x)))) + ((double) (((double) log(1.0)) + ((double) (2.0 * ((double) (x - ((double) (((double) (x / 1.0)) * ((double) (x / 1.0))))))))))))));
}

Error
Bits error versus x
Try it out
Your Program's Arguments
Results
Enter valid numbers for all inputs
Derivation
  1. Initial program 58.5
    \[\]
  2. Taylor expanded around 0 0.7
    \[\leadsto \]
  3. Simplified0.7
    \[\leadsto \]
  4. Final simplification0.7
    \[\leadsto \]
Reproduce
herbie shell --seed 2020192 
(FPCore (x)
  :name "Hyperbolic arc-(co)tangent"
  :precision binary64
  (* (/ 1.0 2.0) (log (/ (+ 1.0 x) (- 1.0 x)))))