Why is central differencing usually more accurate than forward differencing for smooth functions at an interior point?
MathematicsLevel: IntermediateDerivative and probability numerics
Why is central differencing usually more accurate than forward differencing for smooth functions at an interior point?
Hints
- Expand both sides in Taylor series.
- Odd terms behave differently.
Answer
Symmetric evaluation cancels the leading first-order truncation term, giving O(h²) rather than O(h) error, at additional evaluation cost.
Source and licence: SciPy — numerical differentiation · BSD-3-Clause · answer: Quantoros editorial