Negative order of accuracy

294 Views Asked by At

Suppose we analyze the order of accuracy of a finite difference approximation of a derivative,

$$f'(x)=\frac{1}{2h} \left[f(x-2h) -4f(x-h) +3f(x)\right]$$

and we conclude that the order of accuracy is 1 because our analysis would indicate that there is one $O(h)$ term.

Now, what if instead of the above formula we tried to analyze

$$f'(x)=\frac{1}{2h} \left[f(x-2h) -4f(x-h) +4f(x)\right]$$

Then our analysis would give us an additional (to $O(h)$) term, $\frac{f(x)}{2h}$, or $O(h^{-1})$.

What would this signify in terms of the order of accuracy?