I was using This Wolfram Alpha Page to calculate the partial derivatives of the sum of two gaussians. It believes the answer to be: $$ e^{-(b-x)^2/2 c^2} $$ but the working shows it as ending on: $$ e^{-(x-b)^2/2 c^2} $$ Which is what I had thought the answer was.
Is this a bug in Wolfram's engine or have I missed an obvious step?
It:$$(x-b)^{2} = (b-x)^{2} \ \ \forall b,x \in \mathbb{R}$$