I stand corrected. I still don't understand why, but the graph in figure 7.20 of Wizo's link shows the effect. The effect of the headwind more than compensates for the decreased climb performance when at less than "the best" nil-wind Vx. Probably similar to the max range speed increasing in a headwind.
I learn something every day!
