Would it be better to trade the “Curve Fitted” strategy or the “IN-Sample” strategy?

A minor tweak to the “IN-Sample” strategy creates a nearly perfect “Curve Fitted” result.

Here are the 2 equity curves.

In-Sample strategy results (in the red box). Curve Fitted Strategy (no red box)

SidewalkAerobics 

Would it be better to trade the “Curve Fitted” strategy or the “IN-Sample” strategy? A minor tweak to the “IN-Sample” strategy creates a nearly perfect “Curve Fitted” result. Here are the 2 equity curves. In-Sample strategy results (in the red box). Curve Fitted Strategy (no red box)

Conventionally the in-sample vs out-of-sample divide isn't necessary once you're done with all your hypothesis testing, at which point you can curve fit to all of the data or any subset of the out-of-sample, except you don't test any hypotheses based on that.