I'm skeptical of that "proof". Specifically, for example, one of the first statements in the proof says:

> We ask our question, call it question 1: “Does there exist a strategy that statistically significantly (after accounting for possible data mining) makes money (after accounting for transactions costs)?” An answer to that question essentially tests each of the possible 3n strategies.

which is non-sequitur. It's like asserting the only way to find the shortest path between two vertices in a graph is to exhaustively check every path, because waves hand.

Also note that paper doesn't appear to have undergone any form of peer review.