Note that xkcd 759 is about something subtly different: you work from both ends and then, when they don't meet in the middle, try to write the "solution" in such a way that whoever's marking it won't notice the jump.

I know someone who did that in an International Mathematical Olympiad. (He used an advanced variant of the technique, where you arrange for the jump to occur between two pages of your solution.) He got 6/7 for that solution, and the mark he lost was for something else. (Which was in fact correct, but you will appreciate that no one was inclined to complain about it.)