I'm not a lawyer, but I am a programmer. I would never swear that my software works so correctly that I will use the actions of the application as proof that a user did check a box. If he claims he didn't, well, he just may be right. I've seen enough weird things happen not to trust any software to behave correctly all the time.
Now my technical reasons may not be relevant for a judge, but I doubt judges are very trustworthy of technology (they should have heard plenty of cases where technology failed and they will be users of all kinds of awful consumer technology, that sometimes, suddenly fails for no clear reason) and won't think favorably about a company with these kinds of practices. Perhaps I'm naive, but those things combined will lead a judge to believe you over the company.
That one, single checkbox, and the fact that you have to
manually check it, is the legality upon which the entire