Trite phrases become trite through exhaustively correcting common errors. If the error wasn't common, or (more to the point) wasn't actually an error, it wouldn't be trite.
So, no, the plural of trite phrases is not rebuttal. But when the phrase in question is a rebuttal, then the triteness is irrelevant; it's still a rebuttal.
But if you want a slightly more long-winded rebuttal, here you go:
Statistics never prove anything. Seriously. At their best, when properly applied, they can tell you how likely something is to be true. In this case, what we'd like to know is if this guy's abnormally good play was simply a fluke. The best statistics can tell us is that there is an X% chance it wasn't. We have good data on the population of his previous plays, so, given an objective measure for the goodness of moves, we can tell a good bit about the distribution, and get a reasonable estimate of the probablity of playing one excellent (grandmaster-level) move. So we can give an estimate for how much of an outlier his play actually was. As far as the "we shut off the cameras and he started playing normally", that's pure anecdote. Correlation in a sample size of two is utterly useless in demonstrating anything. Now, if you toggled it off and on and off and on, and got a dozen data points, it might be helpful. But as it stands, all you can say is he started playing the most probable moves again at around the same time. That in no way counters the possibility that his good play was merely a statistical anomaly; it's imply the expected behavior, until your sample set is large enough to show correlation