Apparently the advice assumes your opponent knows that most people choose rock and that your opponent accordingly chooses paper to beat your rock, so you should choose scissors.
But how many people engage in that reasoning? You might as well conclude that your opponent would choose scissors on the above reasoning and then choose rock or that your opponent would add in that iteration and then choose scissors to beat the opponent's paper. It's ridiculous. The only way this will work is if you have real empirical evidence about how much reasoning of this sort certain kinds of people engage in and some ability to figure out which category your opponent would be in. But no one can do that. It's true that it's not quite a game of chance, just as poker isn't. But that doesn't mean there's any rational way to win.
Of course, what you really ought to do is not play Rock, Paper, Scissors at all. It's much better to play Rock, Paper, Scissors, Spock, Lizard.