You can put mathematical flesh on this concept by reading up on the statistics behind the "multiarmed bandit". You get suboptimal performance if you never take any risks. Since you can't 100% quantify all "risks" involved in watching a bad movie or the joys, you can't just plug numbers in and crank the crank to determine exactly how often you should explore for optimal results, but you can still get the general gist that if you don't roll the dice every so often you won't get the best outcome.

It isn't even necessarily about watching "bad movies" on purpose. Even in the space of movies, one dice roll I've made is to watch the best movie of some genre I'm not particularly into. Even if it doesn't "convert" me to a fan of the genre this has often paid off for me. Another example is a lot of people get into musical ruts very early in life, often I think not even having a clue what the diversity that is out there is like. It doesn't take much to just make a random leap across the landscape to somewhere you've never been to check a new space out.