If you liked sarcasm, you'll love irony

"If you liked 'Gladiator', you'll love 'Emperor: The Gates of Rome'!" said the blurb on 'Emperor: The Gates of Rome'.

Now, leaving aside the questions of whether or not I did like (or love) either of these things, let us consider the implications of this type of assertion.

We can suppose that people who like something also have, in general, a tendency to like similar things. Furthermore, we can usually observe that the more similar things someone likes, and the more they like them, the more likely they are to like another similar but as-yet unexperienced thing.

This is how (and why) Last.fm works. I tell it that I love, let's say, HammerFall, Zero 7, and Savatage, and it suggests that perhaps I might like Nightwish. Correct, Last.fm. Go to the top of the class.

Imagine a band (or any other item within your subjective frame of reference) as a set of variables. The meaning or value of each is irrelevant, but what we are saying is that for several variables x, x(Nightwish) is similar to x(HammerFall), x(Zero 7), and x(Savatage), and that for several other variables y, y(Nightwish) is either dissimilar to y(the rest), or unknown. If there are many more xs than ys, it predicts that I will like Nightwish. Otherwise, it predicts that I won't.

So, the point is that whilst you can predict that someone might like something if they love something similar, you can't predict that someone will love something just because they like something similar.

I suppose that this entire entry could consist of the first paragraph and the previous paragraph, but it is early in the morning and I have nothing better to do.

Also, if you loved 'Gladiator', you might like 'Emperor: The Gates of Rome'.