Friday, November 11, 2011

Bi-Modality in Ratings: A Measure of Idiocy

Hyperbole can be defined as 'exaggeration for dramatic effect'. If you rate any game at 100% or 0%, then you are an idiot.

My previous post mentioned that there are a suspiciously high number of 10/10 and 0/10 ratings for, well, a lot of games recently. Troll community notwithstanding, game companies have overhyped a lot of releases recently. Gamers are increasingly finding themselves let-down every time a game comes out. Games are increasingly hailed as the messiah by half their audience while derided as an insult to human excreta by the other. Of course neither are true.

I like to blame things on game companies, but let's get specific. Publishers. I can't help but feel this one is your fault. I think developers
will release something good, even if misguided, given the time and feedback they need. But the hype machine destroys feedback and schedules. When's the last time a game was as good as its trailer? I remember a certain game trailer that could have taken an award at Cannes . . . it had a sense of desperation and tragedy. A family's grave drama unfolding in reverse slow motion, sudden transitions and realization by the viewer. It had emotional impact. And the game had . . . well it had some zombies. Also a bug-ridden release.

How do we account for these disappointments? Well, I take note of bi-modality.
Most things follow a normal statistical distribution. That means for game reviews, if you stack all the scores up on the x axis, they'll make a hump centered over the average score, which should also be the median and mode. Since the scores are confined by the 0% and 100% limits, and people are biased to rate things between 50% and 90%, there is of course some skew. But the plot should still just look like a hump on one side or the other of the axis.

Instead we get this crap.

Fig. 1: A Decent Game Fig. 2: Internet Fight

On the left you have a game to which the community has responded normally (heh). They might be disappointed, but the ratings are not flooded by defensive fanboys and raging betrayed fanboys. Let's call this curve "Nerds play games".
On the right you have a game which is either OMG THE BEST GAME EVER MADE I CRIED AND QUIT MY JOB, or perhaps OMG THE WORST GAME EVER MADE I CRIED AND QUIT MY JOB, depending on who you ask. We can call this curve "Nerds get mad a video games".

The correct measurement for this is called "excess kurtosis". But since not everyone plugs Metacritic into Minitab, allow me to propose a basic formula for overhype polarization:

Idiocy = (Perfects paired with Zeros) * 2 / AllScores

In other words, the number of pairs of perfectly bad and perfectly good reviews you can make is directly related to how overhyped the game is . . . and how idiotic its community is.

That's right. I said it was the fault of publishers overhyping their product. But here's the truth. We fall for it. Then we defend it or scream about it. But we fell for it.

No comments:

Post a Comment