Three bullshit statistics

Here are three stats that often pop into my life that don’t explain exactly what they’re suggesting they do:

1. Rotten Tomatoes ratings: I used to think that a high RT rating was a pretty-much-guaranteed indication of a film’s quality, but then something happened. John Wick got 85% (I think it was over 90% at one point). I wasn’t the only one to be puzzled by this, but my explanation is slightly different. RT bases its score on whether a movie has had a positive or negative review, so a 6/10 review has a much value as a 10/10 and a 0/10 is the same as 4/10 (roughly speaking). So a great/shit movie that polarises (eg Inherent Vice – 71%) would get a lower mark than a so-so movie that doesn’t (Furious 7 – 82%). Is Furious 7 a better movie than Inherent Vice? Yes, if you’re an 8-year-old boy. Overall RT will point you in the right direction, but tread carefully.

2. The second one is also to do with movies. Cinemascore asks people who have just seen a movie to rate it on the school scale of A, B, C etc. As you can see, it’s pretty hard to get below a B, which is kind of odd considering how many awful movies there are out there. For example, Paul Blart: Mall Cop 2 gets an amazing 0% on Rotten Tomatoes but a reasonable B- on Cinemascore. Clearly the reason for this is that people who have just seen a film are the ones who wanted to see it in the first place. The legions of people with double- and triple-digit IQs who stayed away from PBMC2 weren’t forced to go and see it so that their opinions could give a fairer score. That means only the people who thought they might like it in the first place were asked and, surprisingly, they were broadly in favour.

3. The last one is about football, although I have a feeling it applies elsewhere. When newspapers report that a team has had two wins in ten games or three defeats in their last seventeen they also mean that the team has gone through three wins in eleven games or four defeats in eighteen. They’ve taken a point they want to make (this team is shit/good) and found the most illustrative fact to back it up. Obviously if they’ve lost three in nine that doesn’t sound as good as three in seventeen, so they go to the outer limit of the number of defeats, which means the next game that they decided not to include must have been a non-defeat. Obviously this doesn’t necessarily change the overall impression of the team’s form, but it is an interesting example of ‘not the whole truth’.

It seems I’m not the only one in this mood.