Support for the netflix rating system

I have lists of the books i’ve read, and for a long time i have given each book a rating out of five stars. Giving a book five stars is particularly useful because it allows me, when i’m asked if i’ve read anything interesting recently, to quickly grab the list and see what i’ve rated highly.

The trouble with this is that i’m then not sure what the stars are actually measuring. Should it be how much i liked the book? But there are books i have hated and would still want to recommend to people. Conversely, there are books i quite enjoyed but wouldn’t want to read again, nor to recommend to anyone else. When people occasionally look at my ratings, i find myself having to disclaim that the rating isn’t just a measure of how much i agreed with the book.

Once i was showing a friend my system and they asked what i did if i wanted to give something a half star. A half star? I’d specifically chosen to rank things out of five stars because i just can’t think in ten-step or twenty-step granularity. Mostly i default to giving things four stars, and if a book is particularly notable it gets five stars, and if it was much of a muchness it gets two or three stars. One star is reserved for the lowest of the low.

Having two possible rankings (two or three stars) for the same sentiment feels like a flawed system to me. So i set about thinking about what i could do instead.

The first system i was thinking of adopting was the rotten tomatoes system with two categories. On rotten tomatoes, if a film receives more positive reviews than negative reviews, it is ranked as “fresh”; if there are more negative reviews than positive, it is ranked as “rotten”. This was a good starting point, because i can quite easily say whether i thought reading a particular book was a waste of time or not. However, such a simple system lost the big advantage for me that was being able to easily pick out those books that were really worthwhile when asked about what i’d been reading.

Although it seems obvious now, i did have to wait for netflix to spell out the solution for me. At the end of a film on netflix, it asks you whether you liked or didn’t like a film, much the same as rotten tomatoes. I still feel that this language is not exactly right, and i would rather be asked if i felt empty or not after watching a film; that is, there is lots of media i can enjoy, have a good time reading or watching or listening to, but still feel that it didn’t really make me think or enhance my life in any real way. And that’s okay, but maybe i won’t be seeking out more like that in the future.

The crucial extra thing that the netflix ranking has though is a double-thumbs-up button. If something was really good, you can show that it wasn’t just palatable, but genuinely enjoyable. This solves my problem! I minimise the cognitive overload of having to pick a rating, but still have a way to highlight the really good stuff.

And so with the power of find and replace, i turned all my five star ratings into double-thumbs-ups, all my four star ratings into thumbs ups, my one stars into thumbs downs, and went through the two and three star ratings manually, deciding on a case-by-case basis what they should be—mostly it ended up being thumbs ups, and most of them were originally three stars. It turns out i very rarely gave things two stars.

I didn’t take any backups so my original, more precise system is gone forever. But i don’t mind that, because i don’t think that the increased precision was actually resulting in a higher accuracy.

We’ll see if i regret this decision, but for now, it feels like a little bit of weight has been lifted off my shoulders.