With the New Year upon us, it is time to look back at 2015 and compile all the ratings into one post and see how I’m doing dispersing the ratings. I had 236 ratings on my site in 2015. Below is a Histogram showing all the frequency of all my rankings in 0.5-point bins.
This is nearly an ideal distribution of ratings. A peak in the 6.5-7 range then tapers off each way – higher and lower ratings. This tells me that my average rating is somewhere in that range. Going forward, you can expect that a rating below 6.5 is below average, and perhaps not really worth your time.
The distribution is not quite ideal, though, as the 5.5-point bin is a little low and/or the 5.0-point bin is a little high. This is a little surprising to me as I thought that I rated more things in that range as a 5.5 rating. To me this means that I could round to a 6.0 and someone might like it, just not me. The other item of note is the bump in rating at the 9.0+ rating – more on that later.
A significant portion of my ratings were related to Hallmark movies. If we remove all the Hallmark ratings, there are 97 ratings remaining and this is what the Histogram would look like:
This is less ideal, with many more ratings above 6.0 than below. I believe there is a selection bias, for instance almost half of these ratings are from my What’s on My DVR post. Obviously, only TV shows that are rated above 6.0 are here, because if it rated below 6.0, why would I continue to watch it? I would love to remove this post to see what the remaining Histogram looks like, but then we get into small sample size issues, which I would like to avoid. The selection bias also explains why there is a bump in the 9.0 bucket (in addition to the DVR post, I also included restaurant reviews from prior years that I liked). Next year, when I look back at 2016, hopefully this will also be a nice smooth curve like the one above.
The main takeaway here is that a rating of 6.5-7.0 is an average rating and ratings below that should generate caution. There is also a selection bias with my prior ratings, but that is largely a function of introducing the best of prior years into the ratings; going forward I hope this selection bias will be largely mitigated. There will still be shows or books that I refuse to watch because they look terrible (like last night’s premier of Shades of Blue), so my ratings will still skew slightly upward because I won’t even bother to rank the worst of the worst.