One reviewer, no algorithm

FAQ

Why a four-point scale?

The four-point scale forces the reviewer to make a direct call whether or not to recommend something. It breaks down the question of “how good is this?” into two questions:

1. Is this good or bad?
2. How bad/good is it?

This fits perfectly for recommendations because the first question is the main one: should you watch this? The second question is more about sorting, and knowing what I’d recommend you watch first.

Systems with a “neutral” rating allow a reviewer a non-answer to this question, and a quick way to escape making a recommendation for “meh” content, which is abundant. In “meh” cases I incline towards a rating of 2.

Systems with more granular ratings, say a 10-point scale, have a few issues. First is that it is unclear where the “average” for a reviewer lies. Do they consistently rate things as 8s, and a 7 is actually a bad review? Hard to tell and varies wildly for different people.

It works well for an aggregate of user reviews because the quirks of how different users rate things cancel out, but for a single reviewer it is a bad fit.

Why should I care about your opinion?

You shouldn’t. My opinion is just that, my opinion. It’s based on the thousands of movies and TV shows I’ve watched across about two decades, but it doesn’t make me an authority.

Most recommendation engines are flawed because they are based on what the aggregate of other users liked to determine what to recommend to you. This naturally leads to an abundance of similar content and sameness, especially if you use a streaming service where the content is limited and often repurposed and reordered into a lot of different categories: “Popular Picks”, “Trending Now”, “Just for You”, “We think you will love”, and so on.

This leads to doomscrolling a sea of uninformative movie posters that repeat endlessly.

The alternative? Find someone with strong opinions and see if they disliked some of the stuff you also disliked, and then see what they actually enjoyed. This has a better chance of yielding something you will actually watch and enjoy than trusting a recommendation engine.

Why are there unrated movies and shows on the lists?

Generally the unrated ones I have watched once or twice, but many months or even a year ago. They are in my pending “to rewatch” pile and I will review them at some point. But they were obviously high quality enough for me to remember the main outline of the story and characters, and that I enjoyed watching them.

Why is the “Where to watch” section wrong?

All the data about where you can watch a movie or show comes from JustWatch, through TMDB. It is fetched live when you open a review, for the country selected in the header. Catalogues change daily and the data is not always complete, so treat it as a strong hint and double-check on JustWatch if it matters.

How long did this take?

I started taking notes in 2017 about some movies and TV shows I enjoyed or disliked, originally to keep track and know what to rewatch or skip. Then I went long periods without updating the list but still watching movies and shows.

How representative are the lists?

Less than I would like, especially for the kind of “hidden gems” which I believe have the most value for prospective watchers. I have mainly compiled these lists from memory, but often also checked top lists from multiple sources and handpicked my own favourites.

There are other lists in mind that I’d like to make at some point, and of course I welcome any suggestions of lists and inclusions on the existing ones.

How soon after watching do you review?

Reviews don’t always happen close to the watch date; they are sometimes days, weeks or even months apart. I have watched probably another thousand movies and shows that are not in the reviews simply because I have forgotten about them or remember them only partially.

Generally if I watched a good movie or show long ago I have rewatched it at least once and reviewed it.

Data sources

Ratings, comments, features and lists are mine and live in a public data file. Posters, release years, runtimes and synopses come fromThe Movie Database (TMDB). Streaming availability is JustWatch data, served through TMDB and fetched live for the country you pick in the header. This product uses the TMDB API but is not endorsed or certified by TMDB.

The site is a set of static files: no accounts, no comments, no server that remembers you. The privacy policy lists the handful of third parties involved.

Contact

Suggestions for lists, a title you think I got badly wrong, or a factual error such as a poster that belongs to a different film: write tohello@tvtrove.com.