Our goal is to gather relevant critical analysis of anime in a single place in order to provide both a database of reviews and a tool to understand critical consensus.
Every piece of work that receives a minimum amount of reviews has a score between 0 and 10 based on whether or not the reviewers recommend it. We provide a ranking of all works in the database, but more importantly, rankings ordered by type and genre. The seasonal and studio pages also order the works by score.
How it works
Reviews on The Anime Reviewer use the TAR recommendation system. The final verdict is one of 5 recommendations described bellow:
Must Watch: This piece is strongly recommended by the reviewer.
Watch: This piece is recommended by the reviewer, with reservations.
Neutral: The reviewer is neutral about the piece.
Pass: The reviewer recommends skipping this piece, with reservations.
Hard Pass: The reviewer strongly recommends not viewing this piece.
What to expect from reviews
There are criteria reviews, and subsequently reviewers have to follow in order to be indexed or posted in The Anime Reviewer.
Due to the serialized nature of television and the extent of most works, there is no minimum amount a reviewer needs to experience a piece before making their analysis. A verdict can be reached as early as:
- The reviewer has enough information about the direction and quality of the piece to make their veredict; or
- The reviewer has enough information to predict with confidence based on previous experiences the overall direction and quality of the piece based on their analysis framework;
Only the original piece is used as criteria and any adaptation of the work to other languages and demographics but the original are ignored. Recommendations are based on technical and artistic merit and not on public acceptance.
Genre awareness. Reviews take into account genre, target audience, the medium it aired and other contextual elements. In order to provide a fair comparison between different genres
No ironic reviews or "So bad it's good". Those can be seen in the analysis themselves, but the final recommendation needs to be the sincere recommendation of the reviewer.
Reviewers are vetted based on a subset of their reviews. For this reason, there might be reviews on the site that do not follow these criteria. Please let us know at dev@theanimereviewer.com if any review on the site does not conform to the rules estabilished in this page.
Why not use a numbered score?
Although the computed score for each work is numerical, there are several issues with trying to make humans use a numbered score system, specially when more than 5 possible scores are used.
It makes it harder to choose an outcome. If you are rating multiple things from 0 to 10, you need to have a clean understanding of what every number mean, as well as how they compare to each other. And most humans would have a hard time keeping all that in mind.
You cannot give a score without having a lot of other items as reference. The nature of numbered scores is that you need a frame of reference to build them. It means the first things you score are likely to need revision after you number of reviews increase.
It gives the illusion of an objective scale. Humans are very bad at objectively ranking things, specially to the precision of decimal points. The result of critical analysis is never a number. It can be converted to a number afterwards, which is what most people do when using a numbered scale.
Numbers do not convey meaning intrinsically. What is an 8 for one person is a 6 for another. It is very hard to get into a consensus of what numbers on a scale from 0 to 10 mean for everyone.
People do not use all numbers. It is common to see people either focus on extremes or limit numbered scores to a smaller subset. eg.: In a scale from 1 to 10, use only 7,8,9 and 10.
Those are some of the reasons we believe using TAR recommendation makes communication between reviewer and public clearer and more direct. Of course most reviewers who also use other platforms have their scores over there, and sometimes we might have to translate those scores into our model. And since those review scores were made with a numbered system in mind, they might be not accurately translated into the TAR recommendation system. Once we have enough data we might look into it.