My most viewed posts on this blog, not counting one anomaly, are my reviews of movies, books, video games, etc. So I thought I’d talk about how I write them and, more significantly, what I think an effective review should do.

The easiest way to begin is to categorize reviews by usefulness. I based the idea on Paul Graham’s disagreement hierarchy.  Reviewing something isn’t exactly the same as arguing, but they are related in the sense that both should be about working to improve something, so there is a fair amount of overlap.

[In the following x = any movie, book, game etc. being reviewed.]

Tier 1 – The Most Useless Review

“This x sucks. I hated it. Why would anyone think it was good? What a terrible piece of work.”

This review is useless to everyone except the reviewer. It does not explain what the flaws with the x were, nor does it even give us an idea of what the reviewer looks for in xes.

Note that a really good writer can dress up a useless review very nicely, so much that you might think it’s a useful review. For example:

“x is a piece of unmitigated tripe, the likes of which I am sorry to have ever had the displeasure of enduring. It is a blot on the [whatever x’s genre or medium is] landscape.”

This sounds kind of funny and clever, but it doesn’t say anything helpful.

Tier 2 – A Mostly Useless Review

“I loved this x! It’s the best one ever. Everything about it was terrific.”

This is close to being as useless as a Tier 1 review. In fact, I was originally going to group it as a Tier 1, but then I realized that there are two important differences. First, while a Tier 2 might be useless to everyone else, if I read a Tier 2 review of something I made, it would at least make me happy.

And secondly, if somebody wrote a Tier 2 review about one of my books, it would be more useful to me than a Tier 1, because if I know someone liked something, I can figure out how to produce more things they’ll like by sticking to that formula. Whereas with a Tier 1, I have no idea how to produce something they will like. It’s completely useless feedback.

That said, while a review like this is mildly helpful to the creator themselves, it’s useless to all other consumers. They have no way of determining what was good about the x in question—all they know is this reviewer liked it for some reason.

Tier 3 – The Diagnostic Review

“I liked the parts of the x with these characteristics. I didn’t like the other parts with these characteristics. The second half was stronger than the first. It reminded me of y at times, but lacked its [some element of y]. Fans of z will like it.”

This is getting somewhere. This review tells us some specific elements about the x that they thought were good or bad, and provides a reference point that we can use to start to form an idea of what x is, and what type of audience it would appeal to.

This is an informative, workmanlike review, and if I write a review this good, I feel like I’ve done my job. But there is still one higher form to aspire to:

Tier 4 – The Corrective Review

“The central flaw of x is its [some characteristic], which could have been fixed by the following adjustments.”

Or, if it’s a positive review:

“What makes x work so well is its [some characteristic]. This separates it from [other things in x’s genre or medium] to make it truly effective.”

The best type of review not only diagnoses what is good and/or bad about an x, but states why these elements are there, what could be done to fix them if they are bad, and how to replicate them if they are good.

The best review, in other words, not only tells you the negatives and positives of x, but provides sketch blueprints of how to make more and/or better xes. If it’s a bad x, a good review tells you what went wrong and how to fix it. If it’s a good x, it tells you how to replicate its goodness.

Not surprisingly, Tier 1 reviews are the easiest to write and Tier 4 reviews are the hardest.

Also, it’s important to realize that context matters. Sites like Amazon encourage short reviews, and on average a Tier 1 or 2 will be much shorter than a Tier 4. A Tier 4 is an extended argument, in that you have to make a series of statements that follow one another logically.

Realistically, I think a Tier 3 is the most you can ask for from anyone who isn’t a professional writer or trying to become one.  Tier 4s take too much time. The most useful Amazon reviews I see are all Tier 3s. My eyes glaze over when I see a really long Amazon review, so even if it’s extremely helpful, it probably isn’t worth posting there.

On blogs or sites dedicated specifically to criticism, it’s a different story. On my blog, I feel like I have license to go on as long as I need to make my point.  (You may have realized that already.)

One thing I noticed once I started thinking about this is how many professional reviewers are only capable of writing Tier 1 and 2 reviews, albeit dressed up with five-dollar words and clever turns of phrase. How are they getting away with this? Well, people don’t read professional critics solely to get information, they also read them to be entertained. And nothing is more entertaining than a well-written negative review.¹

This is dangerous. There’s a strong incentive to write negative reviews, because they tend to attract more attention and comment. And if you’re focused on being negative, you’re very likely to lose sight of the thing that makes a Tier 4 review: ways to make the x in question better.

When faced with a choice between being helpful and being entertaining, critics feel a strong pull towards being entertaining. I’m guilty of this myself—writing a scathing but unhelpful review is way more fun and more rewarding (in terms of page views and comments) than writing an extremely useful review.

I’m not saying you can’t be witty and entertaining when you write a review, but that you should be careful not to lose sight of the larger goal. The job of an x critic is to figure out how to make good xes and avoid making bad ones.

Looking back at reviews I’ve written, I’m not sure that I’ve ever achieved a Tier 4. Probably the best one I’ve done so far is this review of Planescape: Torment. I say it’s the best because it’s the only one where I figured out why I liked it so much as I was writing the review. I felt like by writing it I was finally solving a mystery I’d been trying to answer since I first played the game in 2010.

This brings me to a final point about Tier 4 reviews: writing one will probably mean going over the x in question more than once. Maybe I’m just slow on the uptake, but I usually have to read/watch/play something at least twice before I can give a useful review of it. This is another reason Tier 4s are so rare.

Footnote

1.Mark Twain’s brutal review of James Fenimore Cooper’s novels is a good example of this.