Why we don't score out of ten

The short version: every score here starts at ok, and stars have to be earned.

• ok ✓: no problems and no surprises. Not a bad mark.
• good ★: better than it has to be.
• great ★★: worth paying more for, or taking a detour.
• exceptional ★★★: the kind of thing we tell everyone about. Rare.
• avoid ✕: it failed us, or it was not worth the money or the time.

Why 8/10 sounds like meh#

Have you ever caught yourself in an online shop thinking that 8 out of 10 looks like a bad deal? That a 7.3 on Booking feels below average? That a restaurant with 3.9 out of 5 on Google is not worth the visit?

Scores are supposed to help us choose, but they often mislead, because the scale itself is skewed. On paper, a scale from 1 to 10 has its middle at 5 or 6: not bad, not special. In practice almost everyone, writing or reading, treats 10 as normal. A thing that simply works as it should "deserves" a 10, and every point below means something was wrong with it. So 9 out of 10, or 4.5 out of 5, already reads as just ok, and anything lower reads as bad.

You know it from Uber: five stars is a normal ride, and four stars is already a warning for the driver.¹1 Drivers whose average falls below the minimum for their city can lose access to the app.

Out of ten points, we really use three: 8, 9 and 10. Everything from "fine" to "the best I have ever had" has to fit into them. That makes scores hard to give, hard to read and easy to misread, and the ones that suffer most are the things that are actually good. An average restaurant gets 4.3, the really good one where they take proper care of you gets 4.8, and on the scale they blur together.

A scale with a visible top makes you count down#

A scale from 1 to 10 or from 1 to 5 is not wrong in itself. But once the top is in sight, readers and writers alike start from it and count down. Every score becomes a question of what is missing instead of what is good. It does not matter whether the top is a 10, five stars or a word like "excellent".

Defining the middle does not fix it either. Edge, a games magazine, scored from 1 to 10 with 5 as average, and it still became known as harsh: its scores sat well below the average critic's.

Our scale has a top too: three stars. The difference is where a score starts. Nothing here starts at three stars and loses them. Everything starts at ok and has to earn its way up, and we never show empty stars, so there is nothing missing to count.

What we do instead#

If the top of the scale is the problem, and ok has drifted from 3 out of 5 to 4.5, why not start the scale at ok and give stars as a bonus?

When something is ok, it gets a ✓. That is not a bad mark at all: it is fine, it does the job, it is just nothing special. When a thing, a trip or a product is really good, it gets stars on top: one ★, two ★★ or, rarely, three ★★★. Extra ratings, such as reliability, value for money or the views on a trail, use the same words and marks.

We borrowed the stars from the Michelin Green Guide, which rates sights as ★ interesting, ★★ worth a detour and ★★★ worth the trip.²2 The red Michelin Guide does the same for restaurants: one star is worth a stop, two a detour, three a special journey. Most restaurants in it have none. Michelin only marks the best, though, so everything else stays without a star, whether it is fine or awful. We split that blank in two: ✓ for ok, and a cross ✕ for things that are really bad and better avoided.

The verdict says the rest#

The stars say how good it was. Whatever they can't say, the one-line verdict next to them does: a ★★★ trek can be "stunning, but once was enough", and a plain ✓ tool can be "nothing special, and I'd buy it again tomorrow".

Why scores never show in lists#

A score says how happy we were with a thing or a trip. It should not decide whether you read about it. Often the bad experiences and the lessons teach the most, and a ✕ next to a title in a list would only make you skip them. So lists show what a post is about, and the score waits inside the article.