Look back through anyone's game ratings and you'll usually find the same thing: a huge pile of 3.5s and 4s, a handful of 5s, and almost nothing below 3. Most people don't only play good games. Rating on the spot, with no rules, just drifts towards polite. A year later, the numbers can't tell you which of those forty 4-star games you'd actually recommend to a friend.
This guide is about fixing that. It covers how to rate video games on a five-star scale so each star has a job, when half-stars earn their place, and a few habits that keep your ratings honest over time. None of it requires a spreadsheet.
Why most game ratings bunch up at 4 stars
There are a few ordinary reasons ratings pile up in the same spot.
The first is timing. People tend to rate a game right after the credits roll, when the ending is fresh and the slow middle chapter has been forgotten. Finishing something feels good, and that feeling leaks into the score.
The second is sunk cost. If you put 60 hours into a game, giving it 2.5 stars can feel like admitting you wasted a month. So it gets a 3.5 instead.
The third is borrowing someone else's scale. Critic scores on Metacritic and OpenCritic cluster in the 70s and 80s, and a lot of players quietly convert that into stars. A 75 becomes a 4, and suddenly "fine" and "great" live on the same line.
None of these are character flaws. They're what happens when there's no definition behind the number.
What each star should mean
The simplest fix is to decide in advance what each rating means, in plain words you'd actually say out loud. Here's a scale that works well for games:
- 5 stars: I'd push this on a friend. It changed what I expect from games, or I'll be thinking about it in five years.
- 4 stars: I'm glad I played it, and I'd recommend it to the right person without hesitating.
- 3 stars: It was fine. Good parts, forgettable parts. I don't regret it, but I wouldn't chase it down for someone else.
- 2 stars: Mostly not for me. Something kept me going, but I'd steer people elsewhere.
- 1 star: I wish I'd spent the time on something else.
The important line is between 3 and 4. A 4 is a recommendation. A 3 is not. If you hold to that one rule, your ratings will already be more useful than most.
It also helps to rate the experience you had rather than the game's reputation. Elden Ring can be a masterpiece and still be a 3 for someone who bounced off it after ten hours of dying to the same knight. That 3 is honest, and future you will appreciate the honesty more than a score that matches the internet.
When to use half-stars
Half-stars are useful, but they're also where the drift creeps back in. A good rule is to treat them as tiebreakers, not as a separate scale.
Use a half-star when a game sits clearly between two definitions: a 3.5 is a game you liked more than "fine" but wouldn't quite recommend without a caveat. A 4.5 is a game you'd recommend to almost anyone, that just missed the "five years from now" test.
If you notice most of your ratings end in .5, that's a sign you're avoiding decisions. Try rating the next few games in whole stars first, then only adding the half if the whole number feels wrong.
Rate what you played, not what you finished
A lot of people don't rate games they didn't finish. That leaves a strange gap: the games that lost you are exactly the ones worth remembering, so you don't buy the sequel or recommend them by accident.
It's fine to rate an unfinished game, as long as you note that you didn't finish it and roughly how far you got. A 2-star rating after 12 hours of a 40-hour RPG says something real. It says "this didn't hold me," which is useful information when the next game from that studio comes out.
The same goes for games that don't have an ending at all. Multiplayer shooters, city builders, Stardew Valley saves you've had for years: rate them for what they've been to you so far, and update the rating when that changes.
Wait a day before you rate
The easiest habit to adopt is a short delay. Finish the game, sleep on it, then rate it.
A day is enough to let the credits-high fade and for the parts that bugged you to resurface. It's also enough for the good parts to show whether they'll stick. If you're still thinking about a boss fight or a line of dialogue the next morning, that's worth half a star. If you've already forgotten the plot, that's worth knowing too.
If waiting a day means you'll never get around to it, rate immediately and add a note to revisit. What matters is that the number isn't made in the glow of the final cutscene.
Check your ratings against each other
Every few months, look at your ratings side by side. Ask one question: would I swap any of these?
If you gave a game you barely remember the same 4 stars as one you've recommended to three people, one of them is wrong. Adjust it. Ratings aren't contracts, and changing your mind after a year is useful information in itself. Plenty of games grow on people long after they finish them, and plenty shrink.
It also helps to look at the overall shape. If you've rated 60 games and 45 of them are 4 or above, either you're very good at picking games or your scale has drifted. A healthy spread usually has a real middle: a decent number of 3s, some 2s, and 5s that feel rare.
Add one sentence
Stars are good for sorting. They're bad at explaining. The single best thing you can do for a rating is add one sentence about why.
It doesn't need to be a review. "Brilliant first half, the last dungeon dragged" is enough. So is "played this with my brother over Christmas, the co-op was the whole point." A year from now, that sentence will tell you more than the number does. If you want to go further, our guide to writing game reviews you'll still want to read in five years covers what to put in a longer note.
A scale for yourself first
Ratings get more useful the longer you keep them, but only if they mean the same thing in year three as they did in week one. Decide what each star means, treat 4 as a recommendation, rate what you actually played, and give it a day.
In Perthro, ratings go from half a star to five, and your profile shows a histogram of every score you've given, which makes it easy to spot when everything has quietly turned into a 4.