clock menu more-arrow no yes mobile

Filed under:

About Reviews

Polygon's review scores and "the bump."

The era of the game as a static product is essentially over.

Every day, we hear developers and publishers refer to games as "platforms," which is accurate. Not only because so many games have a regular schedule of downloadable content built into their business model, but because at their most basic level, games are often very different even a month after their release. The "day one patch" has become commonly used vernacular in publisher-to-press and publisher-to-consumer communication. Games are also more complicated technologically than they've ever been, which means that over time, issues may become known that were not apparent at the time of release. Multiplayer games depend on users to be their content. If no one plays an online shooter, then its value decreases to gamers who unwittingly buy in late.

Reviews are valuable to our readers, and we believe review scores are part of that. But reviews are also inherently anachronistic. At Polygon, we operate on the supposition that reviews are a critical evaluation of a game on the day of its release, which the score we assign reflects. As a primary means of informing our readers, a static review score has a very limited shelf life. But review scores also serve as a foundation for a site's legacy and credibility with its audience. They're one of the ways that readers understand where a site is coming from. This makes changing review scores a taboo topic.

But we can update them.

Polygon's reviews and database have been built based on the idea of updates, or "bumps," as I've called them. If a game changes in a substantive way, we can add an update to our reviews that informs you how and why, and we can modify our scores accordingly. This will appear on the reviews in question as a timeline of that game's evolution and our corresponding recommendation (or lack thereof). The original review score will never vanish or go away, but our readers will be able to better understand where our opinions as a site reside over time for games we review.

We do not guarantee that we'll be able to do this with every game, and whether we do so or not is solely at Polygon's discretion. We will act in what we consider the best interest of our audience, while being as fair as possible to the developers and publishers who pour time and money into the games you play. If a game sees substantive improvements that make for a better experience, we want to reflect that. If a game is less worthy of your limited time, we also want to reflect that.

How we score

Polygon uses a "20-point" review scale. Specifically, we rate games from 1-10, with half point intervals. Our scale differs somewhat from other sites, but should still be relatively easy to understand for new readers.

Games are not scored until a review is written and finalized. Once a review is complete, the reviewer meets with a group of senior editors to determine which score on our scale properly reflects the text as written. We do not write with scores in mind.

Polygon's commitment to readers with regards to scores is as follows: You should not be surprised by the score you find at the bottom of the review you read. The score will be a logical afterthought to a well-developed opinion. And in consideration of those readers who don't care for scores, we have left them at the bottom of our reviews.

Polygon's review philosophy

The most limited resource any of us have is time. You will never have more of it, and we want to help you avoid wasting it as much as we are able. While money is obviously an important consideration, Polygon's reviews operate on the premise that time is paramount, and our scores reflect that. Our reviews seek to help you make informed decisions on how to spend the limited time you have. The breakdown of our scores below elaborates further on that.

I know that's a lot to take in. If you have questions about our reviews system, you can email us at

Our review scale

A score of 10 is the highest recommendation we can give. 10s represent ambitious games that succeed in ways few games have, and that we expect will be part of the gaming conversation for some time. These are the "must-plays." However, this is not a "perfect" score. We've never played a perfect game. Except for The Legend of Zelda: A Link to the Past.

Nines are fantastic games that can be recommended without much in the way of qualification. They may not innovate or be overly ambitious but are masterfully executed, or they may be ambitious in scale and scope with competent execution.

Eights are great games, and easily recommendable with caveats in mind. They're examples of consistently sound design, or a novel concept well-developed around a functional core. A game that executes well enough to be remembered, even if there are better contemporaries.

Sevens are good games that may even have some great parts, but they also have some big "buts." They often don't do much with their concepts, or they have interesting concepts but don't do much with their mechanics. They can be recommended with several caveats.

Games with a score of six have good parts, but uneven overall execution. Prospective players should know what they’re getting into before they dedicate time and commitment.

A score of five indicates a bland, underwhelming game that's functional but little else. These games might still possess quirks or aspects that appeal to certain players.

Fours fail completely in one of three ways: design, execution or basic functionality — or they fail a fair amount in all of them.

Threes fail completely in two of three ways: design, execution or basic functionality — or they barely achieve a baseline in any of these.

A two is the lowest score a playable, working game can get. These fail in design, execution, and basic functionality, though a motivated player can finish the game.

A score of one indicates that Polygon review staff believe said game doesn’t properly function. Most reasonable people will not be able to finish a game with a score of one due to massive technical, design, and execution problems.