Opinions on Review Score Scales?



  • By review score scales, I am referring to using a 10 star system, 5 star system, etc. to review games.

    I assume most people are comfortable with either the 10 or the 5 star scale, because that is what I see people using 90% of the time. But from a consumer standpoint, using reviewers to partly help influencing a purchasing decision, I feel that both scales are lacking.

    Starting with the 5 scale, which I think is definitely better than the 10 scale, my main issue is the difference between rating a game 1/5 vs. 2/5. Regardless if a game gets either of these scores, the reviewer's opinion already seems to be communicating that they had a negative experience with the game. If I see a game get a 2/5, it tells me the reviewer did not like the game.

    I feel that the 10 scale is the worst, as the balance feels completely off with how practically all reviewers seem to handle it. Anything from a 1/10 to a 5/10 tells me the reviewer didn't like the game. Is there really much value in differentiating what makes a 1/10 vs. a 3/10? Is it important to have so many different levels of "bad"?

    Wouldn't a 4 scale system make more sense? A 4 scale system would not necessarily manifest similarities between an educational grading scale, as 5 and 10 scales seem to. It is slightly more adaptive than a thumb up/thumb down system, as you can differentiate between good and really good. The 2/4 feels different enough from a 1/4 in communicating a mediocre experience vs. a negative experience.

    To highlight some questions:

    Do you think there is an important distinction between reviewers giving a game a 1/5 vs. a 2/5? Is it important to have the range of 1/10 through 5/10 in communicating a negative experience?

    Do you think the aspect of giving a game a number score has any inherit benefit over reviews that do not use any numbers or stars? Do you prefer reviews that do not use a number scale?



  • I think there's a problem with the way review scores are used specifically in gaming, however I still favor the 20 grade points, some from 1 to 10, with 0.5 intervals.

    Why do I like this system? Because I think it allows you to convey a lot more of the quality of a game, any more than that is just useless, which is why I find 100 based scales and scores like 87 completely meaningless, and less than that I feel it doesn't give enough room to convey an opinion accurately in some cases.

    I do agree with your point about low scores, and I generally think that it is mostly understood that anything below a 5, is basically indicating something is not functional, so everything at that range does lose value and the scaling becomes less important, but I don't think this is a big issue.

    I'm ok with reviewers simply giving recommendations instead of scores, but I'm still partial to scores, even if I can't say there's a specific reason why it would be better, I think it is useful if only for data collection purposes, but personally I give much more weight to opinions from reviewers I know and the actual critique, for example I simply don't trust Huber reviews and I know that I'm very likely to agree with Jason Schreir's so I adjust my expectations accordingly.



  • Eurogamer has an interesting 4-point scale:

    • Avoid: The game is either broken or plain crappy.
    • "Blank": The broadest category, no particular recommendation. It has flaws, it's not the best of its kind, but there's still some good stuff in there so your mileage may vary, just read the review and see if it appeals to you!
    • Recommended: You can't go wrong with it, as long as you have an interest in the genre, it's doing a lot of things right.
    • Essential: This is a classic that will stand the test of time, you should play it even if you're not usually a fan of that sort of game.


  • I think review scores are best when they give people a very basic idea of the quality of a game. One review score can not accurately convey all the little nuances that make a game what it is. Reading or watching the actual review itself, is what will help someone with there purchasing decision. That's why I prefer a smaller point scale like the 5 star system or 4-point scale.



  • I like 5 point scales when I think about scores for personal scoring, but I can see why a published review would use more points. That said I'm the type who believes a score doesn't have a finite resolution, it just gets progressively harder to judge how you feel about it the more decimal points you try to add (and what actually goes into the score, how much weight to give specific feelings, etc.). Even a 5 point scale can feel like you are inconsistent with yourself sometimes!

    Here's why scores with a large number of points is appealing:

    A lot of people like the idea of ranking all the games (or all the movies, etc) they've played, in order from best to worst. If you do this even part way, you end up with so much ordered data that it is easy to appreciate the idea of an absolute score on even, say, a 100 point scale. This is because once you imagine a few anchor points in your list where you think an "80 game" would be, or a "90 game" or a "50 game", etc. then you can get a good estimate of any game's score by applying a little subjective judgement for how far between the two nearest anchor points it lies.

    There's an interesting side subject here on scale types: [1] [2]. Ordering a collection of games (ordinal scale type) is simpler in terms of information content that giving them all score (interval and ratio scale types). With an ordinal system, the only meaningful operation is less-than, greater-than, and equals. However, a score system will usually be interpreted, at minimum, as an interval system. This means it is meaningful to talk about numerical sums and differences: a 9.0 game is not only better ("greater than") than an 8.0 game, but also this difference the same interval of quality between an 8.0 game and a 7.0 game.



  • My favorite scale is probably the 20 point one like EZA uses now. I think it allows for a more accurate "expression" of what the reviewer feels, while still limiting them just enough so they can really nail down their feelings. My main objective of reading reviews is to know how exactly the reviewer feels, so more simple scales like Eurogamer's 4 rating system is less useful for me. Let's say that game A is bad, so it gets the Avoid rating from Eurogamer, and it gets 3.0 from the Allies. Both accomplishes the objective of teeling the readers to avoid that game, but the Allies' scale gives me more of the impression of just how bad it actually is. Of course that impression has already been given on the actual review script, but the added accuracy of the scale really adds a lot for me personally.



  • Being a teacher I like dat 100 point scale.
    Game gets a 50? Sure it might have some redeeming qualities but it has done most things incorrectly and is just barely scraping by.
    In dat 70 range range? You're making some good points but maybe you completely ignore proper formatting.
    90-100? Good job you are in the top percentile of the class. You did your research and polished your work to a mirror finish. Maybe the rest of the class could take a few notes from these guys.