The dead horse woke up again in the DMoM&M-thread. As far as I see ratings the are roughly like this (multiply with 10 for the 1-100 scale):
*/1.0-4.9: Crap
**/5.0-5.9: Flawed
***/6.0-7.4: Ok
****/7.5-8.9: Good
*****/9.0-9.9: Excellent
So if a game gets *** or 6.5 I expect it to be ok, slightly flawed perhaps, but also entertaining for some. It’s obvious that half of the 1-10 scale is used only for crap, and people generally don’t care about the exact degree of crappiness.
This might not be optimal for a ratings purist, but for most people it wouldn’t really matter if all the action was between 9-10, with 9.1-9.49 mening “crap”, 9.5-9.59 mening “flawed” and so on. The interesting part is that the rating is easily understood, because as far as I can tell the main purpose of a single numerical rating is to easily and quickly tell very roughly what the reviewer thought about the game. If you want to know more in detail how the reviewer made up that number you could try reading the review or start flaming on various message boards until you get the general idea.
Changing the scale in such a way that a 6.0 on the usual scale becomes a 4.0 is nothing but misleading. To me it is as useful as saying that “flawed” is the new “ok”, “ok” is the new “good”, and “good” is the new “excellent”. It just doesn’t help anyone at all.
Saying that you want 5.0 to represent an average game is silly, because people can’t relate to an average game. If you have a rating of all citizens there will be some point of reference, but with games it’s impossible to say what the average would be of. The average of all games sold, or maybe released? I haven’t even played 10% of all fps-games, so I have no idea what the average would be in that case, but my guess is that it would be a pretty crappy game. This would of course mean that we’re back in the 7-9 range for most games that people actually know about.
The 7-9 scale is actually quite intuitive for me, because here in Finland, the school grades are 4-10 (for some reason), with 7 being average, 8 good, 9 excellent and so on. I don’t know about the rest of the world, though.
I just don’t like using the number system at all and thus I don’t use it on my site. It really doesn’t convey anything to the reader yet it’s the most visable and saught after piece.
Also (going out on a limb here), I assume that those that are more serious about thier game purchases will read the review anyway to reach thier own conlusion.
Yes, exactly, it’s the way it matches up with the grading system. A 5 will always be a failing grade in most people’s minds. If you want to use the full range I think it’s a good idea to use the star system – 2.5 stars means ‘ok’ to my mind.
Exactly. Take darkstar one for example. A pretty damn mediocre game, really, but I bought it and enjoyed it because I’m starved for space fighters.
I still like daily radar’s miss/hit/direct hit system. Basically it’s thumbs down, thumbs up, or thumbs way up, with the latter reserved for the half-life2’s and oblivions. It had the distinct advantage of not being arbitrary.
I still think a letter grade system is the most intuitive, though I’d be happy with a thumbs-up/down approach too. Everyone knows an A is excellent and a C is average, or at least I think they do.
An A-F letter grade is both intuitive and allows for the use of the entire scale.
I think a “D” would be a lot more reasonable for Dark Messiah (the game that sort of kicked up this debate this time) than a 4. Just because the last time I remember seeing a game get a 4 from a major publication, I think it must have been one of IGN’s scathing reviews of a HeadGames’ EXTREME game line.
CGW (or GFW:TOM) has turned it’s review section into a place that reviews other peoples reviews. They go a step further than merely printing other scores, they publish quotes within the body of the GFW:TOM (or CGW) reviews.
I don’t want half my review space taken up with "…this is what Blah Blah Kasan at Gamespot trenchantly observed", or “…Woopdee Doo at Game Mag worried that…”.
With compilation sites like Meta Critic and Gamerankings, not having consistent scoring can skew overall averages from sites like this. Tom’s 4 out of 10 score (which would’ve been a 6 out of 10 with any other site) is going to impact their overall score at Meta critic and gamerankings.
Averages being skewed is obviously a pretty severe problem. I think the best solution is to simply have every reviewer and review site rate every game an 8; this will also ease the server load on averaging sites by reducing the calculations necessary. I don’t see a downside.