Okay, so this might have been done to death, but I think we can still squeeze some interesting stuff out of it.

A few questions to touch upon. Feel free to make your answers as pompous and long-winded as you see fit.

Generally, do you like assigning review scores?  Do you like to see them in the reviews you read?
Do you have a preference towards a particular scoring system?
What purpose does a score serve?  Who are they for, and how are they used?
Does it devalue the writing itself?
On what basis do you assign a score?  At what point do you start thinking about what score a game will get?

Go!

The consumer side of me finds having a score in reviews I read useful to get a quick and dirty sense of whether a game is worth considering for purchase. The writer side of me often finds the process of trying to arbitrarily shoehorn a score onto a more complicated review I've just written tedious at best.

I hate the normal score system with the numbers. That's why we are using on our website a system with: Crap, Bad, Avarage, Good and Amazing. The reason why we are doing this is because of the websites are using scores like 6,7 or 6,9. What's the difference between those two scores?

But I still prefer the context instead of our score.

Pretty similar to a scoring system I was quite familiar with (and in favor of), rating games from 1 to 5. In my opinion, it is pure bullshit to classify games in one hundred different levels of quailty, for that is essentially what you're doing when using a scoring system from 1 to 100. You can't rationally explain why a 6,9 would be appropriate where a 6,7 would not suffice, so don't make the distinction.

I think even ten distinct scores are quite a lot. Of course, there's a significant difference between an eight and a nine out of ten, but that's mostly because the score-quailty relation isn't nearly linear. I'd argue this difference is about as big as the gap in quality between two titles scoring 2/10 and 6/10. Besides, mediocre games are often rated six or seven out of ten. To me, a six is enough to pass, so don't hand it out to games you think aren't worth playing. What's the point in a one-to-ten scoring system if you're not using half the numbers?

When I wrote for GameCore Network, where we used a scoring system from one to five, the differences between those scores were so obvious, the entire crew agreed on the score we gave most of the time. I think that's very important: if your colleagues disagree, certainly your readers might, and if they do, there's little use in handing out scores at all.

"You can't rationally explain why a 6,9 would be appropriate where a 6,7 would not suffice, so don't make the distinction."

I'll have a go:

One of the numbers is higher than the other. The 6.9 goes to the better game.

Alright, I'll be serious.  We use a hundred-point scale at Resolution for a handful of reasons.  The first, most important one is that we ran a survey and around 65% of our audience liked that degree of precision.  Which surprised me a bit, actually, since I always consider our audience to be the sort that read the words and aren't fussed about the numbers.  Secondly (actually, this was firstly, technically) we wanted to mirror the mags we grew up reading, which tended to use a hundred-point scale.  Thirdly... oh, there's probably a thirdly.

But I actually don't mind the hundred-point scale, for a few reasons.  There's something extremely powerful about the high end of the spectrum.  That difference between a 91 to a 93 to a 95 to - dare we go there? - a 97 tends to mark a pretty remarkable section of high-quality gaming.  You could argue, of course, that anything in the 90s should be a must-buy, but a lot of people only dip into gaming occasionally, or don't have a lot of money, so need to know what we think really are the best. And, of course, the nature of a hundred-point scale means it's very obviously subjective.  You're right: it's tricky to say what's a 67 and what's a 69, and I'd hope that the result is people understanding that it's one writer's educated opinion.

The thing is, the anti-percentage crowd always seems to level that same argument. It assumes there's a quantifiable difference between two games that should be reflected in the mark, but that's not how I score things. It's a gut reaction.  You don't sit there, as a consumer, playing a game and taking notes about what's good and what's bad.  You just think about whether you're having fun / being affected / whatever.  It's an instinctive reaction, and if you approach it like that, you can totally differentiate between the two numbers.

( I also think it's probably less true these days that publications only use half the scale.  We seem to have got the hang of that one now.  And the other thing to remember is that the very worst games often don't land on our desks, for fear of that reaction.  Trying to pull the average into the '5 range' is ill-advised, because the "average game" - as in, the quality we most commonly see - probably is better than the middling score. If that makes sense.)

Personally, I use a grading scale. F to A+ It would work like a numerical scale, but I really dislike the whole number thing, especially since numbers can get really weird. What's the difference of a 6.4, 6.5, 6.3?

The other thing I try not to do is compare games to each other. It's like trying to compare apples and oranges. They are totally different things and should be graded separately.

When I have a chance, I'll go into a more detailed diatribe on the grading scale I use.

I think review scores are a necessary evil for a mainstream site. The problem is that I want to, like someone said above, utilize all the numbers, not just the top half of them.

We used to use A+ through F, then we moved to a 1/2-star-to-5-star rating system. Recently, I had a PR guy from a very, very successful publisher freak out because we gave his game a 3.5. Now, to me, 3.5 out of 5 indicates a very good game, but when you have a site like Metacritic or GameRankings translating scores, that "3.5" becomes a "70." And, as we all know, 70 is a bad score for a video game.

But it's not my fault that all these sites have skewed video game scores way the hell high. Am I supposed to give most games four-to-five stars to avoid giving them a 70 on Metacritic? To me, the scales are not equal. "80" makes me think of a decent, B- game, but "4/5" makes me think of more of a B+/A game - more of a 90 than an 80. Yet that is lost in the translation.

At least in years gone by, most movie reviews used a star system, and 2/4 or 2.5/5 was an average movie, worth seeing. With the current game scene 2/4 or 2.5/5 is translated as "50," which is an F. So, really, half your available scores are an F. Ridiculous.

Hey it's math class here! Just kiddin'. I prefer a grade under the piece of words I have just written, just to clarify what my point is on the game and especcially if you can't make out if some is just enthousiast or just being sarcastic.

@Nick

Yeah, Metacritic's pretty fucked up, as we all know. I do wonder, though, whether the best way to avoid having your scores skewed is to just mark out of a hundred and be done with it.

If Metacritic wants to use your scores, it will. And it will change them without giving you a say. It's pretty set in its ways like that. So not awarding marks isn't going to help, 'cause Metacritic will just start making shit up. If Metacritic is an issue for you, you've got to start jumping through hoops.

Of course, more preferable is that you stick to your guns and suggest the publisher looks at, and makes the effort to understand, your own scoring system to get an idea of how much you enjoyed a game. And perhaps put in a plea that this horrible publisher stops using aggregator sites to work out whether their developer made a good game or not. It's the height of laziness, totally inaccurate, and wildly unfair on the people who pour their effort into the creation of a title.

I definitely try to use the whole scale of whatever system I'm using. There's nothing I detest more than (well, there are many things, but few that crop up so frequently in my life) than 7/10 equating to "Average". 5 should be average for funk's sake, 4/10 shouldn't mean "Awful", merely "Flawed" or whatever.

That said, since I started properly reviewing games I have come to hate attaching an arbitrary digit on the end. Useful as a quick guide, perhaps, but awful otherwise. Case in point: my Elven Legacy review. I gave it a 4, which to me isn't an awful score, but there's a fanbase out there for games like that which I'm simply not a member of who couldn't give a shit about most of what I complained about.