Forums > MobyGames > 67 doesn't seem right for a B-
yellowshirt (1583) on 8/11/2016 5:00 AM · Permalink · Report
That's just way too low. It should at least be a 70 because when I see the orange on a review I immediately assume it's a negative one, which B- is certainly not. Maybe make A+ 100, A 95, A- 90, that way the B- gets a 3 point bump to 70. I've never seen anyone give a game a "F+ or F-" so I don't know if it's worth having those so maybe that will possibly balance it out.
Harmony♡ (21848) on 8/11/2016 5:54 PM · edited · Permalink · Report
Strictly speaking, orange is for mixed/average reviews, but I do agree. 70 is the minimum score for a blue/good review, which lines up with what I think when I see B-.
That said, I wasn't here when this scoring stuff was determined, so it's possible there's a good reason for it being the way it is now.
Victor Vance (18099) on 8/12/2016 3:28 AM · Permalink · Report
I think it is just mathematically converted.
Tracy Poff (2094) on 8/15/2016 8:51 PM · Permalink · Report
It's specified in the code, but, yes, it's basically just evenly dividing up the scores from 0 to 100:
'A+' => 100,
'A' => 100,
'A-' => 91,
'B+' => 83,
'B' => 75,
'B-' => 67,
'C+' => 58,
'C' => 50,
'C-' => 42,
'D+' => 33,
'D' => 25,
'D-' => 16,
'F+' => 8,
'F' => 0
Cavalary (11445) on 8/16/2016 12:08 PM · Permalink · Report
With how scores generally are given, with 70% = average, I'd go more with smth like
A+ / A = 100
A- = 95
B+ = 90
B = 85
B- = 80
C+ = 75
C = 70
C- = 60
D+ = 50
D = 40
D- = 30
F+ = 20
F = 10 (could also be 0, but you don't usually get scores of 0 on any scale)
Sciere (930479) on 8/19/2016 6:25 PM · edited · Permalink · Report
Edit: disregard what I wrote before. It seems we use the same interpretation as Metacritic and they probably thought a lot longer on it than we did, since their averaged rating carries a lot of weight in the industry. I don't like the letter grade system at all, but I wouldn't change it if they do it that way.
Tracy Poff (2094) on 8/19/2016 6:49 PM · Permalink · Report
I think there's a fallacy in your reasoning there. Those letter grades in the US school system correspond (more or less) to something like a percentage of correct answers on a test, but letter grades on games are basically not different from star ratings--they don't generally correspond to some particular rubric with X% = grade letter Y, and even in the odd case where the reviewer is using something like that, they are very unlikely to be using a conversion that corresponds to school grades.
There are practical reasons not to do that, too. Under the system you listed, a game rated F--the lowest possible rating--would have almost exactly the same effective score as a game rated three stars, or 7/10. A hypothetical game rated by every reviewer with an F--essentially the worst game possible--would beat out Mario Party 5 and most other 'average' rated games, and one rated with most C and D would beat, say, Kirby's Dream Land. The problem is that every other rating system would be zero-based, but the letter-grade system would start at 69/100, so it'd be impossible to make meaningful comparisons.
That said, it might be better in the case of letter grades especially if we simply showed the actual rating, perhaps with the title text ('tooltip', hover text) set to display what we covert it to (the 'normalized score'). This way we keep the ability to rank the reviews according to how positive they are, and calculate overall critic scores for games, but we also clearly show the rating as given by the source. What do you think?