A little statistic I made has stirred up this discussion between Tapuak and me once again. Not that we two disagree much on them, but we've talked about the meaning of these numbers quite often already.

Ratings on TGOD currently go from 0 to 6 (this being the best), so mathematically, 3 is the 'average'. The average rating of the featured games is 4, though, probably for the simple reason that people are usually reviewing games they like rather than dislike.

What I'm curious about is this: How do you all interpret it if a game gets, for example, a '3' or a '4'? If you had to base your decision whether to try a game or not purely on this rating, where would you stop to bother?

Pretty unrealistic situation, but if there wasn't even a genre indicated, I think I'd try anything scoring 4 or higher. Any other information given would probably have more influence though, I really don't think the ratings are that important.

My interpretation: 0,1, and 2 not really worth trying, and 3 possibly being an OK game if you like the genre. 4 would be a decent game worth a shot for anyone, 5 or 6 would mean good or excellent. The choice between the last two seems more a matter of the reviewer's personal taste than actual game quality.

It seems that the real problem occur when people see the number 6. Most normal people who visit the site will never think about the 0 unless it is utterly obviously stated that the 0 is included, and then again some might still neglect it. This mean that, for the majority of the people reading the reviews, the middle bar is not 3 but 3.5. 3.5 rounded up would be 4, so 4 would be good and 3 under average...

This could be solved, I guess, if you had numbers ranging from 1 to 7 instead, with 4 as the midbar.

I'd consider 5 or 6 to be playable, 3 or lower not, and 4 maybe.

I agree on this. This is my interpretation, too.

Edited by dregenRocks at 13:29 on December, 23rd 2004

Eagle: Very good point! The 'problem' is I am, after all, an IT person, and 'we' count starting with zero The more emotional argument how it came to zero being used is that there are games which aren't worth anything. A score of zero perfectly represents this - 1 is already something. Still, you're certainly right when it comes to casual readers.

I'll just try to explain what I mean when I give out these ratings. Some of this might be a little surprising, but I'll make it as clear as possible.

0: total crap, not worth to be in existence
1: has something to it, like professional production values, a somewhat nice idea behind it, but is in the end not seriously playable
2: not that bad; certainly has some redeeming value, but also very serious flaws which, at least at some point in the game, make frustration or boredom overwhelming
3: good game, definitely worth playing for longer than half an hour, but absolutely nothing special to it - a game made 'by the book'
4: very good game, almost all important aspects come together in a good way; a safe recommendation
5: excellent game, a 'must-have'
6: pure genious, the best of its kind

Of course, these definition don't go along with the standard definition of 'average'. If 'average' is measured counting all games in existence, it could never be a '3', because (as Tapuak so correctly pointed out), I can't even stand playing the vast majority of games for more than five minutes. However, there is no way to have different steps for such games as it would be the case if '3' were the 'true average' (at the same time leaving only little room for different ratings for good games). That is why I settled for having the 'average 3' not as the average quality of all games, but as something more positive - way more positive, in fact.

I don't know whether this is somehow understandable. It somehow fits what Mole said, but Johann's interpretation is definitely in contradition to it. So... what would be the most logical choice?

I don't really see anything wrong with the current system. It makes sense to use the average rating more as an average of games on the site, not an average of all games ever made. But then your average should be 3 and not 4 as it is now...Review more crap so the actual calculated average of all games on the site equals 3?

I don't think it's such an important matter anyhow. The conclusion is that 'the average game' on the site is a very good one even by your standards, which is a good thing, right?

Correct. I agree on that again. After all the title of this site is
The Good Old Days
and not
The Average Old Days...

I meant a what would be a logical system to have, not how to get the average rating down Of course, this single number isn't that important in the end, but it should still be understandable, and I was wondering how people interpret these numbers. As I suspected, at least part of the visitors do it rather differently than it's intended. Then again, I guess this problem arises with any system...

Maybe you can add specific titles to each rating, like in the german/austrian school-system:

0 - No rating possible
1 - unsatisfactory
2 - unsufficient
3 - sufficient
4 - satisfying
5 - well
6 - very well
(...for example...)

or would this be to pretending? On the other side, it's the natural aim of a reviewer to pretend something. (?)

Following my own interpretation at the beginning of this thread Mr Creosote, a good and easy way to fix this would to keep your own scale of interpretation but add 1 to every # so 0 would not exist anymore and 7 would be the best.

It would change absolutly nothing in all the reviews and would fix the average I mentioned to those "average readers".

I myself am completely against you when you say that 0 can exist for games. If you have a game, then it can't be "completely worthless". Even the worse games have something good on them, else they would not have past the stage of coding at all. That would be a worthless game, since it can't be played whatsoever.

My reasoning is this, from the viewpoint of NOT me, but just someone browsing the site casually:
-There are quite a few sites with games on the net, with lots of different games, and I can't play them all.
-I'm just browsing here, so I will probably download one or two games, and those will be the most fun games I find
-Therefore, I don't think people will even consider downloading many games with a lower than 4 rating, except if they already knew the game or read some positive reviews elsewhere.

This does open up the discussion if people base their download decisions on the ratings...

Looking at our statistics here, the decision is mainly based on whether people already know a game (my interpretation of the fact that the most common games are also the most downloaded ones). The rating is another pretty influential point, seeing that for example a virtually unknown game like Proliferation (rating: 5) does get many downloads as well.

Still considering what you others wrote, so I'll reply later

Ok, I thought about this a little more over the last few days. Above all, we have to keep in mind that a rating is always purely subjective, making its worth highly questionable. It would get better if several people discussed all ratings and then each gave their own as well as a 'compromise rating' by all of them together, but that would require a core team to seriously play all games - which would be a full time job, so it's impossible.

Good idea in general, but just small titles won't help much. Maybe I'll put up a longer explanation (like the one I already posted here) on the site, so that those who do care (as opposed to the casual reader Johann was talking about) can get the information.

Mathematically logical, but thinking about it, I'm convinced '3' will be seen as the average of a '0-6' scale by the casual reader, too, because '3' is half of '6'. Yes, not numerically correct, but still most likely valid.

That reminds me a lot of what a once famous gaming magazine (I think it was Amiga Format) said. They seriously claimed ratings below 30 (they used the percentage scale) can't be possible, because being finished and published alone already indicates a basic amount of quality. Even then, I totally disagreed with that - every game starts with zero and has to earn additional rating points in my opinion.

All perfectly understandable and most likely true, but then again, when did I ever care much for visitors like that Seriously, though: I don't think there is a solution for such readers, because no matter what the rating scale is, there will always be games which have a higher one and the ones which are 'just' good.

maybe a color coding scheme would help? Or a graph that indicated the 'earned points' and not its rating?

Could you elaborate on this?

One would wonder why you ever oppened a thread about the subject when it is no surprise (at least to me) that you use your status quo and decide to do nothing on the matter.

I mainly opened the thread to hear some opinions how the current ratings are interpreted passively, and not necessarily to develop a new system. Also, you know that I never take any opinions but my own seriously

Color coding could be as follows

0 = black
1 = red
2 = orange
3 = yellow
4 = what ever color best fits between them
5 = light green
6 = green

And then something like a bar graph where an outline shows how much of 6 six would acumulate to somthing like
----------
|XXXXX...|
----------

And let the color vary with the amount of points it has earned.

Maybe it would be interesting to put another bar with the points that the averidge game in the same group have.

As sure as pie!