So I've been looking at the game ratings and I've noticed that yoyogames uses a six star system, yet there is like one game that even has five. Also there's a lot of games that have a four stars, which I think are not as good as some of the classics, such as jumper or seilkius. Also its not fair for games with a few ratings to be judged the same as those who have been rated more times.
So I think that yoyogames oughtta fix their rating system.

While some suggestions can be helpful, vague suggestions are not.

Warrior, glad to see you're back.

Hasn't this topic been made a bunch of times though?

-PF

You have a point about games with few votes. Maybe games that are rated less than... say 5 or 10 times will be left as unrated, and current rating not shown. This would as least prevent evil downraters to vote 1/6 on a perfectly good game =).

No it wouldn't.

It would prevent their vote from mattering...

...Until four more people made ratings.

-PF

Or the same member with 4 more accounts.

Anyway. Getting rated is overrated.

I agree 100% with the text I made bold. The most annoying thing about the system is that there allways is many bad games with one 6-stars-vote at the top of the gamepage. Its annoying when you want to play a really good game and hits a "hit the clown" thingy.

Also I would really like to see a more controled "Staff choice" system. It would be wonderful if the yoyo-staff played every game that has a high rating, and also deleted some of the worst games from the "Staff choice" list. If the list could have a max of 100 games, the moderaters would have to take bad games off the list.

- Jenner

All I see here are people complaining about how things are sorted but I don't see anyone actually proposing different algorithms for ordering the things on YYG. If it's so "trivial" to fix, then you propose an algorithm that sorts by two separate criteria in the way you want.

Very well then, here are my suggestion:
-First off the rating system should be 5 stars instead of 6
-The amount of stars should be based on the score compared to other games rather than the rating itself
-games should be rated a certain amount of times before they can get to the top of the page, 20 votes for example
-if the game has been rated by less than 100 times, one star should be taken away if its more than 3 stars

This doesn't address the problem at hand and is, for the most part, totally irrelevant. Just pretend it's out of 5 and 1 star is actually 0.



An interesting thought, but it's really only a cosmetic change. The games still have to be sorted by their rating so you're just rescaling the number of stars without changing the relative ordering.



Seems reasonable, but it does make it more difficult for new games to show up unless people actually go looking for them (I can't remember if there's a "most recent" category but it's likely that there is). I fear, however, that this might only encourage people to solicit bogus votes from their friends or fake accounts, but there's nothing that can realistically be done about that.



Why? Is a game with 100 votes really guaranteed to be that much better than one with 99? I mean with a lower limit a line has to be drawn somewhere for showing up on a highest rated list, but this suggestion artificially and unfairly changes the ordering based on the number of votes. A game with a poor to mediocre rating could easily show up before a higher rated game with just a few less votes.

I still don't think the issue has been properly addressed.

Seriously. What if you are the first person to stumble upon an excellent game at YYG and it is truly desrving of a 6-star rating? Well, too bad because since 100 people haven't been there and rated it, it is not rated as high as it should be. That doesn't make any sense at all.

Perhaps there could be a section sorted by RPDs. (Ratings-per-day) That way, if your game gets immediately shot down, at least for a couple of days it will have a high RPD to give it some chance of staying up there.

And/or they could have a section of Random games with at least 5 plays. (to dis-include all the n00by games that really don't deserve any attention)

If you guys spent half as much time working on your games to make them worthy of 6 stars as you did complaining about how your game never got a fair shot, we wouldn't need to be having this conversation.

The reason why I'm complaining is because so far no game has gotten 6 stars

If a game is truly deserving of 6 stars it will get more than 100 people to vote for it

That is assuming the average GMC user has that many friends to begin with, LOL.

I'd say recognition for your hard work is more than a mere cosmetic change.Its more fair to compare GM games to other GM games, not professional games, like some people currently do.

People tend to be more picky on a 6 star system and it also makes it harder to judge. I agree its not that relevant but I do think it would help.

The criteria by which people evaluate games cannot be changed by the software. The only way to change that is to go to their house and beat them with a big stick until they agree to evaluate a game based on your criteria.

Using the ratings that people submit to determine a game's ranking relative to the ratings submitted for other games is a cosmetic change because it does nothing to change the ordering of the games, which is what you say you want to do. For example, giving the highest rated game on the site 6 stars and the lowest rated game on the site one star still leaves the order of "highest rated" the same, which is what you apparently have a problem with. Essentially what you're proposing is a normalized scale similar to the way IQs work. The median game by rating would be at 3.5 stars and the best and worst would be at 6 and 1 respectively with the remaining games distributed appropriately in between. As I said, it might be interesting, but it won't have any effect on how these games are rated, just what their rating appears to be. That is unless every time someone wants to rate a game they have to go through the entire list of games hosted on YYG and mark off which are better and which are worse.

I agree with Yourself. And I don't think anything discussed in this topic can be changed, without adversely affecting something else.

The most common complaint seems to be "games I like aren't rated high enough, and lousy games get higher ratings". Unfortunately, the hallmark of a democratic rating system with many choices is that most people won't like the result. Changing the scale, or the number of votes needed for ranking, won't change the ultimate ordering. Why should it?

And regarding those handful of bad games that temporarily get rated 6/6... I say don't worry about them! They have their moment of glory and then disappear. They don't have any long-term effect on the other games.

I said this once when this was brought up before:

"Two" ratings: Like it, or no rating at all. More likes could basically mean its a better game. Anything being said could be handled in the 6 star system for individual ratings for graphics, sound, gameplay, etc. More detailed information can be given in comments and reviews.

I agree with KC LC though, just ignore the crappy games with 6 stars. Whatever to them.

Just get rid of the rating system.

Base the most popular games off the amount of downloads.


I don't get why we even need a rating system, it just creates bias towards which games are good, and which aren't.

I like IMDB's method, which is keeping a movie from showing it's rating until there has been at least 5 votes.

Better yet, lets make sure all games on yoyogames are high quality. Keep all of the staff picks on. Delete the rest. Give the members who have created a staff pick the ability to choose whether a game makes it on to yoyogames based on quality. But they can't put a game on if they've made it. So, to reword that if someone makes a game, and can vote. They can't vote for there own game.

Could be a fair system?:

x = votes
y = rating (every time it is rated the number they rated it is added to this variable)
z = number of unique downloads
q = number of stars possible

( ( y / x ) * z ) / q;



It's not really perfect, but maybe all our mathematicians on the forum can pull together and hopefully make a new formula. HOWEVER, I do agree with FredFredrickson in that you should stop complaining about this issue and just make a great game worthy of six stars.

it is confusing having two different rating systems, maybe they should be merged?

Also in order to avoid people rating games that they dislike 1 star just so their vote lowers the score, could there be a bar showing how many people voted for each value, as opposed to the average?

Mentioned previously in the topic was the idea of having the game stay at "unrated" until it gets a certain amounts of votes. However, the game would not be visible. So what you (Yoyogames, doubt they even read this topic, though) could do is have all games with less than 5-10 votes be treated as a having 3 stars. Once it hits 6-11 votes, it's real rating will show. That way, rating it six stars won't make it appear at the top of list until it gets 5-10 more six star votes. Same for getting one star.

Ok so let me try that out, lets do it with the game "Visit"
Ok, so its been rated by 191 members. That means x=191
Its been rated with a 3.9 (lets say 4) ok so its been rated with 4 stars, so y=4
Its been downloaded 5164 times so z=5164
And it uses a 6 star system so q=6
all put together its ((4/191)*5164)/6= 18.02...
Not good, but nice try

You may settle with a bad rating system, but I sure don't. Its not only bad to the users, but it also gives yoyogames a bad image.

Precisely.

Good games get noticed, bad games get trashed. Spammers try to make your game seem bad no matter what, but if your game is truly good with a lot of time spent into it, people will just ignore the pointless spam.

Now, about the rating system. It's fine. I, personally, always like to have a rating scale either 1-10, or 1-100. But oh well. That's just me. I think the only way YoYo could improve their rating system is if they only allowed their staff members that play the games(if they have specifics.) to rate other's games. Still, that could maybe only be it.

the only reason that ****ty games have a better rating than all the really good games is because jealous ******s make multiple accounts so people can play their games

Its not only that. Its also the fact that no game has or will ever have 6 stars

Of course they won't. It is strictly impossible to create a game which everyone would view as perfect by any standard.

The only problem I have with the rating system is the granularity(is that the right word?), it appears to me that the games are only sorted to the nearest tenth of a star. I believe it would be better to go to the hundredth or thousandth... if they are already sorted like this it would be nice if it was shown.

There are professional games that do have a perfect rating, just look up Zelda ocarina of time. The same should go for GM games, but like I said, that will never happen if we keep comparing them to other professional games. The only way we can change this, is if the rating system compares the score to the rest of the scores of other GM games, rather than taking the rating system literary.

Having a perfect rating in a game review, is not the same as having a perfect record, after hundreds of people have rated a game.

But YYG already does this. Under "most popular" the games are rated highest to lowest. The highest rated game gets 1st place, with all the other games listed in order of their ratings.

Isn't this the same thing as the "relative rating"? It compares the scores of all the games to each other, just like you want.

It only takes one person not liking it to any degree to take away the "perfect" qualifier. I've met people who don't like Ocarina of Time. Therefore, it's not rated perfectly.

A common alternative star rating system is one based on a 'scoring system' then star ratings are based on wether the score reaches a required bounbary for a particular star 'level'.

e.g. (Just an example can obviously be modified)


Then you have arbitary boundaries:


The games can then be ordered exactly on the score they have got.

Well currently the highest rated game has the highest place, but not the highest rating. If the rating were compared to other GM games, games on the top spots would have 6 stars. Like Yourself said earlier its a cosmetic change, but I think its an important one.
I also like the idea of ratings not being shown until a certain amount of votes is reached as mentioned earlier.
And there are some games that have the same amount of stars as the best games without really deserving it, such as
http://www.yoyogames.com/games/show/39857
My idea was that a star be taken away from the actual rating until it reached 100 ratings, but most people didn't like that idea

You're giving stars based on the number of stars you got. It doesn't make sense.

Why? That's what is done at the moment.. a star rating is given based upon the mean value of the stars rated.

Well let me get this straight
if a game has 5 stars it would be
1*3+2*2+3-4-5*2=-4?

No?

That would be if it had 1 - 5star rating, 2 - 4star ratings, 3 - 3star ratings, 4 - 2star ratings and 5 - 1star ratings ???

Well give me an example cause I don't get it

You want to wage warfare against the rating system? Here's our attack plan:

Write a review.

I have one idea. Maybe their should be a staff rating as well as a player rating. The most popular section would still remain on the front page and be based of players votes but a new section, Staff Plays, or something like that would be based on what a staff or moderator (or respectable person). This would still leave out many games but those are pretty much left out already and the more popular game will have ratings that aren't based on overraters or 1 star raters.

Really?

OK example:

Obviously it's bias towards games that have been voted on a lot.. perhaps a limitting factor based on the number of votes would be best introduced to make this bias less.

The word is 'biased.'

Of course its always going to be bias.. but you can lean the bias in the desired way to get the desired results. In my opinion looking at it at the moment, I personally dont think it looks very good for yoyogames with only games with 4star ratings on their main page and no 5star ratings.

I've been messing around and the best I've come up with so far is (with new variable n included being the number of votes):


Then the boundaries are changed to:


Plug it into excel or gm or something with some values and test it, it probably still needs some work.

Surely having some relativity between rating and downloads/plays would make sense.

That's not really a good way of rating.. then the games that just 'look nice' or interesting will get rated a lot higher than the ones that are actually nice. And this can pretty much be found using the 'most played' tab.

At first I would have complained that this should be on the YYG forums, but then I realized the amount of thinking this requires and decided to hold my tongue. Heheh, sorry YYG forumers, just a friendly joke.

I say that we should take the quick-rate system away and force people to write reviews. On top of that, enfore a minimum character number for the review's pros and cons; in this, people not only have to play the game, but have to know a bit about it.

Bla bla blah bla bla blah blaaa blah bla bla bla bla blah bla bla blah blaaa blah bla bla bla bla blah bla bla blah blaaa blah bla bla bla bla blah bla bla blah blaaa blah bla bla bla bla blah bla bla blah blaaa blah bla bla. There you go a 50 word essay on why that wouldn't work. Just kidding, i think that would reduce the amount of rating a lot and might leave some great games from being played if no one rates it at all.

How would it reduce the number of plays? People are already complaining that great games have low scores, meaning even with the low rating they are played. With no rating at all, someone who wants to check new games out will find it. By your logic, 0 games on the YYG site have been played because they all started with no rating.