[Baby Bonnie Hood]

What the title says. What determines the game that'll be displayed every time the main page is refreshed? I noticed that Talon's Tale never showed up until I reviewed it. Is there some kind of system?

Asking just out of curiosity, and totally not because it never features any of my games.

[mzxgiant]

It has to have at least one review, and the average reviews must equate to at least 4 stars (it might be 3 stars, I don't recall off the top of my head). Once it has narrowed it to that set, it randomly selects from there on each refresh.

[Baby Bonnie Hood]

Ahh, it's automatic.

And yeah, minimum is 4 stars. You think it should be changed to at least 3 stars? There are plenty of good 3-star games like Chronos Stasis and Catacombs of Zeux that don't get featured.

[KKairos]

You could also give them 4-star ratings yourself if you believe that's what they deserve. That said, I don't think it's a bad suggestion because I think sometimes games get rated lower or higher than they deserve based on only one rating. Example: Micah's Geneva demo being a featured game because of a review he gave it.

[Lachesis]

Perhaps the featured random games could be manually selected? It sounds like a pain in the ass until you realize that there are only about 50 games at maximum that should be appearing there. We could also have voting and stuff for games that aren't added that people think should be there.

[Dr Lancer-X]

Quote

Perhaps the featured random games could be manually selected?

That's how it used to be. The reason it's now a 'random game of the moment' is that the old Featured Game was never updated on a timely basis.

[Lachesis]

I mean, you'd just queue 100 games or something and they'd randomly rotate just like they do now. Decent new games might take a while to get added, but that's because new games hardly ever come out.

edit: of course, our dev bailed out on us, so this isn't ever going to get changed anyway

[mzxgiant]

Lachesis, on 20 January 2011 - 04:19 PM, said:

edit: of course, our dev bailed burned out on us, so this isn't ever going to get changed anyway

fixed that 4 u

[Baby Bonnie Hood]

Loco by zombieguy is a great game, yet it's not featured. Why? Because amidst all the high-starred reviews, someone gave it only 1 star (though it claims it was supposed to be a 4). Thus, it drops below 4 stars and falls off the random game radar forever. Now, is that right?

[zzo38]

I think there should be a small chance of selecting ones that to not meet the threshold, too; like 95% it selects from the minimum 4 stars ones with reviews, 5% it selects from all records in the database (other than hidden records) (regardless of reviewed or not, average rating, and so on). This way is possible someone will notice other one in case it is wanted to review those, or in case some of the ratings are disagree by some people sometimes.

This post has been edited by zzo38: 15 April 2011 - 08:27 PM

[Lachesis]

I think we should assign star ratings to areas of the normal distribution, i.e. |x|=[0,.5] for 4 or 5 stars, =[.5,1] for 3 stars, =[2,3] for 2 stars, and =[3,inf) for 1 star and unreviewed games. Ideally, you'd derive the Std.Dev. from the equation for the normal distribution manually (no pre-made functions or hard-coded approximations, because those would make sense).

[zzo38]

Lachesis, on 15 April 2011 - 09:09 PM, said:

I think we should assign star ratings to areas of the normal distribution, i.e. |x|=[0,.5] for 4 or 5 stars, =[.5,1] for 3 stars, =[2,3] for 2 stars, and =[3,inf) for 1 star and unreviewed games. Ideally, you'd derive the Std.Dev. from the equation for the normal distribution manually (no pre-made functions or hard-coded approximations, because those would make sense).

Yes, the normal distribution might work, try this. This way, it also solves the problem, possibly better than what I have, maybe.

[Wervyn]

Lachesis that was beautiful, I just have to applaud you for a fantastically successful troll.

No really, the actual solution is to implement the infrastructure necessary to easily administrate reviews. But since no one really feels like doing it...eh.

"I'm a master of subtlety! Shabam!"

[zzo38]

Wervyn, on 16 April 2011 - 07:44 PM, said:

Lachesis that was beautiful, I just have to applaud you for a fantastically successful troll.

No really, the actual solution is to implement the infrastructure necessary to easily administrate reviews. But since no one really feels like doing it...eh.

"I'm a master of subtlety! Shabam!"

Not everyone can agree on the review either; I think a solution is a normal distribution, is much better solution. Although deletion of reviews that contain real nonsense is also something that should be done (but only if it is really nonsense, and something you just disagree with doesn't make it to be nonsense).

In the mentioned case, whoever rated it 1 star obviously made a mistake (read the message, the reviewer obviously meant to put 4 stars), so it should be corrected to 4 stars. However, if someone does genuinely give it 1 star, it should remain with 1 star.

Actually, I think unreviewed games should be considered to have 2 stars for the purpose of normal distribution, not just 1 star. This means games will be reviewed more quickly, in which case it may be given 1 star if that is so doing, or 2 or more stars if that would be the case instead.

This post has been edited by zzo38: 17 April 2011 - 05:24 PM