I do not know if a decision has been made to implement a map rating system in the map DB. I do not think one is necessary, but if you decide to code one, I have a system in mind that should give relevant and accurate results(this is not intended to be a feedback system, but for people that are looking for maps to play).

I will try to flesh out the idea more: Start with a few questions like these:

Preferred game length in minutes ( 5 20 60 )
Preferred team size( 1 2 4 )
Preferred game aspect ( macro both/either micro )
the algorithm would classify these as ( dislike / N/A / like )

These questions are treated by the rating algorithm just like rating a map, the only difference is they wont show up in search results or suggestions(it will just include them in the judgment when weighting the ratings of other users).

There will be 2 ratings for any map: premise, and execution. The latter will not affect the ratings of other maps, but the former will. Each member will have a few variables with respect to every other member that has rated any of the same maps. The first variable is relevance, ranges from 0 to 1, the second is the number of maps rated the same, the third is the number of maps rated differently. These variables are updated every time you rate a map that the other has rated, or vice versa. Relevance = similar ratings/dissimilar ratings.

Now here is the fun part, when browsing or searching maps, the rating displayed is 10*(1+summation(each positive rating*respective relevance)/1+summation(each negative rating*respective relevance))/(Number of ratings)

If there is no database entry for a user pair, relevance defaults to the median relevance score of the person searching or browsing.

Erm... break it down for me more.
Each member has a relevance variable that shows how much they vote alike to other members? So the objective is to find maps that other people that votes just like them enjoy?

Post has been edited 4 time(s), last time on Sep 8 2007, 4:54 pm by isolatedpurity.

yes, it give more opinion weight to people with closer tastes.

IMO, you should just keep it simple and have it so that people can rate, and make comments, other than that I see no real reason to make it more complex than it needs to be.

It is not for feedback, it is about finding maps you like. I have never seen a useful map rating system before, and I think this would simply be an improvement, in the event a rating system is included in the map db.

If anything, a simple 1-5 star rating would be nice. I remember you could rate up to 10 in the old database, but in this new database I would like it to be able to show maps in a certain order. Like say, by, most downloads, highest rating, etc.

If there are basic rating categories needed, you could take a look at the ones we in Clan Oo use..

Here is the rating categories.

You could do that before, I'm certain.

I like the concept. I'm not all to certain of the practicality though.

For a public-oriented system: keep it simple. The problem with breaking down judgement of a map into categories is that the importance of different attributes varies from map to map. For almost any element of objective quality (terrain, trigger design, trigger complexity, professionalism, originality, etc.), there can be found a good map that ignores that element and isn't any worse for doing so. It would be wrong to give a map a lower rating for not attending to an attribute that is irrelevant in its design.

For an individual-oriented system: the idea of weighting the ratings of others by how similar their ratings are to your own is a good idea, as it would provide a fairly easy way to get a general idea of what people who like the sort of map that you like, like. But be sure to also make visible an "objective" rating that does not vary from person to person; otherwise people will get confused when they talk about the maps. Other options that should at least be considered are a "reset" option, that causes your previous ratings to be disregarded (in case your taste in maps changes), and an "ignore" option, so that your ratings won't be corrupted by people who happen to like some of the maps that you like, but for different reasons, rendering their opinions of other maps misleading.

i didn't bother figuring the mechanics of the system out, but "more opinion weight to people with closer tastes" is definitely cool

hm.. So you want a system that allows people to 'weight' the rating system when they look for maps, so that they see maps that are best suited for them?

It's a pretty cool idea mathematically, but I have a feeling that a good categorization/labelling scheme would do just as well for finding maps that are suitable for you. (Given, I guess that's a subset of what you're proposing anyway...)

Besides, part of the fun of searching a maps DB is stumbling across maps you weren't looking for

Hell, I'll probably give it a try. A good map though is, generally, universally a good map. And a crappy one = a crappy one. This might only be useful in the gray area in between, me thinks.

This is basically the only way I could think of to rate and rank mixed genre maps.

As I said before, I think map rating should be kept down to 2 values: premise and execution. The feedback system should be broken into categories: triggering, terrain, etcetera, and should have no bearing on rating.

I was wondering if you could keep "Originality" on one of those factors?

That would be implicit in the premise.