forum

I think Rating Spread should be removed [duplicate]

posted
Total Posts
14
This is a feature request. Feature requests can be voted up by supporters.
Current Priority: +0
Topic Starter
LuigiHann
This map has been deleted on the request of its creator. It is no longer available.
anonymous_old
They are dickheads. All seven of them.

I'm not sure if the rating spread should be removed as much as the rating system redesigned, though. A "thumbs up"/"thumbs down" kinda thing would be cool.
Ekaru
This map has been deleted on the request of its creator. It is no longer available.
Takuma
This map has been deleted on the request of its creator. It is no longer available.
CheeseWarlock
I'd be fine with a "like/dislike", and that would pave the way for weighted voting:

voter.relativeweight = 1 - ( abs(voter.likedmaps - voter.dislikedmaps) / voter.numvotes)

The voter's vote weight, defaulting to 1, is decreased by the difference between their + votes and - votes with respect to their total number of votes.

100 liked, 0 disliked: no weight.
4 liked, 2 disliked: 67% weight.
1 liked, 1 disliked: 100% weight.

Edit: maybe even have a slight weighting advantage to people who vote more often?

voter.relativeweight = 1 - ( abs(voter.likedmaps - voter.dislikedmaps) / voter.numvotes) + 0.15 * log(voter.numvotes)

Maybe I'm overthinking this. Oh well. FUCK YEAH PSEUDOCODE
anonymous_old
With the current system weight wouldn't work for me because I rarely vote anything other than 9/10 on maps I like. I fell bad voting lower, because I know it affects the mapper's rating (even slightly), and I don't want to bring them down at all.

I'm not sure if in a thumbs up/down system I would thumb down, either, so weighting wouldn't work there either.

Also, with your little formula Cheese, if you vote only once, you'd have no weight at all. D=
Derekku
I really don't care much about ratings. I never pay attention to them. :/

A different system would probably be better though, so +some kind of support from me ;p
Ekaru
@ Takuma: strager's post wasn't there when I made my post, he ninja'd me. So there! *stick out tongue*

Maybe we should have a poll about this. Just a thought...
Mogsy
This map has been deleted on the request of its creator. It is no longer available.
Gabi
no.

there will always be trolls, no matter what system, so either remove voting completley or keep it as it is.

EDIT; maybe make it seperate vote for song, and vote for map, however that will never happen lol
awp

Gabi wrote:

EDIT; maybe make it seperate vote for song, and vote for map, however that will never happen lol
Even if that doesn't fix anything, I find this interesting.

If you want to know what people really think of a map, have it Approved instead of Rated. Not really; it's slightly biased toward favouring because the only people who play Approved maps are the people who want to; they don't feel "obligated" to play something just because it's a ranked map.
mm201

CheeseWarlock wrote:

I'd be fine with a "like/dislike", and that would pave the way for weighted voting:

voter.relativeweight = 1 - ( abs(voter.likedmaps - voter.dislikedmaps) / voter.numvotes)

The voter's vote weight, defaulting to 1, is decreased by the difference between their + votes and - votes with respect to their total number of votes.

100 liked, 0 disliked: no weight.
4 liked, 2 disliked: 67% weight.
1 liked, 1 disliked: 100% weight.

Edit: maybe even have a slight weighting advantage to people who vote more often?

voter.relativeweight = 1 - ( abs(voter.likedmaps - voter.dislikedmaps) / voter.numvotes) + 0.15 * log(voter.numvotes)

Maybe I'm overthinking this. Oh well. FUCK YEAH PSEUDOCODE
A bellcurving approach would work even better IMO. Calculate the mean/standard deviation of each voter's vote breakdown, then standardize all their votes to, say, mean 7, standard deviation 4.

Any kind of vote transformation would rape the database, so I find it unlikely to happen.
Gemi
This map has been deleted on the request of its creator. It is no longer available.
Stefan
I think this Thread can be added for now as Duplicate since t/61247&start=0 is better and more accurate as this one.
Please sign in to reply.

New reply