Caedes

Non-art Website Issues

Discussion Board -> Non-art Website Issues -> Please Explain Distribution

Please Explain Distribution

RUBrite
09/29/05 8:44 PM GMT
Hi everyone. I love this website, but am totally confused by the new c-index distribution and score. I've looked at a number of images today and the numbers shown in the distribution simply do not conform to the distribution average shown - - most index scores being lower than the spread of scores would suggest. Would someone please explain how this number is calculated? Thanks.
0∈ [?]

Comments

Post a Comment  -  Subscribe to this discussion
XYZ
09/29/05 9:07 PM GMT
I think the average is made and then a few other things are worked into it to make the C-Index, I don't think the numbers you see are the only things are used for the C-Index.
0∈ [?]
::kjh000
09/29/05 9:21 PM GMT
I'm not sure but I believe the c-index is a calculation of the statistical trend in vote distribution rather then whipping up the mean value of the ratings. I'm sure downloads etc. are weighted in there too. It's a new system so we're not really sure I guess. ^_^ Only The Maestro himself can tell you for sure.
0∈ [?]
scionlord
09/30/05 10:53 PM GMT
isn't their weighting built in anyway?

Basically we don't know. :D
0∈ [?]
'Study the past, if you would divine the future.' - Confucius
RUBrite
10/01/05 12:00 AM GMT
It would seem that there must be some weighting going on scionlord, as I viewed a number of abstract (my favorite) images today and the distribution itself would suggest a significantly higher average than what is reflected by the c-index score. It would be helpful to know how this is calculated.
0∈ [?]
+Samatar
10/01/05 12:03 AM GMT
Unfortunately it would also be helpful to the vote attackers. Hence caedes has (wisely in my opinion) opted not to reveal too much detail about exactly how the c-index is calculated. Basically though it is weighted dependant on how individuals normally tend to vote.
0∈ [?]
-Everyone is entitled to my opinion- Visit the new improved rescope.com.au
prismmagic
10/01/05 12:29 AM GMT
You mean Sam that their a people on the site that would do that kind of thing. so to hurt someones ratings.
0∈ [?]
Art is the perception of the creator. Meaning is the perception of the viewer. acceptance is the perception of society.
MiLo_Anderson
10/01/05 1:01 AM GMT
have you lived in a hole?:P
0∈ [?]
"A piece of toast with butter always lands butter side down, and a cat always lands on its feet. What happens if a piece of toast is tied butter side down to the back of falling cat? Does it hover above the ground in perpetual indecision?"
EmilyH
10/01/05 2:02 AM GMT
I am also confused. People have told me my pics are good, yet someone keeps voting 1's and 2's on them. So they are low, yet they have also received high votes from somebody (probably the commenters who thought they were good). So if several people vote a 1 because they are pissed, and one person votes 9 and one person votes 8, the 1's win just because there are more of them even though it was good?
0∈ [?]
RUBrite
10/01/05 2:12 AM GMT
"It is weighted dependant on how individuals normally tend to vote?" Boy, that's hard to do accurately unless you have an awful lot of someone's votes to establish a true pattern. All I know is that I tend to be drawn to abstract and computer generated art and I'm seeing a number of such images with distributions leaning to the right, yet their c-index scores are leaning way left. I've seen this pattern on many different images by many different artists and it doesn't make sense at all to me.
0∈ [?]
::regmar
10/01/05 3:23 AM GMT
The reason (I believe) for the distribution panel, is to allow you the artist to see the deviant votes and ignore them. You should also watch your download number and compare it to your view number. Among my images a 3-2 ratio seems to be good. Also I often have images that get trashed by vengeful children, but when I see the number of downloads I am comforted.
0∈ [?]
ж Regmar ж
RUBrite
10/01/05 4:52 AM GMT
Doesn't the number of downloads simply reflect if viewers looked at the image full size? If images are now looked at full size as a mandatory condition for voting, do downloads really mean that much? I'm confused about this too. Are "selective" downloads counted as well as "random" downloads and how are "selective" votes being tallied and considered in the overall evaluation of images? I know when I vote in the booth, I rarely get to vote on images of the type I like and am required to vote on many of the type I would not normally look at. So I am now "downloading" images I would not normally take a closer look at. I would think that would diminish the significance of the number of downloads.
0∈ [?]
+Samatar
10/01/05 4:54 AM GMT
I am fairly certain veiws in the voting booth aren't counted.
0∈ [?]
-Everyone is entitled to my opinion- Visit the new improved rescope.com.au
+tbob
10/01/05 5:17 AM GMT
Future posts of this nature should be posted in the "Non-art Website Issues" section of the forum not here.
0∈ [?]
EmilyH
10/05/05 2:59 PM GMT
Hey, where did the distribution index go?
0∈ [?]
::RobNevin
10/05/05 11:28 PM GMT
I think the distribution index got too many zero votes, became upset, deleted all it's images and left the site with a bruised ego.
0∈ [?]
You're invited to tour my gallery ••• บนบนบบนนบนนบนบบบบนนบนบบบบนนบนบบบ
::kjh000
10/06/05 11:03 AM GMT
That's too bad I really liked that feature.
0∈ [?]
::regmar
10/06/05 1:09 PM GMT
I also liked the feature. Using it I was able to "factor out" the anger votes when looking at my c-index so as to produce an "r-index". This helped me weigh the relative merit of my image compared to the mediocrities that have 90's. Wasn't that the whole point of using it? Did something happen that we weren't aware of?
0∈ [?]
ж Regmar ж
::WENPEDER
10/06/05 4:18 PM GMT
Ditto that, Reg. The Distribution helped me understand where most raters viewed my images, as opposed to those who seemed to have a bone to pick, for whatever reason. Now we just have a number which is often low, with no distribution reference to help us make better sense of that number. My guess is that more things are going into the C-index than raw score averages and, because people were expressing confusion over the lack of continuity between the C-index and the distribution scores, the decision was made to take the distribution down. I, personally, would rather see the distribution than the c-index or would like to have a better idea of what exactly goes into calculating that index..
Wen
0∈ [?]
EmilyH
10/07/05 3:58 AM GMT
"I, personally, would rather see the distribution than the c-index or would like to have a better idea of what exactly goes into calculating that index."

Same here. That way, we could get a better feel for if one or two people are casting protest votes and the majority of people like our stuff, or if we really need to improve on something.
0∈ [?]

Leave a comment (registration required):

Subject: