I am really confused about these C-Index's

One of my images was a 91:100, and today it jumped DOWN to a 82:100
Another went from 89 to 87, and one of them went from a 78 to an 84 which I thought was not a very good pic compared to the others. I just wanted to know how this could happen. They don't seem to be very consistent.

c-index may change drastically, it depends on many factors like the votes, number of views, etc ...

look on this topic 'Official: c-index re-normalization '. it's one the second page of the
Discussion Board.

Thanks for the heads up, nothing is consistent around here... >_<

Life isn't consistent so why should a website be. ;)

if the c-indexes are changing very drastically you are probably looking at the initial c-index, and then looking at them again the next day. the initial c-index of an image is usually a false one, especially if you uploaded your image close to the time of the c-index calculation. for instance you may have only 1 vote on your image and its a 9 and 4 downloads out of 4 views when its calculated. You get 5 votes and it shows u the calculation it made when it had that one vote which is really high, but when it gets calculated the next day with 7 votes averaging 70 and 10 of 15 downloads to views it will be much lower. Typicaly wait a day or so to get a good idea of what the "stable" c-index will be :)

Thanks for the input guys, but my image has about 300+ views! but alot less downloads, maybe that's why....

Take a look at the number of votes on your Caedes Control. The fact that you have loads of views but not many downloads will also be a big factor in the c-index calculation. Having said that, with the latest formula, votes is more important than the download to view ratio:

cindex/100=1+e^(-favorites/10)*((downloads/views+2*rating)/3-1)

Yea, I saw the formula, thanks.....

I wish they would choose a formula and keep it! :)

by "they" you mean "he" and its only been change once
I think it should be changed again personally, to perhaps (downloads/views +3*rating)/4
but thats just my personal opinion, the system works pretty well as is.

I think it was changed twice actually, unless that post that CaptainHero notified me of was old, who knows...

just once, the captains post was the change caedes made to the cindexes recently

If I update my favorites list and replace some old with new, does the older removed image lose any points? Hope not.

I would imagine that as soon as an image is removed from someone's favourites list it will lose points. Having said that, the newer formula is geared less towards favourites anyway.

Ooh. ;( Don't want that to happen, then again, don't want 50 pages of favorites either. hmmmm. Thanks Matthew.

I am very wary about putting any images in my faves (just as I don't hand out many 10's). However, I must confess I very rarely revisit my favourites gallery! I think I might do that now actually...

I would think that the favourites (I put the "u" in for all you European types) should play a very big role on how an image is ranked, votes next, and downloads over views last. Does anyone agree? Then again, I am the kind of person who likes things small, simple, and clean. I do my best to keep only one page of favorites, but I have seen many who don't. I have one picture with an 88, but that will change as soon as it gets its second vote.The c-index is designed to be comprehended by much more complex mathematical minds than myself, so I won't even try to debate it. Thanks for all you who do, especially caedes who makes and changes it, you make the site so much better, even if the rest of us can't understand. :-)

The favourites used to be weighted more, but that was reduced with the new c-index. I think that is a good thing.

Favourites should be a factor, but not too much. The view:download ratio is important and the actual votes are very important.

the cindex is a very simply formula stuffinstuff, if you want an easy way to think about it. Consider it (minus the favorites) the average of your downloads to view ratio and your raw rating. accept now its the average of downloads to views and raw rating and raw rating again :p

the favorites part of the formula simply brings the number closer to one, think about it as making the difference between the c-index and 100 smaller and smaller with more favorites (the difference between the c-index and 100 is divided by a scale related to the number of favorites). but I guess that part is a bit more complicated to visualize.

I guess some graphs would be in order if we were to discuss it in much more detail. =)

Well there are a lot of images that really deserve to have much higher C-Indexes. Too bad though...

You can just tell that all the effort does not deserve a 67 but more like an 89 atleast.

The whole point of changing the formula was that there were too many bog-standard images with high c-indexes.

well I wasnt trying to go into detail caedes, just attempting to make it easy to "visualize" the math for non mathematical minds :)

C'mon, draw up some graphs, noobguy!

dont worry I did some 3d graphs of the new formula when it was released just to see how hard it really was to get a 99.

I would say its near impossible if but only for a few minutes if you submit your image right b4 the c-index calculation for the day, and you have a couple really high votes and all the first few people who click on your image looked at the full size

but an image with a 99 after being on the site for 24 hrs is simply not going to happen.

Hmm..... Lets just beat this post to a pulp :D
Is there a way to unsuscribe for a topic I made?

Click "Edit Profile" under the Control Panel on the left and then click on "Edit Subscriptions" in the middle ;)

You can also click on the same link "unsubscribe to this discussion" near the top of this page.

Ah, I see it now, thanks.

Just remember, as I have said before, when it comes to moving images to the permenant gallery (and, at current, removing them) the c-index is not really a factor. So don't worry about it too much if you think your images have a lower c-index than they might deserve.