Attached Link: http://news.morningstar.com/news/ms/Investing101/riskybusinesstwo.html

We have changed how the rankings work in the gallery. The formula is now based on identifying the standard deviation of the rankings instead of the overall average. Once we have calculated the standard deviation, we then multiply it by 2, add and subtract that amount from the true average ranking to create an average maximum and average minimum ranking. This determines the bell curve of acceptable rankings. Anything exceeding the maximum or minimum is excluded and the new average is based upon all rankings within these two limits divided by the number of rankings within these two limits. I have provided a link that better communicates standard deviations if anyone has any spare time and is excited to read about this. :) We hope this is going to provide a ranking system that members find to be more accurate and fair. In order for this to work appropriately, we have to have multiple rankings on one image. To be considered for the ranking, an image must have 5 or more rankings listed. Those with 4 or less rankings on an image will not be displayed in the ranking areas. This does not change anything else with the galleries. This is only related to the Best Ranking areas. We have added a statement within the Best Ranking areas to communicate what image is going to be listed in the ranking area. "Only images with 5 or more rankings will be displayed." Thanks to all the members who communicated this issue to us and was patient until we were able to work through this process. thanks tammy

Thank you for the explaination, I will go over this several times, lol. But, I definetely grasp the concept. AgentSmith

Contact Me | Gallery | Freestuff | IMDB Credits | Personal Site
"I want to be what I was when I wanted to be what I am now"

?

Bryce Forum Coordinator....

Vision is the Art of seeing things invisible...

lol Maths huh Zhann ;) Basically, among other things this removes any obvious outliers which might be having a disproportionate effect on the ranking. As an example, say an image has 7 rankings. 6 of them ranging from Almost There to Great, and one score of Hmmmm. This new method takes the Hmmm comment out of the equation when calculating the overall ranking, which better reflects the opinion of the majority of those who gave a ranking. Make sense?

Ahh so less of a chance for trolls to mess things up. Yes, I did some work with SD in physics, it's definitely a better method of ranking. Great job guys.

Obviously doesn't dent the popularity contest aspect since they post as a group (and I know nothing can be done about that), but I appreciate the effort even though I don't, now that I think of it, participate either as a ranker or rankee (though some people would say I'm pretty rank irregardless).

This sounds like a much better way of determining an honest ranking. I do understand the concept of it and it makes good sense to me. 8) ~ Swade ~

There are 10 kinds of people: Those who know binary, and those who don't. 

A whiner is about as useful as a one-legged man at an arse kicking contest.

Makes no nevermind to me since I don't bother with rankings. It's just too high school.

TammyMC: Please accept our award for the best written gobbeldegook on the net............

Available on Amazon for the Kindle E-Reader

All the Woes of a World by Jonathan Icknield aka The Bryster

And in my final hours - I would cling rather to the tattooed hand of kindness - than the unblemished hand of hate...

:)

I have no idea how this works from your explanation, Tammy. I have 6 rankings on an image; on the image page, it says Average Ranking is "Excellent". On thumbnail page, Ranking is "Interesting". Image next to mine has 3 rankings. On thumbnail page, Ranking is "Excellent". ?????

Sgiathalaich

Maybe not a very good system : if an image has a very bad note from the beginning (from trolls let's say) it will not be possible to modify the note by further votes ! Extreme case I know ... Wait and see how it works Yves

Finding some bugs and working through them. Thanks for letting us know about them. tammy

Hm! I never thought Standard Deviation would be used for Rankings! Even though I think it's not a full proof method, but I would love too see trolls out of here. Would have been great if you (tammymc) guys take off those words and put just plain simple numbers. You know, 3 or 4 out of 10 is much easy to accept then "Hmmmmmmm", "almost there" or "Interesting"! I wonder, why would you use words for ranking, when you are doing the same thing in comments!!!

Yves - I don't understand your point. In your scenario, by the time the minimum 5 rankings have been received, assuming the first was a very low one from a troll and the remainder were all nearer the high end, then the troll's marking would become an outlier beyond the limits of the bell curve and would be omitted from the overall ranking. This would occur irrespective of when the troll posted, first, last or somwehere between.

Of course it isn't perfect. As has been pointed out, "block voting" from a group of trolls or more likely a troll with "multiple personalities" could still sway the ranking - in the extreme case more strongly than before, as a mass attack could cause the fair votes to actually be discounted as lying outside the accepted region thus having absolutely no impact on the overall result! Hopefully though such cases would be glaringly obvious and would be reported for suitable exclusion action against the individual(s) concerned.

I do agree with Nish's point though. The lowest few terms are so pejorative that I think few other than trolls would consider using them, preferring not to rank in such cases. Obviously since you are computing standard deviations, you are converting these verbal rankings into numerical equivalents. My guess is that these are in a simple arithmetic progression by unit decrement. If so this could be construed as a bit misleading, as the verbal representations don't seem to me to form an especially linear progression between the extremes, and in any case are open to individual interpretation, a particular problem in a multi-lingual community such as this. I would have thought simple numeric marks out of ten would be easier to grasp. I am still of the opinion that removing anonymity from the coting would be an effective deterrent to trolls.

Sorry - last sentence "coting" was a mistype for "voting", but in fact I of course meant ranking rather than votes anyway.

For those who are un-aware, the ranking system WAS a 1-10 numerical system, but it was recently changed to the verbal rankings. I personally don't care for the current ranking system for the exact reason that Flycatcher stated. I might love an image, but it might still "need work" but I rarely rank images anyway.

flycatcher, I agree my scenario is not realistic (lot of trolls voting against a picture from the beginnig). In fact I don't really believe it could happen in real ;=) A most serious point could be to completely change the system, just considering the number of votes (it should not be possible to give a note between 1-10, 1 vote = 1). Then divide this number of vote by the number of visitors, thats a counter of visitor appreciation of a pic. That should be very close to my opinion to the n,umber of comments divided by the number of visits. My 2 cent, Yves

I don't think you can figure in the number of visitors vs the number of rankings. Most rankings and comments come only in the first 24-48 hours or so. But visitors (viewings) can come anytime.

JR221 - thanks for that. I haven't been around here long enough to have encountered the old numeric system, nor indeed the time when I believe rankings were not anonymous. I wonder why these two changes were made - I'm sure it was with good intentions, but they both seem, at least with the benefit of hindsight, to have been for the worse.

It's a cool way of doing things, my only comment (doing a lot of work with stats) is that the nominal outliers are normally removed with a 6 sigma limit, rather then a 4 sigma limit. However the tighter (4 sigma) limit chosen will be more effective at removing trolling. I'd love to see the rankings no longer be anonymous, but how much extra data would that load on to the servers??

----------

Toolset: Blender, GIMP, Indigo Render, LuxRender, TopMod, Knotplot, Ivy Gen, Plant Studio.

"Obviously doesn't dent the popularity contest aspect since they post as a group" Do not the yahoo group junkies and such have to create accounts before they can vote? Why not impose a one week to month wait period before a newbie can vote on an image (but still view, post pix, etc etc.) This way, even if all the junkies got new accounts anyway, most would forget about voting later on, and the image would slip into oblivion. Of course this would not stop those who already exist,but it's a start to keep new problems from cropping up, yes? /P