Squash the trolls! Rant/Site suggestion

01/08/2003 11:30:17 AM · #1

Something has to be done about the troll voters here. It's completely ridiculous that high quality images are getting 3's and below. I, personally, do not feel that those votes are an accurate reflection of how the world views our photos (except in extreme cases). The trolls are significantly skewing the rankings and should not be allowed to degrade the quality of our votes.

My suggestions to remedy this problem:

1. Select a number of voters in each voting week (say 20) who have voted a 3 or below. Have the moderators review the votes and corresponding photos and decide whether or not those voters were voting in good conscience. If not, allow the voter to continue voting...but have their votes not count at all in the average scores. It would also be nice to have a weekly announcement of how many trolls have been erradicated...so we know progress is being made.

2. Change the voting scale...from 1-10 to 1-7 possibly. This could be a totally wrong idea as I am horrible at math. But I'm thinking that in the big picture...this would reduce the impact of troll low votes as the disparity between possible high scores and low scores would be smaller.

Any other ideas?

bamaster

Canon EOS-D30

01/08/2003 11:41:17 AM · #2

Or how about just knock off the lowest five and highest five votes? Kinda like a curve.

jimmythefish

Canon EF 17-40mm f/4.0L USM

01/08/2003 11:42:12 AM · #3

I really, really like the system used by photo.net where a comment must be made if a rating of 1,2 or 7 (1-7 scale) is chosen for the photo. The person's name is displayed by both the rating and the comment. I think something similar should be put into place here.

FranziskaLang

Nikon D100

Tamron SP AF 28-75mm f/2.8 XR Di for Nikon

01/08/2003 11:45:28 AM · #4

i don't consider myself a troll voter. but i do use the whole range. isn't that why we have a whole range? i mostly leave a comment on my low votes, but that's personal choice. forcing people to comment won't make a difference because i can vote a 1 and say "good" in my comment. or "sucks" or anything like that.

overall, these votes are few and far between and skew the results for everyone alike. i think just live with it.

oh, and if people vote 1s and 2s on more than 50% of the pix the vote is already discounted.

Sigma 18-125mm f/3.5-5.6 DC for Canon

01/08/2003 11:45:52 AM · #5

The simplest way to rid the site of troll voters is to make the vote you give to each photo public and visible to the photographer...

Kavey

Canon EOS-400D Rebel XTi

01/08/2003 12:02:39 PM · #6

I've voted 5 3s and 10 4s in the current Song Titles challenge but no 1s or 2s.

But I do use them and do feel I can distinguish enough levels for at least 9 voting scores...

I know that people see photos very differently.. but I do wonder about some of the very low scores given to what are pretty good photos.

I know some votes are excluded because of voters not achieving the 20% target or because their votes are thrown out for other reasons - question is do these votes still show on the voting graphs next to each image after results are announced??

Kavey

Patella

Canon EOS-500D Rebel T1i

01/08/2003 12:21:14 PM · #7

I use the 1-3 range as well as the 7-10 range. And while I like to think I can vote "responsibly" I know I've seen photos I didn't care for at all in the top 20 and photos I really dug score in the last 20.

Any time you force people to justify their vote, or even, in my opinion, make it so that it's possible to identify their vote, you're simply going to get vote inflation.

Here's the way I'd think: "If I vote this photo a 1, I'll have to justify myself -- so screw it, I'll give it a 3 and avoid the whole issue."

Then we'll have people complaining about the number of 3s being handed out, and then the number of 4s, and then....

Froober

01/08/2003 12:48:06 PM · #8

Okay, chat's been going bonkers and I've gotten another idea (from Gordon). Maybe the biggest issue is that people aren't voting on the same scale. Could the 1-10 scale be too broad?

For example, you take a bite a pie (mmm...pie)...and someone wants to know how you would rate it. You would most likely say...it's terrible...it's okay...it's good...it's very good...or...it's delicious-out of this world. That's how I operate everyday...I don't know that I can accurately use a scale so broad as 1-10. Using the pie example and a scale of 1-10...I would have to respond based on:

1. This is toxic waste...call hazmat asap!
2. OMG! I need a to-go box...my mother-in-law has to try this!
3. Unable to respond...running to bathroom retching.
4. Look up from rubbing pie into shoe...this is great polish!
5. It's okay...slip some bites to Fido.
6. Pretty good...just leave the crust on your plate.
7. Yummy...if only it was ala mode!
8. I need this recipe!!
9. This is awesome...mmmm...flick a tiny speck of burnt crust off.
10. Break into the 'When Harry Met Sally' diner routine.

I don't think I have the time (or inclination) to think that much about each and every photo...my apologies. I usually vote on a scale of 5-10..a few 4's...and even fewer 1's. So...maybe a 1-5 scale would be better...I dunno...

Thoughts anyone?

Gordon

RIM BlackBerry Curve

Canon EF 85mm f/1.8 USM

01/08/2003 02:35:09 PM · #9

Originally posted by Froober:

I think if that works for you, then that is great and you should continue to vote that way. I also think that the other scales other people use are also okay. It all pretty much balances out in the end anyway with 'good' (for various values of good) pictures coming to the fore.

The problem isn't that 1-10 is too granular, it is that really people vote on about 6 or 7 different things, and give varying weightings accordingly - to be scrupliously fair we should probably give each picture a score in half a dozen categories - not that I think that this is a realistic approach.

E.g., there is a 'does it meet the challenge' scale of voting, there is also a 'is it technically good' scale of voting and there is a 'is this a powerful image' scale of voting, and a 'is this well composed' scale of voting and a 'do I personally like this subject/ arrangement' scale of voting and a 'I'm having a bad day' weighting factor

The 1 - 10 result is an aggregate of all these and it is no surprise that everyone gives different weight or consideration to each of these factors, based on personal bias, experience level and taste. Any attempt to 'impose' a more meaningful scale than 1 bad to 10 good isn't going to improve things in any significant way so I wouldn't think changing it would make a whole lot of difference. You can still tell that if you get lots of low votes that you got lots of low votes - regardless of what you change 'low' to be

jimmythefish

Canon EF 17-40mm f/4.0L USM

01/08/2003 02:42:04 PM · #10

Not to push praise on photo.net too heartily (as there are things I don't like about their system, too) but the idea of giving a separate rating on aesthetics and originality is also a good one. It gives the opportunity to trash or reward the technical details while still giving good ups to the effectiveness...

I think a good thing to do here at DPChallenge would be to award two scores - one for technique, one for expression, and combine them for the total score. This way, people get more feedback and voting is more clearly defined. I don't think 2 values would be any harder. In fact, I think it would make things easier on most people who are conflicted about images with great impact yet lower technical merit, and vice versa.

Message edited by author 2003-01-08 14:44:43.

lisae

Olympus C-720UZ

01/08/2003 04:24:26 PM · #11

Hehehe... Nothing ever changes around here :). Even after the site upgrade!

Solutions that have been offered before:

- Grades for different qualities (at the very least "technical" and "artistic merit" type division, with "meets the challenge" as another possibility).

- Different scales than 1-10

- Ranking the photos instead of scoring them

- Voting more like "good", "neutral", "bad" and then ranking on the number of "good" votes, qualified by "neutrals", then qualified by "bads" (like the olympic medal system).

- Being able to tag a photo with little labels like "most creative", "most weird", etc. (yearbook style but not really) so people have more information than just the score and comments.

- De-anonymising the votes

Or then there's the idea that we should all just live with it. The really bad low voters are cleared out with a script anyway.

bamaster

Canon EOS-D30

01/08/2003 04:31:49 PM · #12

Originally posted by lisae:

The really bad low voters are cleared out with a script anyway.

I didn't know this... can you 'splain?

karmat

Apple iPhone 7

Canon EF 70-200mm f/4.0L USM

01/08/2003 04:33:48 PM · #13

Or another suggestion -- we ALL become trolls, then it becomes boring, and it doesn't show up so badly! :-)

Tony, if someone votes with an "odd" pattern (like all ones, or something like that) their votes are cast into the sea of fire and brimstone, never to be counted in that challenge.

GeneralE

Canon PowerShot S3 IS

Canon 12x IS USM (S3 IS built-in Lens) -- 36-432mm (35mm EQ)

01/08/2003 04:59:20 PM · #14

I've always wanted to run another algorithm/analysis where you compare a person's vote on a picture with the average vote for that pic, and then calculate that voter's average difference from the average vote (I suppose that's a form of standard deviation, but I don't want to be accused of calling people nasty names).

People who consistently vote a certain amount lower than the site average, or have too many votes which are "too low," could be made subject to suspension or a "re-education program" like in the old days...

KarenB

Canon EOS-70D

Canon EF-S 18-135mm f/3.5-5.6 IS STM

01/08/2003 04:59:42 PM · #15

I agree a decision must be made. And I totally agree with Froober's original post that excellent photos are getting 2's and 1's. There really is no reason for this. I am not saying you have to be psyched about every photo and give 'em all 9's and 10's, but what is with the 1's on a photo that is obviously at least well done?
We can talk about it forever.. what we need to do is come up with a solution.

01/08/2003 05:03:45 PM · #16

i can't really see a solution that would make any difference in the least... what I do see is that the overall scores cast in the members challenges are better than the non member challenges... there is much less of this going on...

making your vote public would make you open to criticism for scores you give, but would it really do any good?

If I gave you a 1 and you felt it was not warranted, what will u do? send me email? I can either ignore the email or i can justify to you why I didn't like your image. Either way, you will still be unhappy :(

Canon EF 28-135mm f/3.5-5.6 IS USM

01/08/2003 05:07:43 PM · #17

If I could have any change *I* wanted on this issue, I would ask for non photographers to NOT have the ability to vote. I estimate that most of the low votes on the great photos could be coming from the friends and family of photographers that come here to vote on their buddy's images.

Proposal: You must have a photo in the challenge to be able to vote in the challenge.

Would it make a difference? Dunno...

I think it all washes out in the end. Anything we would change to get rid of the lowball voters would simply raise the bar for everyone equally, would it not?

ClubJuggle

Canon EOS-400D Rebel XTi

01/08/2003 05:09:42 PM · #18

Does anyone else think that forced-conformism is a really bad idea, especially on an artistic site? I'm not saying that trolling isn't a problem, but let's not implement a cure that's worse than the disease.

-Terry

Gordon

RIM BlackBerry Curve

Canon EF 85mm f/1.8 USM

01/08/2003 05:11:09 PM · #19

Originally posted by Karen Bryan:

Perhaps because to some people they are a 1 on their own personal scale of 'bad' to 'good' ? I'm not saying that there are not people who vote 'oddly' or maliciously but in all the cases I've seen, there are often good justifications for the low votes on certain scales.

FranziskaLang

Nikon D100

Tamron SP AF 28-75mm f/2.8 XR Di for Nikon

01/08/2003 05:12:13 PM · #20

i think you hit the nail on the head there, john :)

Originally posted by jmsetzler:

I think it all washes out in the end. Anything we would change to get rid of the lowball voters would simply raise the bar for everyone equally, would it not?

Canon EF 28-135mm f/3.5-5.6 IS USM

01/08/2003 05:12:46 PM · #21

here's an action that I would request to start with:

D&L have this script that cleans out troll voters. I would modify that script so that it generates a list of names of those who get their votes dumped. If the name appears on the list more than once, the account gets suspended for a period of time. If it appears on the list a third time, it gets suspended permanently.

ClubJuggle

Canon EOS-400D Rebel XTi

01/08/2003 05:20:01 PM · #22

Originally posted by jmsetzler:

I'm not sure there is a proper cause and effect here. The members are people who, for the most part, are serious enough about photography to pony up $25 for a year's membership. I want to make it clear that I am not saying that there are not excellent photographers among the registered (free) users, but simply that the percentage of excellent photographers among membership is higher than that among the entire site population.

-Terry

GeneralE

Canon PowerShot S3 IS

Canon 12x IS USM (S3 IS built-in Lens) -- 36-432mm (35mm EQ)

01/08/2003 05:24:42 PM · #23

Originally posted by jmsetzler:

I agree with you except the last point -- I'd be happy to understand how someone finds one of my photos so unrelated to the challenge, or so poorly composed/executed, that it has no redeeming value and is a waste of the viewer's time. I'd ask less for "justification" than just an honest explanation of the reasoning process behind it...that way I could choose whether or not to factor that information into my next shot.

As of now (35+ challenges) I can't recall if I've EVER give a 1, and don't think I'd give any of mine a 1 either (though I'd probably cast some 2 or 3 votes in there maybe).

Actually, I think the simplest change would be to require people to vote on (say) 60% rather than 20% of the images for the votes to count. I think all committed members (with or without cameras) do this most of the time anyway, and the 20% bar is too low to weed out the "friends and family" voters, and it also may not be a big enough sample for "The Alorithm" to work effectively.

How about a poll giving folks 3-4 choices of minimum vote percentage required for votes to count? Maybe 20% 50% 75% 100%

01/08/2003 05:28:55 PM · #24

Originally posted by ClubJuggle:

Originally posted by jmsetzler:
i can't really see a solution that would make any difference in the least... what I do see is that the overall scores cast in the members challenges are better than the non member challenges... there is much less of this going on...

I'm not sure there is a proper cause and effect here. The members are people who, for the most part, are serious enough about photography to pony up $25 for a year's membership. I want to make it clear that I am not saying that there are not excellent photographers among the registered (free) users, but simply that the percentage of excellent photographers among membership is higher than that among the entire site population.

-Terry

The point is not about the quality of the photography.. it's about the number of 1 and 2 votes that are given.. they are unproprotionally better than in the other challenge...

Froober