DPChallenge: A Digital Photography Contest You are not logged in. (log in or register) 
 

DPChallenge Forums >> Challenge Results >> Is DPC being hacked?
Pages:  
Showing posts 126 - 144 of 144, (reverse)
AuthorThread
 05/05/2010 07:17:41 PM · #126
A 5 point voting system would be better in many respects.. But change is always resisted.. :)
 05/05/2010 08:14:56 PM · #127
Originally posted by coryboehne:
A 5 point voting system would be better in many respects.. But change is always resisted.. :)


Reducing the resolution of the voting system will solve nothing.
 05/05/2010 08:23:22 PM · #128
Originally posted by TrollMan:

Odds of getting 5 1's in total and all of those in a row when getting 236 votes is 1 to 732082482176. That is 732 billion. It has happened to me also on more than one occasion on high(er) scoring images that I get almost all the low votes in a row. For example 3 3's in a row and only 3 3's in total after voting is over. Odds for this to happen with for example 150 votes in total is 1:3.3 million.

And this is exactly why I at times have been questioning the voting system.


The odds of it happening aren't relevant, because voting is not a random occurrence - it is a decision made by a person.

The ones could be the result of scripts written to automatically downvote submissions in a contest. The script would run as soon as voting opened unless it was specifically written to do otherwise.
 05/05/2010 08:24:43 PM · #129
I might have skipped a few posts (so if I missed a comment similar to this I apologize)... and this is likely not a popular opinion... but I'm thinking mandatory paid memberships should be required. Instead of free accounts, we can have "basic" accounts. $5-10 or something cheap like that. Just enough to where nearly everyone can afford to "play" and the trouble of paying might discourage some people who come on here with ulterior motives (not related to improving their photography skills).

For those participants on here that are really in a financial bind (and, Lord knows I've been there) a "free" membership can potentially be offered on a case by case basis. Again, it would require an effort that someone not here for "good" might not waste time on.
 05/05/2010 08:34:37 PM · #130
Mandatory paid memberships would not get rid of people who are willing to do anything for a virtual ribbon.
 05/05/2010 08:57:18 PM · #131

Has anyone thought about those among us who delight in the reaction(s) of the early posters "I got a 1! troll!" HA!! I got a 3, yeah, the trolls like me better! Ah, I do oversimplify here. but do look at the opening of all the Challenge discussion threads. They are ALL the same. Low scores. It's the TROLLs again! to the trolls it's instant gratification.

Because we are a complete microcosm of people, photographic people, but still...there are some among us who want to get recognized, one way or another even if it is only at their own computer.

Sick? Heck, I'm not gonna go there.

Adding restrictions to voting, posting, entering. Nope, not gonna do it. A strong sense of humor might help.

Additional insight from an almost retired "Tilter at windmills"? Sure I voted for several challenges before I got the courage to enter one. Wanna dis those of us who just want to learn how to do it?

Of course not.

Okay, carry on.
05/06/2010 12:41:27 AM · #132
Again...

My intent with the post was only to offer a method for weeding out ghost accounts - but only after you would prove the fact by looking for patterns based on consecutive votes which otherwise ONLY would be possible by purposely creating at least 3 ghost accounts which according to the DPC rules are illegal. That's all.

But as the SC has clearly pointed out, there is nothing we can suggest that hasn't already been tried or that you already aren't doing so it's only insulting to them if we try.

The ghost accounts is really not a huge issue to me. I'm just a fairly logical person that react when I see things that aren't possible unless site rules are broken. But the attitude of the SC in this thread has been disappointing and certainly changed how I feel about the site and regretfully downgraded the image the SC used to have.
 05/06/2010 01:02:55 AM · #133
The statistics and psychology of the matter fascinate me:

The goal of the voting system seems to be to rank the images in a challenge according to the general consensus of the voting population in a fair manner. To that end it allows each user 1 vote on a scale from 1-10 and performs a simple average of those votes for a final score. It seems fair, but is it?

By offering users a 1-10 scale it allows images to "earn" all 10 points from a fair voter, and said voter has the ability to distinguish between many levels of quality. Which is great. However, for an unfair, divisive voter (whether high or low-baller), I believe the system actually incentives extreme voting.

It's simple math - a vote of 1/2 or 9/10 affects the outcome (simple average) more than a vote of 4-6. That means that voters who intend to unfairly sway the system can use the extremes and will have more effect on a given image's final score than people who vote in the middle. That seems wrong to me.

So how do you make the resultant score more closely match the voting consensus and less represent individual votes without completely changing the voting system?

A couple statistics ideas that might accomplish this, concerning the mode (the vote that occurred most often for a given image) and outliers (high and low votes for a given image):

- Weigh the mode into the final score of an image. (i.e. final score = 80% simple average, 20% mode)
- Use standard deviation to identify and remove outliers on a given image.
- Discard outliers by some simple rules, like discard the highest N votes and the lowest N votes on a given image (where N is some percentage of total votes for that image).

These may not seem "fair" at first glance, but IMO, in a system where voters have varying philosophies, voting styles, and motives, a simple average on a 1-10 scale gives the individual vote too much pull.

Just for kicks, I may go back and recalculate results given some of the above "consensus emphasizing" filters. One might also predict that unfair voting rose for the DPL, and could hypothesize that the filters would affect results (root-mean-square change in placements) more during DPL weeks than non-DPL weeks.

Fascinating stuff, eh?
 05/06/2010 01:15:18 AM · #134
On the contrary, one could argue that pushing the voting system more toward a consensus result would reduce individuality and unique perspectives. Of course, the very nature of having images ranked by the general public doesn't seem terribly conducive to unique perspectives.
 05/06/2010 01:46:53 AM · #135
Originally posted by smurfguy:

- Discard outliers by some simple rules, like discard the highest N votes and the lowest N votes on a given image (where N is some percentage of total votes for that image).

Fascinating stuff, eh?


I could have sworn I said something like this....

+1!!!!!

BTW Smurfguy. I completely understand and agree with your arguments....I sincerely believe that an RMS rating of photographs would better reflect the ACTUAL view of the voters than just a raw average.....

Message edited by author 2010-05-06 01:58:24.
 05/06/2010 02:11:19 AM · #136
I implemented one of smurfguy's suggestions in the form of an excel formula (new score = 0.8*simpleAvg + 0.20*mostFrequentVote). Here is a snapshot of how that will change the rankings of a recently ended challenge (double exposure) if implemented today. Intuitively, I tend to agree with the new rankings, using the new formula, but I leave it up to you guys to play with:-) (I can share the excel sheet with formula with SC/ Langdon if needed for verification. All you need to do is input published stats from a finished challenge and it will compute the new score).

Click on the thumbnail below to see the new (hypothetical) rankings. Note that I didn't implement any vote scrubbing, only the new weighing scheme to weigh in the most frequent vote (80/20 rule):



ETA: Top two ranks stayed the same for this particular challenge. However, there were three pictures that jumped up the ladder:

Rank 5 went to rank 3.
Rank 6 went to rank 5.
Rank 10 went to rank 6.

Message edited by author 2010-05-06 02:13:27.
 05/06/2010 02:14:22 AM · #137
Originally posted by keyz:
Originally posted by Dirt_Diver:
I don't know if L can block IP's like admins can on other sites but it would be a good idea.

It should be quite easy to block IP's, but 1 IP is not equal to 1 user. You can find 10 user behind 1 IP. Then, to block 1 you penalize 9 others.
Also, people who wan't to cheat can find as much IP as they want, just use web proxy, the onion router or stuff like that.
I really think blocking IPs is not an option.


And some people spoof IP's to act like they are in another country to be able to watch some online tv.


05/06/2010 02:32:55 AM · #138
The idea of consensus filters is repugnant. Has this been forgotten already?
 05/06/2010 06:16:36 AM · #139
Originally posted by Jac:
Why are SC members against Ken's great idea?

Man, I get no respect around here! I'm new, but still. No one's read the post (17) that directly precedes Ken's great idea? He just said it with more authority.
 05/06/2010 07:15:55 AM · #140
Originally posted by cryan:
I personally don't vote below a 4, and that is being harsh for me. I use 5 as my base and go from there. Maybe instead of a 10 point scoring system a 5 point system would work better at eliminating lower scores. Just a thought.


... and such is your right to do so.

I on the other hand have doled out the occasional one and use the entire spectrum of the scoring scale.

I truly cannot agree with your last comment about using a five point system is is has no added value...the only thing we would be changing is the number people will whine, rant and bitch about. Another thing to consider is that when you step either down or up one point on a ten point scale, the ensuing change is far less than it would be in a scoring pattern involving only 5 scores.

Just another person's view.

Ray
 05/06/2010 07:16:22 AM · #141
Originally posted by skewsme:
The idea of consensus filters is repugnant. Has this been forgotten already?


Gotta agree with this view.

Ray
 05/06/2010 07:20:34 AM · #142
Originally posted by coryboehne:
A 5 point voting system would be better in many respects.. But change is always resisted.. :)


Making a generalized comment is one thing, having the ability to provide empirical evidence to support it is quite another thing.

Show me just how this would work and I might be swayed.

Ray
 05/06/2010 07:23:44 AM · #143
Originally posted by clive_patric_nolan:
Originally posted by David Ey:
I saw no difference in combining the head of a little girl with a dog and the head of a monkey with Obama.


Oh, dear god, David. I hope that was tongue in cheek and not quite as offensively idiotic as it comes across.


I have never been known to come to the defence of David Ey, but to be fair he did have that little comment you neglected to add, namely: "Some things just aren't right " which does change the gist of the comment somewhat.

Ray

Message edited by author 2010-05-06 07:24:43.
 05/06/2010 09:54:25 AM · #144
If we all just vote on a 10 through 10 scale, all of these nasty little problems can be avoided.

In all seriousness, Here is how I see it... I don't think trolls meanfully influence the outcome of a challenge, and if they do then their targeting is random. As much as people would like to don their alumninum foil hats and believe that there is an international conspiracy of trolls targeting them and making their lives miserable, that is not the case, trolls are a bunch of rabble rousers and all this complaining on spurs them on. Life is unfair.

In general though, voting, including the trolling, is a pretty efficient process of sorting the entries. Individual tastes will vary, but for the most part, images that photographers thought out and executed well will rise to the top and images that need a little more work will sink to the bottom.

I think the value of this site lies in that there are so many different personalities and approaches. It is enlighting to go the a lot of/all images in a given challenge, it gives (to me at least) a bunch of different ways to tackle the same problem I faced, it often gives me inspiration and ideas for future challenges and my non-DPC work.

If you look at my profile, you see that I have a fairly ridiculous total number of images viewed, but I feel it gives me a pretty good context for my work when I submit it to challenges, and good foundation for my photography in general.

I hope I didn't ramble to far...
Pages:  
Current Server Time: 08/22/2025 06:31:37 AM

Please log in or register to post to the forums.


Home - Challenges - Community - League - Photos - Cameras - Lenses - Learn - Help - Terms of Use - Privacy - Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 08/22/2025 06:31:37 AM EDT.