| Author | Thread |
|
|
10/27/2007 07:17:48 PM · #1 |
Please don't hold it against me that in my day job I am a computer engineer. I am certain that others of you have such deep dark secrets in your lives. Let's be honest here.
I seem to remember in my classes (e.g., statistics, physics), when trying to get a representative set of data we would through out the extremes; this was called normalizing the data. This would be trashing the one highest and one lowest value attained in the experiment. This gave us better results and sometimes eliminate spurious data. Hang in there, I am getting to the point.
People competing in DPC occasionally (perhaps frequently even) complain of inappropriate voting by people with hidden agendas. It is my opinion that: if the DPC would automatically chuck the highest and lowest vote for every entry it would eliminate some of this.
"What!!!", you may say. "Throw away my only 10!? I worked too hard to have my votes thrown in the trash!!!" That may be true, but if you think about it, I believe that you will see the merit of my suggestion. It would also throw away that single evil '1' that is killing your otherwise excellent score.
This is not a cure-all for the fact that people will be mean or for cheating, but it may help. IMO.
What do you think? |
|
|
|
10/27/2007 07:20:06 PM · #2 |
| This has been kicked around numerous times in the past, and it all boils down to the it-ain't-broke, no-need-to-fix category. |
|
|
|
10/27/2007 07:31:22 PM · #3 |
Originally posted by alanfreed:
This has been kicked around numerous times in the past, and it all boils down to the it-ain't-broke, no-need-to-fix category. |
Unless the logo suddenly falls off the page in dramatic fashion the site will always be viewed as not broken. :P
|
|
|
|
10/27/2007 07:34:14 PM · #4 |
As a matter of fact, there is an automated "vote scrubber" that discards votes from voters that are blatantly violating the spirit of voting fairly. Don't ask what the formula is, even the SC does not know that.
Bottom line, I've done a bunch of statistical analysis on scoring, and even the farthest outlying votes are part of a "normal" expected statistical distribution. Further, for those images that score exceptionally high (or low) the vote distributions "bump up against the ends" of the voting scale, and so even tens (or ones in the case of low-scoring images) are not really extreme votes for these images.
|
|
|
|
10/27/2007 07:36:29 PM · #5 |
*giggles at yanko and the visual I just had*
|
|
|
|
10/27/2007 07:43:27 PM · #6 |
Originally posted by JopperTom:
Please don't hold it against me that in my day job I am a computer engineer. I am certain that others of you have such deep dark secrets in your lives. Let's be honest here.
I seem to remember in my classes (e.g., statistics, physics), when trying to get a representative set of data we would through out the extremes; this was called normalizing the data. This would be trashing the one highest and one lowest value attained in the experiment. This gave us better results and sometimes eliminate spurious data. Hang in there, I am getting to the point.
People competing in DPC occasionally (perhaps frequently even) complain of inappropriate voting by people with hidden agendas. It is my opinion that: if the DPC would automatically chuck the highest and lowest vote for every entry it would eliminate some of this.
"What!!!", you may say. "Throw away my only 10!? I worked too hard to have my votes thrown in the trash!!!" That may be true, but if you think about it, I believe that you will see the merit of my suggestion. It would also throw away that single evil '1' that is killing your otherwise excellent score.
This is not a cure-all for the fact that people will be mean or for cheating, but it may help. IMO.
What do you think? |
I thought about the voting many times and have few ideas in my head. One of which is a troll factor.
Troll factor is one voters agreement with others.
For example lets say that this voter votes on 3 images A B and C.
So the troll factor is calculated by his given score against avg score given to same image by others. lets say others give this image 7 and you give it 1. Now this difference contribute to your troll factor but it still does not make you troll.
What would make you troll however is you disagree with most of the images not 5 percent or 10 but may be 50 percent.
So calculate disagreement with all the images you voted, and see how much you agreed with average public.
Off course this method has many problems, one is it discourage you to give extreme scores like 10 and 1, because average is very likely to be around 5 or 6 for an image. Hence not implementable.
|
|
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: 08/28/2025 04:20:15 PM EDT.
