| Author | Thread |
|
|
08/16/2006 06:14:58 PM · #1 |
Ok math geeks, this is a strictly out of curiosity question, let not burn down the site or start screaming about trolls.
Right now the images are ranked by the average vote on each photo. Generally we vote on a fairly random sampling and votes are spread out fairly evenly among photos. That is, FAIRLY spread out - but not truly completely spread out. Some photos may receive 250 votes where others receive 200.
Does this create a significant discrepancy in the way the photos are ranked? For example lets say a photo received 250 votes and had an average score of 6. Then (for the sake of argument) another photo only received 1 vote of 10. The photo that received 1 vote would be ranked higher automatically because it's average score was higher. It would work the other was around too.
Is there a statistical was to sort of "level the playing field" between these 2 photos? I guess it would be some sort of weighted average or some such thing (I'm not a math geek so I don't know what it'd be called). The above examples may be too drastically different to work so what if one had 230 votes and the other 210 votes.
Or would it make sense to look at it more from a voting methods perspective and have the order of pictures to vote on automatically be in the order of lest voted on pictures? In my on little world's theory that would distribute votes more evenly among photos across a challenge.
Does any of this make sense or am I just rambling nonsensical pseudo-math mumbo jumbo right now?
edit: BTW, I'm not even suggestion we change the site, I just stuck this thread in here because I didn't know where else to put it.
Message edited by author 2006-08-16 18:19:56. |
|
|
|
08/16/2006 06:24:32 PM · #2 |
| For small samples (less than a dozen or so total votes), that kind of sampling error can lead to significant problems. For the kinds of vote counts we see in challenges, the potential errors are miniscule. Here's an example that illustrates why that's true... suppose one image gets 250 votes while another gets 300. That's a big difference for our purposes. Now imagine that there are 50 more "hypothetical" voters on the 250-vote photo. Assuming that they come from the same general population that the rest of the voters came from, their average vote will tend to be the same; some will vote higher than average, some lower. The end effect is that the extra votes won't change the average much at all. This is even more true if we ensure that the extra votes are distrubuted across the voting period somewhat evenly. |
|
|
|
08/16/2006 06:35:10 PM · #3 |
So I'm not totally crazy (that's a relief).
I understand that in general it's not going to create a huge difference in scores (assuming it's the typical pool of voters) but the challenges are often times called by the 10ths or even 100ths of a point.
With out doing the math (or being able to) my guess it that it could still re-order some of the places in challenges. |
|
|
|
08/16/2006 06:39:35 PM · #4 |
Originally posted by Megatherian:
So I'm not totally crazy (that's a relief).
I understand that in general it's not going to create a huge difference in scores (assuming it's the typical pool of voters) but the challenges are often times called by the 10ths or even 100ths of a point.
With out doing the math (or being able to) my guess it that it could still re-order some of the places in challenges. |
A few, and more likely in the average-score range where images are separated sometimes by thousandths of a point. For images separated by greater than about 0.02, it's unlikely.
|
|
|
|
08/16/2006 06:46:03 PM · #5 |
ok. Like I said, I'm not trying to stir up trouble. Someone posted a thread about the average scores on cameras today (which seem WAY whacked) and that got me to thinking about the scores in the challenges.
Thanks Kirbic :) |
|
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: 08/29/2025 07:15:25 PM EDT.
