cKleinhuis
|
|
« on: May 01, 2010, 02:06:44 PM » |
|
FractalForums.com Spring 2010 Fractal Art Competition Submission Period has EndedThe Entry Submissions time frame for the "FractalForums.com Spring 2010 Fractal Art Competition" has now ended. This submission period set a new record for Member Activity on the FractalForums, with around 650.000 Page Views!!! We would now like to enter the voting period, which will last until the end of May 2010. We have a total of 130 entries to the three different competition sections. And due to the large amount of entries, please take your time to review each of the submissions and give a reasonable vote for what you consider "best" images per category. Only the Gallery votes count for the results. So, no external embedded video Sites Rating system is taken into account. The ranking will be as follows: - highest average rating ( if tied for same position, then ) - most votes ( if tied again, then ) - sharing of winner placement Again, we ask that you please take your time and review each category of the competition, the sections are as follows: • Mandel Brot -- Only views of the standard z^2+c Mandelbrot set and Julia Sets where allowed, and this section has more than 47 spectacular entries: http://www.fractalforums.com/index.php?action=gallery;cat=38 • Fractal Fun -- Contains computer generated fractal art of any kind. This section was the most popular with more than 64 entries. It is a vast mixture of state of the art Mandelbulb renderings and beautiful 2-D creations: http://www.fractalforums.com/index.php?action=gallery;cat=37 • Fractal Movies -- For the first time in the Forums' history, there was a movies/video section, with a total of 19 entries. You will encounter interesting new locations of the Mandelbrot, plus amazing 3-D fly-bys and inside explorations. Enjoy a total of 40 Minutes of extraordinary Fractal pleasure: http://www.fractalforums.com/index.php?action=gallery;cat=36 Voting Period will end on: 1.June 2010
|
|
|
Logged
|
---
divide and conquer - iterate and rule - chaos is No random!
|
|
|
Pauldelbrot
|
|
« Reply #1 on: May 02, 2010, 02:56:12 AM » |
|
( if tied for same position, then ) - most votes Is that most votes total, or most 5-star votes? The latter might make more sense (assuming that when all is said and done all potential winners have had at least one 5-star vote, as seems likely).
|
|
|
Logged
|
|
|
|
Nahee_Enterprises
|
|
« Reply #2 on: May 02, 2010, 11:20:22 AM » |
|
Is that most votes total, or most 5-star votes? The latter might make more sense (assuming that when all is said and done all potential winners have had at least one 5-star vote, as seems likely). Well, it could be a two-tiered vote counting system, where if there is a tie for the most 5-star votes, then it would then go to counting total votes for that entry. The 5-star votes would be the "highest ranking" vote. And the total votes would be the "popular" vote.
|
|
|
Logged
|
|
|
|
Pauldelbrot
|
|
« Reply #3 on: May 02, 2010, 03:00:10 PM » |
|
Tied for rating and 5-star votes I'd suggest either calling that a tie, or looking next at 4-star votes and then calling it a tie if still the same. Not the lower numbers though; it doesn't make sense to me to pick one over another because more people rated it average, for instance.
|
|
|
Logged
|
|
|
|
Sockratease
|
|
« Reply #4 on: May 02, 2010, 03:17:08 PM » |
|
Not the lower numbers though; it doesn't make sense to me to pick one over another because more people rated it average, for instance.
Actually, I feel that the total number of all votes added up, low and high ratings, is the most important part! Even if it is "won" by a single 1 star rating. The fact that more people cared enough to react at all, and rate the image, counts far more to me than how many people rated it a 5. Even if an image has dozens of low ratings, compared to just a few high ones, it should win! I think that if there are only a few high votes, it is an Elitist Image, and lacks mass appeal. This is not a contest for the High Brows Only! The votes of Non-Mathematicians and Casual Artists who happen to dabble in Fractals count equally as those of the people who write the software I hope you don't take anything I said as hostile... I respect your opinion, I just have different views.
|
|
|
Logged
|
Life is complex - It has real and imaginary components. The All New Fractal Forums is now in Public Beta Testing! Visit FractalForums.org and check it out!
|
|
|
cKleinhuis
|
|
« Reply #5 on: May 02, 2010, 11:20:38 PM » |
|
there is much room for discussion about a voting system, it will stay as it is defined above, i really hope we do not have many tie situations ... and do not forget, it is a fun competition ... just because we have small prices is not worth discussing too much about it, but you are sure right all ....
|
|
|
Logged
|
---
divide and conquer - iterate and rule - chaos is No random!
|
|
|
Nahee_Enterprises
|
|
« Reply #6 on: May 03, 2010, 01:53:19 PM » |
|
Actually, I feel that the total number of all votes added up, low and high ratings, is the most important part! Even if it is "won" by a single 1 star rating. The fact that more people cared enough to react at all, and rate the image, counts far more to me than how many people rated it a 5. I whole heartily agree with you on this issue!!! There is much room for discussion about a voting system, it will stay as it is defined above, I really hope we do not have many tie situations ... ... There definitely will be ties based upon the current number of stars showing (averaged from the total stars). Hopefully there will not be any ties on the other two voting levels.
|
|
|
Logged
|
|
|
|
twinbee
|
|
« Reply #7 on: May 04, 2010, 09:33:10 PM » |
|
I used this technique when I tried to find the best web host ages ago. It may not be perfect, and there could well be a more mathematically accurate way of deciding. But it still works very well. The idea is that simply averaging the votes isn't the best idea, because a low number of votes will naturally have higher levels of margin-of-error. I think Sockratease partially has the right idea in thinking we shouldn't just look at the average score, but the margin of error (/wisdom of the crowds) thing is the real reason for its undesirability Also, we should still take into account the average score to at least some extent like originally proposed. The below formula takes the score into account, but also looks at the number of votes to form the final score, thus forming the reliability of the score. High or low scores, but with only a few votes, will converge towards score = 3. A higher number of votes will maintain more closely the original average voting score. The x=4 part is adjustable. 4 seems fairly sensible, though maybe others can decide on the best value. The ^ symbol means exponentiation. x=4 reliability = (numberOfVotes / x) / (1 + numberOfVotes / x) weightedScore = (score/3)^reliability * 3
Examples: Votes=1, score=4.9, WeightedScore=3.31Votes=2, score=4.9, WeightedScore=3.53Votes=5, score=4.9, WeightedScore=3.94Votes=100, score=4.9, WeightedScore=4.81Votes=1, score=3.7, WeightedScore=3.13Votes=2, score=3.7, WeightedScore=3.22Votes=5, score=3.7, WeightedScore=3.37Votes=100, score=3.7, WeightedScore=3.67As you can see, it's more valuable to have five votes averaging 3.7, than one vote of 4.9 star (but only just). You can have it the reverse by lowering the x=4 to something like x=2. A low number such as x=0.1 is the original rating system taken to its logical conclusion (i.e. only average score matters). On the other hand, a value such as x=100 will be more similar to Sockratease's idea of just counting votes. I think a middleground is fairer, hence x=4. What do others think of this? Is there a more mathematically accurate way still? It seems complex, but really the system is very elegant.
|
|
« Last Edit: May 04, 2010, 10:05:19 PM by twinbee »
|
Logged
|
|
|
|
cKleinhuis
|
|
« Reply #8 on: May 04, 2010, 11:04:04 PM » |
|
it will be a NUMBER, e.g. one 5 star vote and one 4 star vote results in: (5+4)/2=4.5, so with the current amount of votes, i do not think we get a tie
|
|
|
Logged
|
---
divide and conquer - iterate and rule - chaos is No random!
|
|
|
Nahee_Enterprises
|
|
« Reply #9 on: May 05, 2010, 03:24:55 PM » |
|
It will be a NUMBER, e.g. one 5 star vote and one 4 star vote results in: (5+4)/2=4.5, so with the current amount of votes, i do not think we get a tie. I believe the method you have chosen for calculating a winner is a variant of something Einstein stated, which is similar to "Occam's Razor": Everything should be made as simple as possible, but no simpler. or Make things as simple as possible, but not simpler. (see: http://en.wikiquote.org/wiki/Albert_Einstein )
|
|
|
Logged
|
|
|
|
|
Tglad
Fractal Molossus
Posts: 703
|
|
« Reply #11 on: May 21, 2010, 12:29:52 AM » |
|
So currently an entry with 1 five star vote will win over an entry with 99 five stars and 1 four star.
This might be better- score is sum of 'rating - 3', highest score wins.
i.e. 4 or 5 stars will rate a picture up, 2 or 1 stars will rate a picture down.
Having said that, for simplicity I probably would also go with the mean as Trifox is doing. We'll see the flaw if the winner is something with a single 5 star rating I guess.
|
|
« Last Edit: May 21, 2010, 12:40:29 AM by Tglad »
|
Logged
|
|
|
|
Sockratease
|
|
« Reply #12 on: May 21, 2010, 01:04:31 AM » |
|
So currently an entry with 1 five star vote will win over an entry with 99 five stars and 1 four star. I don't think it would work that way. Both examples would still show a 5 star rating on the board. Of all the entries which show 5 stars at the end, the totals would then be compared. If I understand correctly. Then the comparison in your example would clearly go to the one with the highest total number of "stars" all added together. I think... But in a more closely tied example - an image with 4 "5 star" ratings and 3 "4 star" ratings would lose to an image with 3 "5 star ratings", 5 "4 star" ratings, and a single one star rating. Clearly more votes were cast and the total number of "stars" higher in the second one Assuming the average is not the primary criteria... Even I am not certain how that would work - but the important thing is that we have a Judge to decide this for us (Trifox!) and so we don't really need to be overly concerned here. I think next year we should put the voting process and winning entry determination to a vote!! Either that or just select a "panel of Judges" who will declare a winner without having to justify anything to anyone. Just declare the Winners and be done with it
|
|
|
Logged
|
Life is complex - It has real and imaginary components. The All New Fractal Forums is now in Public Beta Testing! Visit FractalForums.org and check it out!
|
|
|
Tglad
Fractal Molossus
Posts: 703
|
|
« Reply #13 on: May 21, 2010, 07:30:31 AM » |
|
The system is already decided this year, but just as a brain activity, here's perhaps a way to compare systems:
If everyone had to vote on every picture then scoring would be easy, just add up the number of stars. However most people only choose to rate a few of the pictures so the problem is a question of how many stars you give to each non-vote. The 3 systems in this thread differ in how they value an un-cast rating, like so:
Type Value of not rating a picture who wins winning picture 1 0 stars highest number of stars A 2 3 stars highest score (sum rating -3) B 3 its rated average highest mean value C
picture A has 3 five stars, 4 one stars picture B has 2 five stars, 2 four stars picture C has 1 five star
To decide which of the three systems you think is best, you just need to decide which picture you think should win the contest.... A, B or C?
|
|
|
Logged
|
|
|
|
|