Currently, if 2 people get the same score, the one who received the most votes gets placed higher.
I see this unsuitable in a couple of ways: 1) Most people vote on every photo 2) People who dont vote on every photo, will probably simply vote on the first half that appear on their "voting session"
I propose it would be a good idea for the "number of 10s received" to be the winning criteria, and if they are equal, 9s, and so on. I think this would be a fairer way to work the problem.
Originally posted by konador: Currently, if 2 people get the same score, the one who received the most votes gets placed higher.
I see this unsuitable in a couple of ways: 1) Most people vote on every photo 2) People who dont vote on every photo, will probably simply vote on the first half that appear on their "voting session"
Can you point me to an example?
My understanding what that it used the full score (beyond just the 3 digits displayed). I don't know what happens if there's a true tie, though.
-Terry I propose it would be a good idea for the "number of 10s received" to be the winning criteria, and if they are equal, 9s, and so on. I think this would be a fairer way to work the problem.
It was my understanding that the pictures appear in a different order for everyone. So mine may be in the top half of one, but the bottom half of another.
Originally posted by jakking: With an ordinal and three decimal places, three hundred voters and 200 submissions, how often can a tie possibly occur?
Statisically highly unlikely. I am, however, VERY suprised to hear that only 3 decimal places are being used. I had always assumed that if there was a tie, it would go down to an unformated floating point.
Originally posted by jakking: With an ordinal and three decimal places, three hundred voters and 200 submissions, how often can a tie possibly occur?
Surprisingly often, as I just went through the standings and found seven this week.
Of those seven, two were true ties, meaning that the sum of the scores divided by the number of votes yielded EXACTLY the same number (places 159-160 and 169-170). I'll have to ask what the logic is on this, but it appears to me that in both cases, the photo that was submitted earlier won out.
The remaining 5 "ties" (44-45, 61-62, 72-73, 139-140, 191-192) appear to be ties when rounded to three decimal places, but when calculated fully are not in fact ties. In all cases, the photograph that scored higher won out, regardless of the number of votes.
Unless we are talking about the top five photos I really can't see it makes much difference...if I move from 127th to 125th due to a scoring method revision I will probably not notice...I myself don't consider moving up 3/1,000 to be statistically significant -- I rather get a couple more comments.
Originally posted by inspzil: You mean to tell me I tied for 159, WOW you just made my day!!!! My whole outlook for the next challenge has been changed.
Okay, don't get me wrong, I'm not Complaining but if this were to happen at 10/11, or 3/4, It would make a difference. I think it would be unfair to give one photo a ribbon and not another simple because one submitted first, or got more votes all together.
Originally posted by konador: Okay, don't get me wrong, I'm not Complaining but if this were to happen at 10/11, or 3/4, It would make a difference. I think it would be unfair to give one photo a ribbon and not another simple because one submitted first, or got more votes all together.
Well, something's gotta be the tiebreaker. All things being equal I think earliest submission is probably tbe best option (this coming from someone who has submitted at 11:58PM Sunday).
-Terry
* This message has been edited by the author on 9/26/2002 7:12:56 AM.
Well, something's gotta be the tiebreaker. All things being equal I think earliest submission is probably tbe best option
Why would that be? I think submission time is as much a function of a person's ordinary work/school schedule as it is a function of their level of perfectionism and a function of their level of procrastination.
Agreed. It would not be right to penalize on the basis of time submitted. What if the person has a shot and mulls it over for a couple of days? Does this mean they shouldn't be sure of their submission and just hurry up and get it in?
I also agree with the tie breaker being # of 10's etc, although I'm not a statistician, so maybe the admins' original rule has something to do with the way votes are processed just prior to the change over...
I also agree with the tie breaker being # of 10's etc, although I'm not a statistician, so maybe the admins' original rule has something to do with the way votes are processed just prior to the change over...
Myself being a database/web developer, I think more than likely the current "first submission" original rule is actually because that's the way most databases order their data by default. Hence, if you sort the results by just score first and there is a tie in the score the database will return the earlier entry first by default.
In any case I think we'll need some clarification about tie-breakers. I could see a lot of hard feelings should there ever be a tie in the top 5 and there aren't any contigencies put into place.
In any case I think we'll need some clarification about tie-breakers. I could see a lot of hard feelings should there ever be a tie in the top 5 and there aren't any contigencies put into place.
I agree. Either that or we could end up like Major League Baseball did during this years All-Star Game with no contingencies for a tie and just call the whole thing off!! LOL
