You need to be logged in to post and to see the uncensored versions of these forums.
I’ve noticed that in some cases, ideas that eventually end up successful in forumbuildr can get destroyed depending on what time of the week they’re submitted. For example, in AskThieves there were three submissions in the Enemy category playing off of the “I am not a crook” line of his. Two of the submissions were submitted early in the week and went through something like five days worth of voting before the other Enemy was submitted. At the end, the third entry had a much higher score than either of the two earlier entries.
From what I’ve been told the voting is done by simple totals, i.e. the highest vote count at the end wins. This means that ideas submitted earlier in the week and have to go through rigorous review are at a disadvantage to those submitted later…and those votes submitted later don’t reflect a the general opinion of the Forumwarz community.
For example, lets say a topic description was submitted early in the week and it gets 23 upvotes and 16 downvotes, for a score of 8. Another entry gets 8 upvotes and has a score of 9. The one that has the score of 9 wins, but the entry it beat was voted on 30 more times.
This is clearly biased and very unfair to people who submit early (or even during midweek), so I have a proposal. Instead of simple totals, we use weighted voting.
The weighted voting takes into consideration the number of votes, not just the final score. Take the earlier example.
The formula would be: (raw score)*(total votes) = weighted score
So for Sub A,
(8)*(39) = 312
For Sub B,
(8)*(9) = 72
As you can see with these weighted values, it’s not a simple cut-and-dry win. Even if there was a sudden voting spam at the end, with six downvotes for Sub A and six upvotes for Sub B, Sub A would still win by a total of 90-42 (Sub A: Score of 2, 45 total votes; Sub B: Score of 7, 6 total votes).
I feel this is a better solution that removing downvoting altogether, and helps to protect older entries from the influence of newer ones. It doesn’t make it impossible for newer entries to win because downvoting still exists for those older entries, it just makes it harder for scammers to win because they have to overcome a far larger number of votes submitted. And if that vote threshold is large enough, it’ll attract moderator attention.
Such a plan could have spared us of Fran is Fat, and I noticed Outside (which was leading in the 30’s last I checked this morning) was downvoted to HECK not too long ago. Organized crime has too much influence in forumbuildr right now.
|Posted On: 09/20/2009 9:59PM||View mterek's Profile | #|
I mean it’s suggested lots of times before since long ago.ChilePepino edited this message on 09/20/2009 10:07PM
|Posted On: 09/20/2009 10:06PM||View ChilePepino's Profile | #|
To my mind it’s earlier-submitted entries that have the advantage – if a sub is good enough for the voting to, on net, be positive then the longer it can be seen for the more of a lead it’ll build up over what comes later. But that only applies for normal entries, not the ones with the backing of a group of co-ordinated voters.
In the case of exploitative voting, slipping it in at the last moment and quickly all voting it up is the best strategy, but only because you’ve got that dedicated group of voters behind it.
Separate the submission period from the voting period imo. Eliminates bandwagon effects, blocks last minute rushes and ensures that everything is available for voting for the same amount of time.man-man edited this message on 09/20/2009 10:14PM
|Posted On: 09/20/2009 10:14PM||View man-man's Profile | #|
This would be true if people paid attention to the time-stamps, but taking my Nixon example that wasn’t the case because the avatars are presented to voters in a random fashion. Earlier on, because there are fewer submissions in the pool that increases the chances of a sub getting voted on and thus pushes its vote total up.
|Posted On: 09/20/2009 10:31PM||View mterek's Profile | #|
I think that if your idea is good, it should acgreat timesulate enough votes by submitting earlier in the week to give you a good advantage over the last-minutes submitters. If you do have to change it, however, I would say this is the best idea.
The only problem is the Style voting, which could be eliminated by letting the people who submit the logo’s pick the style to go with it, but idk.
|Posted On: 09/20/2009 10:31PM||View Skyman747's Profile | #|
I don’t see a problem with this.
Basically, what you’re saying is that if someone hypothetically were to get 500 upvotes and 480 downvotes, and someone else just gets 20 upvotes, that the person with 500 upvotes should win?
That’s absurd. It’s more accurate to go by percentage; if someone can get 20 or even 8 straight upvotes without a single upvote, it’s probably a lot better of a submission than one that gets nearly as many downvotes as upvotes, regardless of the total number.
I don’t agree with your argument and I think of the multiple problems with buildr, this is not one of them.
|Posted On: 09/20/2009 11:07PM||View Shii's Profile | #|
I don’t think this is a good idea. I doubt if we have people who vote are good enough to judge submissions, represent the opinion of the crowd, and vote fairly every week.
|Posted On: 09/20/2009 11:14PM||View quangntenemy's Profile | #|
Outside went from 36 votes to 7 in one hour, whatever advantage you have is going to be ****ed away if people cheat with illegal alts like that. With the weighted vote strategy, even if people use illegal alts to cheat their way toward eroding an incredible lead, they still have to overcome the mbumive numbers of votes a forum like Outside would have won.
When I last saw the current forum’s status, it was at 7 votes. It had only been up for a hour or two, not enough time to acgreat timesulate the 90+ votes that Outside must have had (since AskThieves had 64 upvotes and something like 28 downvotes). With only 7 votes total for the current forum, Outside should have won even if it was at 1 vote by the weighted method, giving it a weighted score of 127.
Though when the situation is THAT BAD the staff should take action, the problem still persists that with simple majority raw scores it is too easy to manipulate.
I had an avatar submission for AskThieves that had a score of 14 during the second day, without a single downvote. By the forum’s publication the upvotes had climbed to 28 but the raw score was 1, because I had 27 downvotes.
Even if you go by the percentage (which I disagree) it’s still a function of when an idea is submitted. Let’s say Tuesdays are designated downvote days. Anyone who submits an idea before Tuesday gets downvoted to heck regardless of how good it is, those who submit after Tuesday are subjected to normal voting patterns. It doesn’t really matter what day of the week it is when an idea is submitted, if someone submits at the wrong time it seriously impacts their chances of winning. The only difference with Sunday is, it’s right before the forum gets published.
As far as numbers go, I back that on democratic grounds. if 900 people voted on a sub, a large portion of the Forumwarz user base had their say in an item’s quality. For a sub that only had 20 votes that’s a far smaller number, if the vote tallies are the same consecutive votes and other arbitrary variables due to the timing of the submission shouldn’t be considered, only the raw score and the total votes.
|Posted On: 09/21/2009 12:41AM||View mterek's Profile | #|