Classic Tease of the Month: November 2011

[Alliteration]

Including a personal nomination from each site member instead of going by ratings is a good idea.

I agree. One issue I can see with it is that we might end up with a lot of nominations, which doesn't lend itself well to a single-round system. Maybe we should require two or three nominations for a tease to be selected? I also think we shouldn't be allowed to nominate our own teases. Another issue is time - do we give a full week for nominations, then another full week for votes? Or do we do 4 days/3 days?

My only gripe with my own suggestion is it only solves the nomination process - it doesn't preclude the possibility of the actual TOTM vote from being manipulated.

I agree with this as well. Maybe we could combine the idea with voting via a post, instead of a poll, and have an admin check IPs of the posters for duplicates?

------------

Here's an idea of my own - what if we lowered the threshold for a tease to be nominated to 3.0, or 3.5, and included every tease over this threshold in a tiered/bracketed tournament? Or maybe, the highest-rated 8 or 16 teases? While this would actually make it easier for a tease to be artifically voted up to be nominated, it would at the same time make it harder for a tease to be artifically voted down to not be nominated. After that, I would think that voting via posts with IP checks would take care of any tease that doesn't belong there.

If something like this, or the above idea, could be implemented, I'd be more than happy to check the IPs myself at the end of each voting period (assuming anyone trusts me with that power ).

[Human]

How would you feel if your vote was thrown out because of some arbitrary criteria?

Depends on what the gain is. In the example I gave, group X would be unaware if their vote was taken or not in the first scenario.
In the second case, they are making positive contribution by accepting that their votes do not count.
I would be fine with my vote not being counted if there is some benifit.

Also, assume that the voting distribution of X is the same as that of Y.

This seems like a dangerous thing to assume. Even if everyone is honest, the distribution is most likely not even, for many reasons.

Correct. This hypothesis needs to be tested.

So, in case everyone is honest, we would lose nothing by restricting the votes to just Y.
X is making no positive contribution. so no harm done

Not necessarily, as "swing votes" come into play here. Say the TOTM is nearing the end of the voting period, with two teases very close - tease 1 has 10 votes, while tease 2 has 9 votes (measured only by Y votes). Meanwhile, a member of Y is debating which tease to vote for. He comes to the conclusion that to him, both teases are equally good, and so decides to vote on 2, putting 1 and 2 into a tiebreaker round.

Now add in all the votes from X. While the ratio of votes for 1 and 2 will remain the same (10:9), if we double the number of votes, we get 20 votes to 18 votes, with the same 10:9 ratio. Now, Y's vote no longer results in a tiebreaker.

Correct. But such events are rare.

The more (honest) votes you have, the more accurate the results will be. Larger sample sizes are more accurate.

And the more dishonest votes you have, the less accurate the results. The benifit has to be weighed against the costs. If we can eliminate 90% of the dishonest votes costing us 20% of the honest votes, it might be worthwhile.

[Alliteration]

Depends on what the gain is. In the example I gave, group X would be unaware if their vote was taken or not in the first scenario.

No they wouldn't, not unless you kept the "must have 50 posts" criterion a secret from everyone. And that's just shady. The system needs to be open, we need to know how it works.

On the other hand, if this criterion was known, we would probably see a massive flood of spam posts from people who just want to vote. That's another reason why this idea just won't work.

Correct. This hypothesis needs to be tested.

I don't think we even need to do that; we can probably infer a disparity based on likely personality traits of those who post often vs. those who don't (such as extrovertedness, for one example).

And the more dishonest votes you have, the less accurate the results. The benifit has to be weighed against the costs. If we can eliminate 90% of the dishonest votes costing us 20% of the honest votes, it might be worthwhile.

I strongly disagree here - I think any solution that costs us even one honest vote is too high a cost. A utilitarian approach is not the answer, especially since solutions have already been proposed that don't cost any honest votes.

[Wyatt]

I am basically a lurker as I don't like to post and gripe if I don't at least have an idea for a solution.

Hiding the poll results until the poll is over has already been suggested and endorsed by a few. Apparently we aren't the only ones that have cheating problems and some other forums have adopted this change. I don't think this is a cure all, only a step in the right direction.

I did a little research and I think I have found a way that SeraphOx could make the necessary changes with some fairly simple hardcoding. I'm not sure this would work for us since I don't know what version of PHBB we are using. This solution was for PHBB v 3.0... I think it or something similar would work for us.

I would appreciate it if 1885 would take a look at this and if it sounds good along with the idea in general forward it to SeraphOx. Maybe Indy could look also since he is somewhat familiar with the guts of the forums.


http://www.phpbb.com/community/viewtopi ... #p12892663


Bill

[Evals]

I believe in the self-culmination of an enigmatic process that reverberates the institutionalization that has haphazardly and inevitably occupied the innumerable and yet still quantifiable squalor that has disarmed this community. The juxtaposition of a unanimous methodology to vote while maintaining anonymity has grossly impaired the ability of superiors to make worthwhile changes to an otherwise ill-fated system. Only a methodology incorporating constant vigilance has the uniform potential to bring reformation to a system otherwise condemned to degenerate to an even more artificial state of non-conformity to this community's humble standards. Clearly, and rather unfortunately, this has become the current situation.

[les]

                              As Succinct as Evals.

[cumhardy]

So, back on topic - any other thoughts from anyone?

I think this thread has become stupid as shit. LOL

[cactusman]

What kind of cheating?

Someone in England is changing their IP and mass down voting the teases while not logged in.

I have seen what appears to be a pattern of down-voting of my recent teases, for what it's worth.
By looking at the RSS feed for Milovana, I saw see that the first vote on one of my teases was a 1 (and then there were 2 5's, it seems). And there seems to be a steady pattern of votes that are depressing the rating on that tease (Topping from the Bottom). I keep seeing the rating go up to 4.4 or so, and then it gets beaten down.

Is the same kind of cheating going on here this time?

[les]

As far as I understand things,
A tease gets a 1 vote when published to enable the rating filter to work after the tenth vote.

[cactusman]

As far as I understand things,
A tease gets a 1 vote when published to enable the rating filter to work after the tenth vote.

Interesting. I don't see why that should be needed, Since the code could equivalently check after 9 ratings instead. I also don't see why a 1 rating is the right way to do it, as that seems to unfairly advantage older teases that were published before that practice started.

[les]

As far as I understand things,
A tease gets a 1 vote when published to enable the rating filter to work after the tenth vote.

Interesting. I don't see why that should be needed, Since the code could equivalently check after 9 ratings instead. I also don't see why a 1 rating is the right way to do it, as that seems to unfairly advantage older teases that were published before that practice started.

I am only relaying what seraph0x posted a short while ago when the whinging about "ryan at Hotmail" teases were at their height

[seraph0x]

Interesting. I don't see why that should be needed, Since the code could equivalently check after 9 ratings instead. I also don't see why a 1 rating is the right way to do it, as that seems to unfairly advantage older teases that were published before that practice started.

The reason I used that little trick is because you cannot divide by zero. Adding 1 to the vote count (which is actually not the same as a 1 rating, but as a 0 rating) allows me to avoid a conditional.

The practice started in 2007 when I added the rating feature. But it's applied at runtime, so it always affects either all or no teases.

A few days ago I added rating precalculation - the precalculated ratings don't use the trick anymore, but so far they are only used for sorting and for the hard rating filter. I.e. the displayed rating still has that extra 0 vote while the rating used for sorting doesn't.

I did a little research and I think I have found a way that SeraphOx could make the necessary changes with some fairly simple hardcoding. I'm not sure this would work for us since I don't know what version of PHBB we are using. This solution was for PHBB v 3.0... I think it or something similar would work for us.

I would appreciate it if 1885 would take a look at this and if it sounds good along with the idea in general forward it to SeraphOx. Maybe Indy could look also since he is somewhat familiar with the guts of the forums.

Good idea!

Unfortunately editing code on the forum is a bit problematic. It's not a good idea to just copy code from a random thread, because there is nobody who will update that code for you in case phpBB gets updated. I looked for a mod that does the same thing, but the only one I found seems to have been abandoned by the author.

So, this may be an option at some point, but not right now where I'm already overloaded with the current technical issues due to the traffic.

@Wyatt: I added the permissions for starting threads and posting polls to your account. You can click quote on an existing TOTM post to get the code, then just drop in the correct candidates. If you need help feel free to ask on the forum or PM me directly.