Examination of removing opinion scoring

[Kai Stromler]

but what's more to be concerned about is that selecting relative to other scores on the member's other videos will tend to promote only one video to the top list per creator

Ummm, I think I've been misunderstood. I'm referring to normalizer per voter, not per creator.

There are attacks on that, but ultimately there is only so much you can do.

Thanks for the clarification; the fairness problem that I had when misunderstanding the system is of course gone. However, normalizing per voter does make the general sort of heavy-tail sockpuppet attack that I had been thinking of before slightly easier to run and more likely to affect people besides the creator running the attack.

--K

[Tom the Fish]

Create an AMV Hall Of Fame

I am so against this idea that it makes me gnash my teeth. I feel there is already too much of an "old boy's network" flavor to this site as is.

Tom

[derobert]

However, normalizing per voter does make the general sort of heavy-tail sockpuppet attack that I had been thinking of before slightly easier to run and more likely to affect people besides the creator running the attack.

Yes, it does. That's why it'd have to be done alongside anti-sockpuppeting measures.

[Arigatomina]

Likes:

1. Derobert's idea to use a reviewer's personalized scoring system to determine the 'distribution/weight' of the scores instead of comparing it to the overall sight average. This stops the comparison of 8 as the 'average given' for those people who actually give a 5 on average, and it doesn't hurt those people who score just like the 'average reviews given' score for the site as a whole.

2. The new 'Great videos' list idea. We still have Phades Required viewing, don't we? Let him decide. Anyone who doesn't like his personal bias can go make their own website. Or better yet, let him update his list and make it more accessible (last I looked, it was hidden on the reviewing page).

3. Making old reviews count less than new ones. I've redone about 50 of my old reviews, but I still have a lot that are scrap material at best. If there was an easy way for people to avoid this problem, I doubt anyone would complain. The only ones hurt would be those who study reviews and learn to be 'as good as they'll get' with their very first reviews. I don't think there are many people like that here.

4. Having an admin score the usefulness of reviews. Most of the complaints are about how a creator will abuse this. But the suggestion was an admin *or* a creator. I have no problem with an outside person scanning my reviews and checking the ones that are obviously more useful. Who knows, maybe it would even get rid of those 'copy and paste text' people who abuse the "Usefulness" list. If an admin (or undisclosed party) checked all reviews, people wouldn't be so free with their abuse (flames included).

Dislikes:

1. Having a weight on popular vs unpopular anime used. I've only found one or two Eva vids I liked, and I dislike the anime. Just because I'm one of those rare people doesn't mean my opinion should matter more than the majority of viewers reviewing Eva videos. On the other hand, most of the videos I've reviewed were DBZ videos. I've gone into *great* detail with most of those reviews, my average score given being a 5 since I purposely sought out beginners who'd need some feedback. I love DBZ. That doesn't mean those scores should be counted less because I'm a fan of the anime - fan does not equal blind. And when I find a rare 'great' DBZ video, I don't want my score to be counted less just because I like the anime - I like a lot of anime, and that doesn't change my scoring method. I can see where this idea came from, but I think it would hurt more than it would help.

**EXCEPTION: If you take into account the reviews given by the people who like or dislike the anime, then I'd be all for it. Someone who marks "dislike" and scores high should be counted more than someone who marks "dislike" and scores low. The same for those who "like" the anime, since the low scorers may be the more picky scorers. The only problem would be the "Anti-fanboy" flamers, and they can be removed if they're obvious enough.

[Bote]

If you put down popular anime videos then the whole idea about a top list loses its sense. Since some creators here just hunt for new animes in order to brag how they used it first, this will just give them more encouragement.

What I recommend before doing any changes to the Top10 is to try and discover multiple account abuses, which are at such intensity that 3 out of 10 first rated videos on the Top10 have about 100 multiple accounts all togheter. I can't even begin to imagine how "helpful" can 30 straight tens be to their ratings. I'm not gonna give names here since that'd resolve in the creation of "Crucify Bote" cult . Just try examining the accounts of certain videos that had got there with only 60 or 70 opinions (of course they have a lot more now). There are plenty of quite obvious indications like "Join Date" and "Last Login" times differ by only several minutes, the Top Fav. list has videos from only one creator and all the opinions given are given to that creator only. Those accounts don't have any videos listed nor are available for dld.

This can take a lot of time I know, but take it into consideration. Some people really have to be discovered and punished. It's the only way to lessen all the illegality (to say it that way) that happened last year and is still occuring.

[Zarxrax]

As far as people making multiple accounts for giving fake opinions... the forum records people's IP address. Does the site do this as well? If the site records IP addresses, you could just see if a bunch of different accounts are accessed from the same IP...

[AbsoluteDestiny]

Does the site do this as well?

Nope.

[Kalium]

As far as people making multiple accounts for giving fake opinions... the forum records people's IP address. Does the site do this as well? If the site records IP addresses, you could just see if a bunch of different accounts are accessed from the same IP...

Wouldn't this be fairly useless for people like AOL users?

[derobert]

Does the site do this as well?

Nope.

Actually, it does somewhat... you can try looking in the timings tables (or get Phade or me to check the apache logs), but we should probably fix it to do a better job.

[derobert]

Wouldn't this be fairly useless for people like AOL users?

A 90% solution is better than a 0% solution. We'd have to investigate accounts implicated by the check to make sure its not just ISP webcaches, two different family members, etc., but it might work as part of generating a list to check.