List manipulation?

Logan wrote: Easy Livin wrote: Logan wrote: What bugs me is the suspicion that people who rate albums one star, while giving a fave band/ artist's album five stars in order to get their fave a higher ranking, haven't even listened to the one star album, let alone given it a fair hearing. I wish that at least one star ratings required reviews or there could be a rule limiting  allowability when it comes to someone rating one album a one and another a five that are clearly competing in a "top" albums list. You're right Logan, there is a risk that we tar people with an unfair brush. On the other hand, when people rate a band's entire catalogue with one star, or rate every album in our chart above their own favourite with one star, I think we have strong evidence of fixing. I agree, and actually noticed a mistake in what I'd said after reading it later but opted no to edit it.  I'm not actually bugged that people suspect people of what I said, I suspect those people of what I said, and that suspicion makes me kind of itchy (it's too late for me to express that better).  It bugs me that people can a rate a one, and we don't even know if they've even listened to the album. Of course one could rate an album anything between one and five without having heard the album, and even review it without hearing a note while making it seem as if one knew it.  In a way, I really would prefer it if a one rating required a review -- some justification given.

Much sense in both posts certainly. As much as we would would prefer written justification for reviews at the polar extremes (1's and 5's) there is that delicate 'non english speaker' critter to contend with.

The only way to close that bolt-hole available to the fanboys would be to put the onus on the prospective member of PA to prove they couldn't speak English. (And we would end up appearing like Immigration Control methinks  by asking them to complete a very basic English literacy exam)

That said, most genuine new members may not have any objections to such a condition if we explained the rationale behind it ?

Easy Livin wrote: Logan wrote: What bugs me is the suspicion that people who rate albums one star, while giving a fave band/ artist's album five stars in order to get their fave a higher ranking, haven't even listened to the one star album, let alone given it a fair hearing. I wish that at least one star ratings required reviews or there could be a rule limiting  allowability when it comes to someone rating one album a one and another a five that are clearly competing in a "top" albums list. You're right Logan, there is a risk that we tar people with an unfair brush. On the other hand, when people rate a band's entire catalogue with one star, or rate every album in our chart above their own favourite with one star, I think we have strong evidence of fixing.

I agree, and actually noticed a mistake in what I'd said after reading it later but opted no to edit it.  I'm not actually bugged that people suspect people of what I said, I suspect those people of what I said, and that suspicion makes me kind of itchy (it's too late for me to express that better).  It bugs me that people can a rate a one, and we don't even know if they've even listened to the album. Of course one could rate an album anything between one and five without having heard the album, and even review it without hearing a note while making it seem as if one knew it.  In a way, I really would prefer it if a one rating required a review -- some justification given.

LiquidEternity wrote: I don't necessarily mean in this case. I just mean it might happen in general.

If the rater only gives 1 & 5 star ratings then it is a sure fire-sign of fanboyism and not a anti-weighting protest. As I said - I doubt these people know about the weighting.

I don't necessarily mean in this case. I just mean it might happen in general.

I think people who don't write reviews are more inclined to rate all the way one way or another, because they feel like their opinion weighs less. So if they want to feel like they made an impact on the rating, they use the full weight they can. Maybe our weakened rating-only weighting is prompting more list manipulators.

I wouldn't have a problem with throttling thread making and rating without a review for a few days, but weighting based on join date is a terrible idea. Making a review 5 times stronger than a rating is deterrent enough, there's no need to weaken a rating further by forcing you to have a lot of them. I wouldn't cry if new accounts weren't allowed to post at all the first few days, but it would be more fair and still effective to only limit them to existing threads and have to write a review for their ratings.

Here's what I do at PF:

I manually assign weights to all the users. If an account was created only to submit a couple of ratings and then never used again, I simply assign a low weight. If the user comes back after a while and becomes more active, I'll see it and will be able to adjust the rating again.

PA has much more users than PF, but I still think it can be done. All you need is a table where users are listed in chronological descending order, with a link to their profile and the weight number and a way for the admin to quickly edit it. You could also list the current number of ratings , min/max ratings and the mean deviation (function "stdev" in MS-SQL Server). Usually spammers can be recognized by submitting both 1-star and 5 star ratings combined with a high standard deviation.

Cheers Guys, some pretty crude attmepts being made to spam the ratings. All the recent  IDs had the same IP address.

Yeah that looks very suspicious:

• solar plexus
• Universs (Dmitry)
• Quilombo1988 (Alberto)
• szmataarmata (Bartosz)
• Magnum59
• sauromat (alexander)
• @thehop
• EduMusic
• philGab
• Tubbs (Josh "Tubbs" Earle)
• diotliz
• jean_delacroix
• Mariachi
• Arkadiusz
• Andy_Fripp (Andret)
• psyk
• arek
• Przemonian
• LambertPL
• MagiKorna
• zatorrr
• xziaja
• morse4life

Cheers guys, matter now dealt with.