Amazon makes it clear that people aren't allowed to vote for their own reviews or comments, period. Simultaneously, though, it has a surprising loophole open to the unscrupulous. If anybody sets up a no-purchase sockpuppet account, he or she would be barred, of course, from posting reviews or comments on Amazon, but would nevertheless be granted full voting rights on the reviews and comments of OTHERS!!!
A self-promoter could set up any number of these phony accounts within a few minutes, and they'd be at his disposal whenever he sought to inflate his own vote totals. This seems the simplest explanation preserving all the appearances and explaining the outrageous vote totals marking the Grady Harp Express these days.
Editor note: you don't even need to amass such accounts: you can create a new one for every 'voting session'. It's that simple: not only don't you need to buy anything from Amazon, you literally need nothing other than a throwaway email address, and even that most likely isn't verified. Takes ten seconds to type in a nickname, email address, and password, and then you're in business.
Subscribe to:
Post Comments (Atom)
14 comments:
This is an amazing discovery! As everyone knows, one must have purchased something on Amazon in order to post reviews and comments there. Yet, as Stanley's brilliant discovery shows, no purchase is necessary in order to set up a voting-only account. This takes just a couple of minutes. One can create any number of them and immediately start voting.
I believe this is the answer to the Harp enigma: he probably creates a few of such accounts at a time and starts using them himself. That explains the gradual growth of his voting totals from the time when another sharp observer, Office Lover, raised his objections, which was about a year ago (then it was about 30 helpful votes) to the mind boggling 150, and sometimes even over 250 today. In other words, while it's easy to set up one such account, it takes some time to create 250 of them.
As far as Harp's huge number of Amazon friends -- along with insistent suggestions by, um, some well-familiar posters that Harp's stratospheric helpful-votes totals come from them, I think Harp likes to amass these Friends for the purposes of "plausible deniability": I truly can't imagine that all these people, Bill Bradley among others, will bother monitoring Harp's reviews and review commentary and, carefully taking turns, one at a time, in groups of between five and ten, vote for Harp's reviews and against the critical commentary under them. Amassing Amazon Friends may be another reason for the graduality of the growth of Harp reviews' helpful vote totals: the two must be coordinated, the vote totals not exceeding the number of Friends at any time.
With all this in mind, I bet the Mystery Voter whose eerily relentless processing of the Harp reviews' votes is none but Grady Harp himself, armed with a gigantic and ever growing number of easy-to-set-up voting-only sock-puppet accounts. Kudos to Mr Nemeth for his brilliant insight: this explains everything.
PS. What could be Amazon's thinking behind this design? Why, in their mind, posting and voting require different degrees of authorisation?
An interesting discovery and I tested it out with a fake email. Voila! I can vote on reviews and comments but I can't post comments. I wonder why they did this?
However, I'm still favoring the theory that he's doint it mechanically. Try timing how long it takes between votes on his reviews/comments as opposed to how long it takes to log out, log in, vote, log out, log in, vote, etc.
You don't need to log in and out, just make a throwaway login, plug your votes, blow away the cookies, repeat. Couldn't be simpler, and yes, this is eminently automatable.
But I think Harp is doing it manually, because a lot of his vote is selective -- remember he not only votes for himself, but also against any critical comments so as to get them hidden/deleted-by-Amazon. That latter operation requires selectivity which a bot wouldn't offer. Well, I mean, a simple bot, and where our brilliant Maestro would get a super-duper Lazertronix AI voteBlaster(TM) bot. I bet he puts on his glasses, gets his elderly ass into a chair, spits on his thumb, and, grunting and squinting, makes a couple of passes voting for his brilliant reviews and against critical comments.
I am truly surprised more people don't do this. Is Harp stupider than everyone? That just can't be.
"I am truly surprised more people don't do this"
It's a lot of work, ony an egomaniac would bother. After ten (I think) votes from one identity you can't affect ranking anymore. If these votes counted, he would have climbed at least one or two rankings wouldn't he?
"After ten (I think) votes from one identity you can't affect ranking anymore" - what do you mean?
It's been discussed several times on the DB's. I have several reviewers I follow regularly, and I always give them helpful votes (because they are helpful). After a certain point in time (ten was the magic number mentioned) those votes will count towards VOTE TOTALS, but will no longer affect ranking. Same goes for the negatives (trolls -- take note).
I think it's a built in thing by Amazon to control voting rings and how they affect ranking. So, Grady may create all these new identities, but after so many helpful votes he really can't affect his ranking much. However, massive HV's do help to get one's votes in the Spotlight Column and that's where everyone wants to be.
Does Amazon spell it out anywhere? I mean, how do people know the workings. L.E. Cantrell (hello, L.E. :-) would call it 'idle speculation'. And, in that particular case, he'd be right.
I know it's been discussed and analized on the DB's to death and the real formula is a deep dark secret at Amazon, as best as I can recall from previous discussions. Feel free to search the DB's or post a new thread, I'm too tired to think further about it tonight :o. And no, in this case it's NOT LE Cantell who analizes it (at least not this time).
No, I don't care to research this either :-) . I do remember reading some official Amazon missive going something like, "if you get a lot of votes we just may use them to promote you in the ranks" or something like that. But the "may" part renders the whole thing meaningless.
I wonder why this is such a secret... probably to make it look like what it ain't.
Folks:
I don't think the central issue here is the oddity of what Amazon chooses to do with positive or negative votes in its ranking system. Instead, the issue is over whether such votes have in fact been legitimately cast, given the discovery that any enterprising person using voting-only accounts can easily inflate his own voting totals and get negative criticism covered over. The opportunites for shameless self-promotion (whatever one's "ranking"), gross product-pushing, and deluding the public by means of phony raves are legion. Though on the surface Amazon has impediments in place against self-voting, the easy ways around these are matters the website seriously needs to address.
I agree; the other stuff is curious but not nearly as important.
Btw, about your "on the surface Amazon has impediments in place against self-voting, the easy ways around these are matters the website seriously needs to address" -- perhaps my opinion of humanity is too low, but it seem to me that this has been done on purpose: so that an occasional visitor would see that he needs to log in in order to vote, and think the voting (in general) is valid, protected, authorised, and so on.
And at the same time, behind the curtains, as it were, they (Amazon) give their beloved shills a completely open/unprotected voting mechanism that they can use to censor commentary (and vote for themselves too).
This is just like this delete-and-repost "oversight" -- I mean, it's an oversight alright, but somehow all these "oversights" boil down to giving more power to the shill and less power to the paying passerby. And I don't see a lot of eagerness to fix them too. Or even to acknowledge their existence!
They (Amazon) did quietly change the "deleted-by-Amazon" method to the hiding of comments. Nothing seems to get deleted outright anymore: but man, that was done sort of on the sly.
Delete-and-repost seems to have subsided recently too: were, um, words exchanged between Amz and their reviewing peons? All our Friends Top Reviewers, including Maestro Harp, of course, used to get rid of commentary by deleting and reposting their reviews in rather large numbers: not anymore -- why? What happened? Delete-and-repost still works, no?
When you're dealing with the corporation and some of the characters we've all been dealing with, it's pretty impossible to have too low a view of humanity. Suspicions that are on the money invariably turn out to be those too true to be good.
At the same time, we've all known individuals and corporations who'll do the right thing, though only as a last resort because they've been exposed and forced to. Maybe Amazon can be shamed into closing the loophole, doing the right thing for the wrong reasons.
"Suspicions that are on the money invariably turn out to be those too true to be good."
:-)))) That's one way of putting it!
Btw, an interesting topic for discussion:
I can understand that some sort of oversight should be exercised over a public discussion venue, such as the Amazon comments area. But who is to be the censor? I think most people would naturally chose someone known as competent and fair: a trustworthy player who'd moderate discourse according to some publicly known and widely accepted rules.
Is that what takes place on Amazon? Nope. Amazon obviously believes that the best censor is a software automaton driven by a mob (at best), or a clique (most likely in reality): the commentary on Amazon gets censored based on the number of clicks on the "customers don't think this adds to discussion" button.
Who are these customers? Isn't the poster finding himself at the receiving end of this mechanism also a customer? I feel this "customers don't find this useful" is a bit of a misnomer.
The real meaning of this procedure is different: what the button should read is this: the ratio of negs ("not-contributing-to-discussion" clicks) to positive clicks is over 3.75 (that appears to be the point when the comment gets covered).
Combined with the ease of voting (effectively, anyone can vote any number of times) this hands the censorship of the board to whoever clicks the most. This could be a clique or a persistent single man.
Does this manner of censoring content make sense? Not to me. And if this is yet another "oversight", doesn't it again, hand the actual power of muzzling anyone to the shill?
It is highly unlikely that an occasional visitor (real customer, btw) will have -- or will bother, or will be able to organise -- a voting gang, or that he'll know how to fudge votes or even suspect that they can be fudged to begin with.
The shill, on the other hand, dwells on the site permanently, knows how the system works and so can game it, has long-term plans and is thus motivated to control the environment; finally, the shill is not alone: there's plenty of them there and their professional, "guild" solidarity will naturally motivate them to help one another, in an organised way if necessary.
So, who's more likely to benefit from the system as it is right now? The lone, unsuspecting, and frequently gullible customer -- or the familiar with the system's "oversights" and well-versed in their use for his benefit shill?
Post a Comment