Cheats on Kinopoisk, look at the accounts closer
Based on the latest news about hacking accounts at KP, I wanted to indulge with rating statistics, which can be partially spied on the site.
It is noteworthy that already on September 18, they wrote on the Internet that other films coming out on September 28 were twisted in a rating, but the KP was stupid for a long time, although all the accounts lowering the ratings there were Novoregs for about two weeks. These users put 2-3 random ratings at registration and another 2-3 at the time of the rating roll (Moreover, the first three were 5-10, and the second 3-4-5, that is, completely pale). I think with the same accounts and screwed up the rating of expectations, another film, which we will not name yet. Why they could not be cleaned up before the 28th is a mystery, on the KP forum, rubbing a few times a day on the comments page on the topic ...
They put marks on the tapes: Not your body (2017), Bobby Hedgehog: Spiky Adventures (2016) and Good Time (2017) , have a nickname of the form IvanIvanov96, are registered for less than a month and have less than 20 ratings.
Such estimates ~ 10k.
Since I last studied web programming in the 11th grade, I didn’t work out the stylishly fashionable youth programming languages suitable for parsing the Internet and used what was at hand, namely MATLAB. Because KP uses dynamically loaded data on the ratings page, it didn’t come up with anything better than adding pens to the end, and the site believes that 1000 ratings will be enough (you can probably play with the parameter
Next, search the file and look for the lines
By saving users' ratings to the id array, we are aware of their profiles and rating pages. To do this, save the pages of the form
And as soon as several thousand pages are downloaded (and this is a long time, you need to play with delays between requests, so as not to catch errors on timeout), you can see what is interesting there. In the course go
This is how the distribution of the number of ratings looked for users with the number of ratings from 1 to 20 and for the rest 21 or more. There is such a noticeable peak of users with 5 ratings.

Accounts were registered for 20 days, starting from August 24th.

If you look for whom these users actively voted, then in addition to the two films mentioned above, one can single out Crimea (2017) and Love pret-a-porter (2017).

To create visibility, more than a couple of hundred votes were given for films from the first thousand and the last popular ones (like Zeropolis and La-La-Lenda), which no longer go to the cinema, the distributions are approximately the same, the ratings are only 3 to 9 with a slide of 7 -8 and a small humpback on 3ke (it's not clear why). In general, I can hardly imagine a person putting 3 Escape from Shoushenk)).
Top 3 ratings, they certainly do not palilis:
Good time: 5215 suspicious estimates

Not your body: 5090/5215

Hedgehog Bobby: Thorny adventure: 5079/5215

And then new, At least the distribution of a no ratings tightened ...
Love pret-a- port: 4432/5215

Crimea: 793/5215 (I wonder if he will soon get into the top of the worst )

Well, as they say on Picaba, all these films are connected by someone Alexey Pimanov

Of the 5217 suspicious accounts, 5215 have a link to the page on VKontakte, so I think people put some toy / application (like see who went to your page) in vk and they are registered with KP with or without their knowledge, or they just have an easy password. If someone wants to study, all the links are here . And I very much doubt that the message from the Kinopoisk administration will reach the users of VKontakte and they will somehow stir and notify the users of the VKontakte administration. In my opinion, KP could well completely clean up these accounts.
UPD1:While looking at the revived bots that began to rate Crimea from 4 to 10, I noticed that the profile information differs from the pages on Vkontakte and when registering through it, a nickname is given from vkontakte, so most likely the contact was used simply as a base for names for registration, without stealing uchetok.
UPD2: If suddenly this page is read by the Kinopoisk employees, I ask them to immediately check the movie’s rating expectations Tankers (2017) . There are a lot of fakes that are now being tumbled over for the Crimea only have it in the list of expectations, which is 12k votes so far. In order not to slow it down on the last day before the premiere ... If they don’t have time, they will have to download the pages of expectations all night and build more pictures and show where the fakes are accurately marked and counted by Kinopoisk.
In the meantime, I look forward to when Film Search will retaliate, which reminds me of on every page:
It is noteworthy that already on September 18, they wrote on the Internet that other films coming out on September 28 were twisted in a rating, but the KP was stupid for a long time, although all the accounts lowering the ratings there were Novoregs for about two weeks. These users put 2-3 random ratings at registration and another 2-3 at the time of the rating roll (Moreover, the first three were 5-10, and the second 3-4-5, that is, completely pale). I think with the same accounts and screwed up the rating of expectations, another film, which we will not name yet. Why they could not be cleaned up before the 28th is a mystery, on the KP forum, rubbing a few times a day on the comments page on the topic ...
They put marks on the tapes: Not your body (2017), Bobby Hedgehog: Spiky Adventures (2016) and Good Time (2017) , have a nickname of the form IvanIvanov96, are registered for less than a month and have less than 20 ratings.
Such estimates ~ 10k.
Since I last studied web programming in the 11th grade, I didn’t work out the stylishly fashionable youth programming languages suitable for parsing the Internet and used what was at hand, namely MATLAB. Because KP uses dynamically loaded data on the ratings page, it didn’t come up with anything better than adding pens to the end, and the site believes that 1000 ratings will be enough (you can probably play with the parameter
prem_max_page
and can manage to save more) and save it as an html file to my disk. Next, search the file and look for the lines
/votes/" class="comm-link"
next to which are the id of the users we are interested in. By saving users' ratings to the id array, we are aware of their profiles and rating pages. To do this, save the pages of the form
https://www.kinopoisk.ru/user/***id***/
and a https://www.kinopoisk.ru/user/***id***/votes/
wonderful team. websave(filename,url,options)
And as soon as several thousand pages are downloaded (and this is a long time, you need to play with delays between requests, so as not to catch errors on timeout), you can see what is interesting there. In the course go
strfind()
and textscan()
with all sorts of pieces of addresses. I managed to collect 3740 ratings sorted by date and another 2245 unique when sorted by user name, for a total of 5985 pages rated movie # 928796 out of 12k of all ratings (could not find 6000 ratings 3-4-5, the remaining ratings were less than 1000, so among there are no gaps) ratings for the other two films with 11k KP estimates are hidden under the pretext of a small number of valid ratings.This is how the distribution of the number of ratings looked for users with the number of ratings from 1 to 20 and for the rest 21 or more. There is such a noticeable peak of users with 5 ratings.

Accounts were registered for 20 days, starting from August 24th.

If you look for whom these users actively voted, then in addition to the two films mentioned above, one can single out Crimea (2017) and Love pret-a-porter (2017).

To create visibility, more than a couple of hundred votes were given for films from the first thousand and the last popular ones (like Zeropolis and La-La-Lenda), which no longer go to the cinema, the distributions are approximately the same, the ratings are only 3 to 9 with a slide of 7 -8 and a small humpback on 3ke (it's not clear why). In general, I can hardly imagine a person putting 3 Escape from Shoushenk)).
Top 3 ratings, they certainly do not palilis:
Good time: 5215 suspicious estimates

Not your body: 5090/5215

Hedgehog Bobby: Thorny adventure: 5079/5215

And then new, At least the distribution of a no ratings tightened ...
Love pret-a- port: 4432/5215

Crimea: 793/5215 (I wonder if he will soon get into the top of the worst )

Well, as they say on Picaba, all these films are connected by someone Alexey Pimanov

Of the 5217 suspicious accounts, 5215 have a link to the page on VKontakte, so I think people put some toy / application (like see who went to your page) in vk and they are registered with KP with or without their knowledge, or they just have an easy password. If someone wants to study, all the links are here . And I very much doubt that the message from the Kinopoisk administration will reach the users of VKontakte and they will somehow stir and notify the users of the VKontakte administration. In my opinion, KP could well completely clean up these accounts.
UPD1:While looking at the revived bots that began to rate Crimea from 4 to 10, I noticed that the profile information differs from the pages on Vkontakte and when registering through it, a nickname is given from vkontakte, so most likely the contact was used simply as a base for names for registration, without stealing uchetok.
UPD2: If suddenly this page is read by the Kinopoisk employees, I ask them to immediately check the movie’s rating expectations Tankers (2017) . There are a lot of fakes that are now being tumbled over for the Crimea only have it in the list of expectations, which is 12k votes so far. In order not to slow it down on the last day before the premiere ... If they don’t have time, they will have to download the pages of expectations all night and build more pictures and show where the fakes are accurately marked and counted by Kinopoisk.
In the meantime, I look forward to when Film Search will retaliate, which reminds me of on every page:
Crime and Punishment
Some users want to influence the rating of the best films of KinoPoisk - Top 250, registering several accounts and / or exposing films to only a few or dozens. Especially for such site visitors, there is a simple rule - several filters are used to calculate the Top-250 rating, eliminating such cheatings (the rating takes into account the votes of regularly voting users, eliminating 100% of false dozens and units).
All attempts of unscrupulous users to “wind up” the rating will be stopped. At best, such assessments will be simply deleted, at worst, the account (or accounts) of the radishes will be banned, access to the site will be blocked, the computer will soon break down, and an unfriendly ghost will appear in the attacker's house.
Then do not say that you were not warned. :)