Twitter takedowns revealed.

AN agency under contract with twitter conducted a study of Twitter spam and purchased Twitter accounts. The information revealed by this study was used to identify fraudulent twitter accounts, where they came from, the methods used to create them and some of the sellers of those accounts. The paper reveals that many of the fraudulent accounts were purchased here at Blackhat World. The paper names names and services that were selling those accounts, and sometimes shows how the same sellers, resold the same purchased accounts several times.

The study is 16 pages, the title is "Trafficking Fraudulent Accounts: The Role of the Underground Market in Twitter Spam and Abuse, " and the link to that study is http://krebsonsecurity.com/wp-content/uploads/2013/08/USENIXtwitteraccountpaper.pdf.

For anybody considering purchasing Twitter accounts, this paper is a "Must read." For anybody that creates those account, this paper is a "Should read."

For those sellers that sell the same accounts: Shame on you.

Happy reading.

2.2 Impact of Fraudulent Accounts:

Todays word of the day is "Miscreants", lol.

Well that would explain the recent panic by some twitter sellers and resellers. Seems Twitter has hired some pretty smart people to create a very good algorithm picking off fake accounts. 99.99% is a success by most standards.

I'd hate to be a Twitter account seller or reseller atm.

Hah, just bought a service off here 2 days ago.

It was crappy, I'm glad I'm seeing those "permanent active followers" dropping off already. Didn't even have bio pics.

lol's..

Narcs. Narcs everywhere.

kvmcable said:

Well that would explain the recent panic by some twitter sellers and resellers. Seems Twitter has hired some pretty smart people to create a very good algorithm picking off fake accounts. 99.99% is a success by most standards.

I'd hate to be a Twitter account seller or reseller atm.

Click to expand...

If you read the paper closely, there are a few things that should stick out besides the math that a person could base an anti algorithm on. When the paper informs of how they detect those accounts, it also informs how to evade detection. Because this thread is public and the paper clearly identifies this forum as being on that the investigators hang on, I will not give what I picked out. The paper does, however, tell one, between the lines, how to evade the new algorithm.

One thing to learn from this study:

Break the pattern as often as you can.

Posted yesterday: http://www.blackhatworld.com/blackh...uys-buy-loads-social-networking-accounts.html

Reminds me of the white paper on FaceBook, and the methods they used to detect rogue activity.

Like most algos that set out to thwart the undesirables, it effectively boils down to this: act like a human.