Dumplings from this Panda!

Oct 25

They are not spammers, what have they achieved?

How many pops did these domains gets by diggs from these suspects? 229. However, just one digg from one these IDs should not make any of the stories by themselves suspicious, so I am now going to list all of the 229 stories and the number of suspect diggs and non suspect diggs. While whether the digg is suspect or non suspect is clear, remember that due to promote_date confusion in digg data, the total number of upcoming diggs of a few stories might not be accurate. Also remember that you are only seeing data as of 11pm CST on Oct 23rd, while this is still continuing to happen.

Link to interactive and detailed version of this data.

Now that each of the story has been given an “ID”, we will use it for our reference. Did you notice that story with ID 1, got only 1 actual digg!. Yes, all it took was guardian.co.uk to submit the story and the rest was taken care (by who?). Any story with or more than 100 upcoming diggs, for sure has promote_date problem in it, so let’s for now leave those stories and crunch a few numbers. Also stories 209, 219 and 221 were excluded as they are clear outliers. For the rest of the stories (leaving out 31+3), 10016 suspect diggs were cast, they also had about 4055 non suspect diggs, but this 4055 is very far high from the reality, due to the promote_time bug/feature.  To get more reasonable estimate of the problem, let’s now only consider stories which needed 60 or less upcoming diggs, as these stories clearly are not a part of the promote_date bug. In this case, 986 diggs out of 1257 diggs were suspicious, that is 78.44% of diggs on these stories are suspicious.

There are a few interesting domains, submitters and stories to note here, which are discussed in a later page.

Next Page

  1. ltgenpanda posted this