Skip to Navigation | Skip to Content

Episode 2, 11 February 2013 

News gets gamed

Alas, we haven't finished with News Ltd tonight.

Here's a story published last October on news.com.au

SPEED UP: Drivers vote to increase speed limits after report recommends limit be lowered

• News.com.au, 8th October, 2012

It was a follow-up to a story in the Brisbane Sunday Mail the day before, which asked readers to vote on whether speed limits should be lowered. The follow-up reported:


Of more than 17,000 votes cast at the weekend almost 16,000 went for the response: ‘No, speed limits should be raised.’

• News.com.au, 8th October, 2012

Sound a bit unlikely? A week later, the reporter who'd written the original story received an emailfrom someone called Russell Phillips.


...I have a confession

• Russell Phillips, 16th October, 2012

The email included a link to this blog:


The Ubermotive Guide to Media Influence

• The Ubermotive Guide to Media Influence, 14th October, 2012

It carried a screenshot of the story about the speed limit poll, and went on:


... I have some news for you: over 15,000 of those votes were mine.

• The Ubermotive Guide to Media Influence, 14th October, 2012

The author of the blog, Melbourne software engineer Russell Phillips, explained that as long as you clear the cookies from your computer each time you vote on a News Ltd online poll...


... there is nothing on their servers that will limit how many times or how often you vote ...You could vote all day, but manual labour is for suckers.

• The Ubermotive Guide to Media Influence, 14th October, 2012

Instead, Phillips used his programming skills to create a cyber-robot, or bot, that voted for him - around 15,000 times.

Having warned News Ltd what he'd done, he waited for a response, and got none. So he kept doing it. But as he explained in a Reddit post, he tweaked his bot first:


I actually wrote a program where for each option someone voted, my program would vote once for every other option, thus maintaining a deadlock.
• Reddit,

And News Ltd would faithfully write up the completely phony results. For example ...


The great vaccination debate: are you for or against?

By Lucy Kippist
November 29, 2012

A poll, run on news.com.au on Monday, asking whether vaccination for children should be mandatory, received more than 170,000 votes. Of those who voted, 50 per cent agreed.

• News.com.au, 29th November, 2012

A hundred and seventy thousand votes? Split fifty-fifty? That same day, Phillips emailed the story's author...


Hi Lucy

I'm writing to inform you that the poll results that have made front page news have, in fact, been synthesised ...

• Russell Phillips, email to News Ltd, 29th November, 2012

Once again, he was ignored. So was a colleague, who emailed next day...


It is ... unbelievable that 170,000 people voted in your on-line poll...
On-line polls are easily hacked. Yours obviously was. When are you going to print a retraction...?

• Email to News Ltd, 30th November, 2012

Still no reaction. So Phillips's bot kept voting - for example, in a poll asking readers about 2DayFM's prank phone call to the Duchess of Cambridge's hospital ward. The rorted result was dutifully reported.


Yet a news.com.au poll, asking whether the presenters should be blamed for the death, was deadlocked at 50-50.

• News.com.au, 14th December, 2012

Once again, Phillips's colleague warned the reporter that the poll had been hacked.


A printed retraction might be appropriate. You might even wring a whole article out of it.

• Email to News Ltd, 14th December, 2012

Well, there was no retraction - but at least news.com.au removed all reference to the rorted poll.

Editor Luke McIlveen now tells Media Watch that he became aware of the poll-tampering on December the 18th


Our IT department has recently upgraded security on all News Limited online polls. While polls will continue to appear in some stories ... News.com.au will not be publishing stories based on the results.

• Luke McIlveen, Editor, News.com.au, 8th February, 2013

But it took a while. Right through January and into early February, Russell Phillips's bot was still finding polls to disrupt on the websites of News Ltd mastheads.

So far, there's been no correction by News Ltd of stories based on the rorted polls, and no account of how they were gamed by a reader. Luke McIlveen told us...


I have not published a story about the hacker's activities because I believe this individual should not be afforded any publicity.

• Luke McIlveen, Editor, News.com.au, 1st February, 2013

Nice try Luke. But we disagree. We reckon he's demonstrated that online newspaper polls are scientifically worthless.

Today, Phillips told Media Watch that he's cracked News Ltd's latest defences - and Fairfax Media's as well. This poll in The Age, he claims...


Was Environment Minister Tony Burke right to reject a bid to protect the Tarkine Wilderness from mining and logging?

• The Age, 8th February, 2013

...was turned around by his bot just last night.

Full responses, and more details, on our website. And next week, we sincerely hope, we'll find other news outlets to shove around. Till then, goodnight.

 

YOUR COMMENTS

Comments (13)

Add your comment

  • Geoff :

    12 Feb 2013 6:00:31pm

    Why can't the online polls be used as a way for readers / subscribers to quickly 'have their say' on stories they might find interesting.
    This SOFTware bloke has ruined online polls for everybody. He probably breached most terms and conditions for using the websites and should be punished accordingly. Instead, MW hails him as a hero.

    Reply Agree (0) Alert moderator

      • Yellek :

        13 Feb 2013 9:03:21am

        The story here is that the rorting is being disclosed to News Limited and Fairfax so they can fix the problem. From the sounds of it his exploit isn't hard to do so the question remains who else is rorting the polls without anyone's knowledge?

        Given the opportunity unscrupulous people WILL rort polls to get their message out so isn't it better to know?

        Reply Agree (0) Alert moderator

      • James :

        24 Feb 2013 1:19:43pm

        Good news ! Online polls CAN be used for a way to quickly have our say !

        News.com.au just needs a proper polling system that is not easily rortable, which they are obviously not interested in.

        You can host a flawed poll and then blame others for the flaw. Fix the poll, save the cheerleader, save the world.

        Reply Agree (0) Alert moderator

  • Ouroborus :

    12 Feb 2013 3:18:15pm

    "Cyber-robot"? Really?

    Reply Agree (0) Alert moderator

  • Peter Allan :

    12 Feb 2013 3:17:41pm

    Welcome back Media Watch, so disappointing you have to take a break.
    Seems Russell Phillip may have been at this for a while longer than some may think.
    Note the poll in the site below, where Bob Katter, with 46%, does remarkably well. However, to help get him from an earlier 36% to that figure, actually required around 1,000 votes out of the last 400, yes 400, lodged. Even earlier results had him polling at just 8%.
    Also interesting, in light of things, listening to former Liberal Party researcher Dr Scott Prasser "airing his knowledge" in the interview with Spencer Howson.
    http://www.brisbanetimes.com.au/queensland/newman-shrugs-off-poor-poll-20120807-23ret.html#ixzz22vKOa7CP


    Reply Agree (0) Alert moderator

  • Deb Davis :

    12 Feb 2013 1:44:09pm

    Just another indication of the fact that people don't want to spend money to get the truth. They want a quick cheap fix. Like fast food. Good research on the other hand takes wisdom, experience and money. Many in business are happy today to tick a box with a project and be given any numbers true or false. Russell will be getting lots of job offers.

    Reply Agree (0) Alert moderator

  • Philip :

    12 Feb 2013 11:11:53am

    Welcome back Media Watch to 2013, and the article on manipulating online polls was most interesting, with three broad outcomes: achieve an overwhelming vote of support for a point of view, and see it written up as if it was real, achieve a deadlock, and for Fairfax Media, totally turn around a poll.
    Clearly these online polls have limitations. But the bigger picture is the severe pressure on newspapers whilst trying to stay in business (with less resources for checking stories), whilst a major on line outfit is to able to make big profits and export them near tax free.

    Reply Agree (0) Alert moderator

  • Brett Harten :

    12 Feb 2013 10:59:24am

    Simon, the difference is, The Drum doesn't write stories based on online poll results! A little bit different, no?

    Reply Agree (0) Alert moderator

  • Simon :

    12 Feb 2013 5:42:05am

    I've been doing the same thing to ABC polls for ages. It's easy to turn around polls on The Drum in a couple of hours. Vote 6000 times while you eat lunch.

    I understand though - that's not News Ltd so nothing for Media Watch to concern itself with.

    Reply Agree (0) Alert moderator

  • Dave Davies :

    12 Feb 2013 2:42:10am

    All I can say is that I love Russell Phillips! Voting polls on websites really don't deserve any serious consideration at all, especially in the form of news story. You have no idea who votes (could be a young child), plus those who vote are only those people who would visit a particular site - and I suspect that most reasonable, intelligent and well informed persons are not going to bother their time on clicking on a news.com.au poll or even visit the site.

    Reply Agree (0) Alert moderator

  • matt :

    12 Feb 2013 12:03:23am

    story worth telling. thanks MW. i have missed u

    Reply Agree (0) Alert moderator

  • Hamish Moffatt :

    11 Feb 2013 11:46:08pm

    Thanks for bringing this to our attention - I found it hilarious. Here is a link to the Reddit article.

    http://www.reddit.com/r/technology/comments/151pur/i_spent_three_weeks_fixing_polls_on_newscomau/

    Reply Agree (0) Alert moderator

      • Bondiboy66 :

        04 Mar 2013 3:14:43pm

        50% of polls are wildly skewed in favour of the proposition suggested by the poll question, whilst %50 show repspondents reject the poll proposition.

        Reply Agree (0) Alert moderator