Global spam e-mail drops after hacker arrests

[DorsetBoy]

http://www.bbc.co.uk/news/technology-11757347

Levels of spam have fallen by almost 50% since August 2010, suggest figures.

Figures compiled by security firm Symantec show that the amount of junk e-mail messages flowing around the net has dropped 47% in three months.

Kaspersky Labs noted a similar fall from July to September, when spam levels fell to 81.1% of all e-mails

The decline was put down to the arrests of those behind spam-sending botnets, and intelligence work that saw other spamming systems shut down.
Server shutdown

In the last few months security firms have scored several notable successes against gangs that own and operate botnets - collections of hijacked home computers.

The vast majority of spam or junk mail is routed through these hijacked machines............(more)

[Lance]

81.1% of all emails stills seems way to high. That must be putting massive demand on servers all round the world.

[Simon]

Indeed, but I guess anything is a help.

[armadillo]

I cannot get very excited about figures indicating massive short term drops. They do not explain how the figures were collected nor relative to what base level they were calculated.

There are two main factors that can cause figures such as these to be misleading.

The first is "regression to the mean". That is, suppose that there had been a large increase in spam in the three months prior to the "drop". Then the "drop of half" would simply be a return to the levels prior to the increase.

The second is volatility.

Suppose that the spam figures are regularly rising and falling. Then, by selecting your monitoring periods accordingly, you can show either a rise or a fall, pretty much as you choose. But there would be no change in the overall pattern.

It is possible that the reduction is genuine. But even if it is, it is not much cause for celebration. With such large numbers of machines recruited as zombies, it is very easy for the numbers of zombies to increase rapidly in a short time. And the arrest of a few high profile spammers has little effect on the slightly less ambitious ones. For as long as the sending of spam is profitable, spammers will continue to distribute it. And it remains profitable because there are still enough gullible people to part with money for what the spammers claim to be selling.

[Rik]

The first is "regression to the mean". That is, suppose that there had been a large increase in spam in the three months prior to the "drop". Then the "drop of half" would simply be a return to the levels prior to the increase.

Are you sure you're not a politician, Dill?

[Gary]

"The first is "regression to the mean". That is, suppose that there had been a large increase in spam in the three months prior to the "drop". Then the "drop of half" would simply be a return to the levels prior to the increase"

This sounds like Gas and Electricity prices really 

[armadillo]

Are you sure you're not a politician, Dill?

Quite sure! Politicians are usually completely baffled by things like that.

[Rik]

 

[armadillo]

"The first is "regression to the mean". That is, suppose that there had been a large increase in spam in the three months prior to the "drop". Then the "drop of half" would simply be a return to the levels prior to the increase"

This sounds like Gas and Electricity prices really 

Indeed. Lots of things follow that kind of pattern. Sometimes, policy is even based on a failure to account for regression to the mean.

The trigger for installing speed cameras always was (maybe still is) to put them somewhere where there has been a dramatic increase in accidents. After the introduction of the cameras, the fall in accident rates is then attributed to the effectiveness of the cameras. In fact, it is just a regression to the mean. Even without the cameras, the likely outcome is a return to the levels before the temporary (or periodic) increase. But by placing them in such areas, the politicians can claim that the policies are having a dramatic effect on increasing road safety.

Such introduction is unlikely to cause harm and is not especially expensive, especially if the cameras are not monitored.

[DorsetBoy]

http://www.symantec.com/content/en/us/enterprise/other_resources/b-state_of_spam_and_phishing_report_11-2010.en-us.pdf   

[Gary]

Indeed. Lots of things follow that kind of pattern. Sometimes, policy is even based on a failure to account for regression to the mean.

The trigger for installing speed cameras always was (maybe still is) to put them somewhere where there has been a dramatic increase in accidents. After the introduction of the cameras, the fall in accident rates is then attributed to the effectiveness of the cameras. In fact, it is just a regression to the mean. Even without the cameras, the likely outcome is a return to the levels before the temporary (or periodic) increase. But by placing them in such areas, the politicians can claim that the policies are having a dramatic effect on increasing road safety.

Such introduction is unlikely to cause harm and is not especially expensive, especially if the cameras are not monitored.

I remember reading about this, Speed cameras as you say have very little to do with reducing accidents, but it makes policy look great, this applies to the horrid percentages for nasty products on TV that reduce wrinkles, piles, makes your shave closer, or your hair look less grey and you get a better job/partner    All these so called facts are never as they first seem, but people rarely look twice to make sure. Its a wealth producing train wreck running on a con.

[armadillo]

 All these so called facts are never as they first seem, but people rarely look twice to make sure. Its a wealth producing train wreck running on a con.

Aren't we a bunch of old cynics, eh?

[Rik]

Yes, it's a requirement here, Dill.

[armadillo]

http://www.symantec.com/content/en/us/enterprise/other_resources/b-state_of_spam_and_phishing_report_11-2010.en-us.pdf   

Good link! Interesting read. There does appear to be a genuine, if fairly small reduction. Though the large short term reduction is indeed based on a higher than normal rate in the preceding period.

The advice the article gives on reducing spam is very good.

I have for many years done what I think is one of the most powerful recommendations in the article. And I get absolutely no spam at all. I use multiple email addresses. I have about 20 of them. I use one for banks etc, another for friends, another for charities, another for on-line purchases from trustworthy vendors. Whenever I think I might get spammed, I create a new email address. If I get spam, I then delete the email address. The fact that idnet offer unlimited email addresses was a very strong factor in my choice to migrate to them (from another ISP that also did) and a reason against choice of other ISPs. I realise you can use third party email providers but it is more hassle and you never know how long they will stay in business.

Very simple. And I do not use any spam filtering and never have. All my idnet email addresses have filtering turned off.

[armadillo]

Yes, it's a requirement here, Dill.

I love this Groucho Marx one-liner.

    I sent the club a wire stating, Please accept my resignation. I don't want to belong to any club that will accept me as a member.
        Groucho Marx
       

[Rik]

 

He had a point.