Stop the Killings and the Fallacy of Selective Attention

Suppose one day you receive an email from a stock broker who offers to send you every week for four weeks a prediction of whether the Dow Jones Industrial Average will go UP or DOWN during the next week. Such predictions, if they were reliably accurate could be very lucrative since people can play the market long or short depending on what they believe the market will do in that following week.

Okay. Several years ago an email scam based on this idea was discovered in the US and broken up by the FBI. Quite a large number of people from a large data base of regular investors lost money to the scammers before they were busted. But here is what happened.

One day, Mr. Jones on the list receives an email from the scammers claiming that they will send him totally for free, four weeks worth of predictions of the Dow Jones. And they say that the Dow will go UP in the coming week. Mr. Jones ignores the email as another bit of spam, but he notices a week later that indeed the Dow has gone UP. He receives that notice along with a second email that claims in the second week, the Dow will go DOWN. Sure enough it does! Still two in a row is nothing, but when a third and fourth email arrive in the succeeding weeks and after he has received four successive correct predictions of the Dow Jones he is impressed enough to send in a requested $100 for future weekly predictions. Several thousand people were induced into doing so after they too received four successive correct predictions of whether the Dow would go up or down in the following week. I might have fallen for it too.

This really happened before the scammers were caught. But how did they do it? How did they convince thousands of people that they had predicted the Dow four straight times in a row?

Here is the answer. The scam was based on something called the Fallacy of Selective Counting.

The scammers had a database with a List of regular investors numbering several hundred thousands. In Week One, they sent half of the people on the List a prediction that the Dow would go UP in the next week; the other half of the people on the List got a prediction that the Dow would go DOWN. Thus, after the first week, half the List has lost interest because they got a wrong prediction. But in Week Two those who got a correct prediction in Week One get a second prediction: half of them that the Dow will go UP, half that it will go DOWN. Thus after Week Two a quarter of the original List have received two successive correct predictions of the Dow. They are feeling sorry they didn't take advantage of the free prediction and play the market long or short. In Week Three they receive a third prediction and sure enough, at the end of the third week, an eighth of the original List have received three successive correct predictions of the Dow! In Week Four one-sixteenth of the original List of several hundred thousand, have received four successive correct predictions. A large number of them sent in their hundred dollars and in the following week bet their life savings on the Dow and the fifth prediction they just received. Half of them doubled their life savings in the fifth week, but half of them lost their shirts.

I invite Philippine Commentary readers to apply the above analysis of a real-world fallacy to the issue of "Extrajudicial Killings" in the Philippines.

There is no doubt that a large number of murders have occurred. Are they related? Do they belong to a pattern because someone is responsible for most of them? Are all those being killed leftists, journalists and church people? What about soldiers, policemen, mayors and local officials, businessmen, farmers, store operators, and others who are not on the lists one normally sees publicized in the Media? The Catholic Church, the Philippine Military and police, have other lists that do not seem to get the same publicity from the Media and the Left.