A (MLM) Skeptic: How to Fake Being Infallible (and how to spot such fakes)

Wednesday, May 29, 2013

How to Fake Being Infallible (and how to spot such fakes)

Derren Brown at the Garrick Theatre, June 2008 (Photo credit: Wikipedia)

Ever run into some guy who claims they invented a system that will do absolutely great in the stock market or foreign exchange market or commodities market or whatever market that involves trading, and he'll reveal the secrets to you for a price? You'll probably dismiss him as crackpot. He offers to sell you the secret, but only for 72 hours (3 days).

What if he mailed you a prediction, and next day, he turned out to be right?

And the next day?

And the next day?

Do you feel a need to buy his system now? As he's withdrawing his offer?

Don't. This is easily faked. And I'll explain why it's fake.

Darren Brown is a well known magician who did a video in 2008 called "The System", where he demonstrated on camera, one continuous cut with 2 angles, that he managed to flip a coin 10 times in a row, and they ALL turned up "heads". This is not done with a trick coin or camera trick or post-production.

Sounds impressive, yes?

What they didn't show you (unless you watched the full episode, which is not in the video), is that he filmed over 1000 coin tosses, over period of 9 HOURS, to create this "10 heads in a row" video.

Not so impressive now, is it?

These "infallible trading system" purveyors have a similar trick.

They offer you a system that they claim offers an infallible signal for trading, probably stocks. And just to show you it works, they'll predict that a particular stock will go up or down the next day. They'll even show you some blah blah screen that says it does predict so (just not the logic behind it). And it's a limited time offer. The price starts REALLY low, goes up every day until at the 4th day it stops being available.

And they send you proof every day that their system works. And it's easily verified that the stock went up or down as they predicted.

So you're feeling that maybe they really did create such a system. Evidence don't lie, right?

No, but your INTERPRETATION of the evidence is wrong. This is how they did it.

Start with a poll of, say 10000 potential people who may be interested in such a thing.

To half, they predicted stock will go up, and the other half, stock will go down.

Next day, they stop sending email to the half that was 'wrong'. And they repeat the process.

Next day, they do the same thing AGAIN.

Now they have 1250 people who believed that this bogus system worked 3 days in a row. That's uncanny accuracy, but only if you're not aware of the overall campaign of 10000. Then you realize it's just randomness and probability at work.