My Favorite Fraud

By Larry Harris

The following scheme is illegal. Do not use it. I present it only so that you can better appreciate the sample selection bias, and so that you can avoid falling for this scheme should a man ever try to target you.

Buy a mailing list of 20,480 wealthy and gullible people. Send them a letter in which you explain that you are a very successful trader who has more money than you can spend. Boast about how you correctly predicted what direction prices would move during nine of the last ten months. Tell some fascinating story how you did it. Then tell them now you crave their recognition for being a market genius, since money is no longer very useful to you. Give them their prediction for the next month.

Divide your names into two equal groups. To first group, predict that the market will rise. To the second group, predict that it will fall.

One month later, write a letter to the 10,240 people for who you correctly predicted the market. Continue boasting and remind them that you correctly picked the market for them. Tell them how much money they could have made if they had traded on your recommendation. Then give them another pick. Of course, you predict an increase to one half and a decrease for the other half.

Repeat this for ten months until 20 people have seen you correctly predict the market 10 times in a row ( 20 = 20480/2^10). If they are not aware of your scheme, they will be convinced that you have correctly picked the market 10 times for them, and they will be prepared to believe that you did indeed do it correctly in nine of the ten months before you started to correspond with them. Such results are extremely unlikely to occur at random. They will swear that you are a genius.

You now have these people wrapped around your finger. If they are gullible you may be able to abuse their trust.

If you also send letters to people for whom you incorrectly predict he market, you can increase your yield. Simply tell them you are only human and cannot get it right every time. You will have 200 people for whom you have correctly called the market nine times and another 900 for whom you have correctly called the market exactly eight times. You probably will not want to continue corresponding with those for whom you failed to call the market three or more times.

The “lucky” ones for whom you correctly predicted the market cannot possibly make the proper inference about your skill without knowing your scheme.