Is this famous investment scam real or fiction?

May 25, 2012 7:48 PM

A crooked stockbroker sends newsletters to 1000 people; half the newsletters say stock A will go up, half say it goes down. Stock A goes down. He sends 500 new newsletters to the people who got a correct prediction about stock A; half of the new newsletters say stock B is going up, half say it's going down. He then sends 250 more newsletters... etc. Eventually there's a handful of people who've gotten eight straight correct picks from him, and those suckers invest their life savings. Has this scam ever been carried out in real life, or is it just a story?

I've found the same story told all over the place; e.g. in the introduction to this paper, scam #1 on this page, and as the set-up for this blog post. And it's easy to find many more examples of the same story. But I can't find anything that refers to an actual case of this scam being carried out. Was it ever actually done in practice? If not, is there an origin point for this story, or is it just economics folklore with an origin lost in the distant past?

posted by escabeche to work & money (18 answers total)

It's in John Allen Paulos's popular book Innumeracy, published in 1988 (or 1990 by some sources?).
posted by LobsterMitten at 7:51 PM on May 25, 2012

OK, I have that book on my shelf. I just flipped through it to find the story. (Astonishingly, there's no index!) Story's there, in essentially the same form, but again with no indication that it's anything but a thought experiment.
posted by escabeche at 8:03 PM on May 25, 2012

Here's a Google Books scan of the part of Innumeracy describing the stock picking scam. This scan doesn't have page numbers; the story is under a subheading "A Stock Market Scam".

Here's a Google Books scan of pp 76ff of Paulos's later book A Mathematician Plays the Stock Market, with the story transposed to sports betting rather than stock picking.

(I don't know if this story was popularized by him, but figured I would link these here as a possible origin point for the wide awareness of this story. I'm very interested to know if this has happened in real life.)
posted by LobsterMitten at 8:04 PM on May 25, 2012

Ok, one more link to the same thing just for posterity: A Stock Market Scam, the short passage from Innumeracy, posted at Paulos's own website in case Google Books blocks the link above.
posted by LobsterMitten at 8:08 PM on May 25, 2012

This is a little side track, but I'm pretty sure Derren Brown did something like this in The System, where he correctly predicted the winning horse for 5 consecutive races. Video explanation here....obviously the show is to entertain, so believe what you want.
posted by Mr. Papagiorgio at 8:47 PM on May 25, 2012 [3 favorites]

I do not know the origin of this story. I am reminded, however, that it (or a close variant) has been mentioned by a comic dinosaur. That was actually the first time I ever ran across that scam.
posted by klausman at 8:50 PM on May 25, 2012

I remember a version of this scam from Mathnet, a segment from an old PBS kids' show. Looks like the episode was "The Case of the Swami Scam" and it was filmed either around the same time as or slightly after Innumeracy came out.
posted by pengale at 9:58 PM on May 25, 2012 [3 favorites]

This is how most of the football pick boiler rooms work.
posted by Marky at 10:07 PM on May 25, 2012

I could have sworn that there was a Sherlock story about this, only with predicting horse races. But I can't seem to find it anywhere.
posted by benito.strauss at 10:34 PM on May 25, 2012

I have seen it done to test the idea, but never seriously.

In practice the list would have to be enormous. Say you wanted to make 100,000 / year and you're selling your fraud at $2000 each. You would have to sell 50 copies (ignore expenses.) If you want to send eight weeks of predictions, right there you need at least 12,800 people on your list every year because 255/256 of them are going to get a bad pick. But, you won't get a 100% conversion rate even after eight weeks. Even with the scam being so persuasive, you probably won't pick up more than 10% if you're a good copywriter. So, you need to add 128,000 people to your list every year and throw out 127,950 of them. There really aren't that many interested, qualified people, and that's a waste of hard-won addresses/e-mails that you probably had to buy or harvest through Adwords.

So, the preferred way to do this is to always be right. That's done by talking about past successes, fabricated if necessary. This is just as effective for most people because the important part is where the e-mail says, "last week we predicted XYZ would rise 17%. It actually went up 28%!" That's written whether the user was informed of the "prediction" last week or not and it's all most people remember. If for some reason it's necessary to make a prediction, making hundreds of predictions and just printing the winners is also effective, unfortunately. Most importantly, even if it's not very effective, the entire list can be used for every communication.
posted by michaelh at 4:46 AM on May 26, 2012 [1 favorite]

Er, throw out 127,500 of them. Please forgive any other math errors as the main point stands.
posted by michaelh at 4:47 AM on May 26, 2012

If not, is there an origin point for this story, or is it just economics folklore with an origin lost in the distant past?

We can certainly date the idea earlier than 1988's Innumeracy; the scam shows up in a 1957 episode of Alfred Hitchcock Presents entitled The Mail Order Prophet. In this case, the correspondent is predicting not just stock prices but all kinds of outcomes, including elections and sporting events, but the principle is the same: send out predictive letters to a number of marks, winnow out the half that "lose" in each round, eventually end up with one mark who believes your predictions to be infallible, and ask for an investment.
posted by Elsa at 8:09 AM on May 26, 2012 [1 favorite]

A crooked stockbroker sends newsletters to 1000 people; half the newsletters say stock A will go up, half say it goes down.

990 of the recipients will throw the letter into the garbage without even opening it.
posted by sour cream at 8:58 AM on May 26, 2012 [2 favorites]

scam real or fiction?

No, this is how you start a hedge fund. Several accounts with different "trading strategies". Then you go with the one which was "successful" to investors and tell them "See, what we can do?"
posted by yoyo_nyc at 12:47 PM on May 26, 2012 [4 favorites]

I learned about this scam from Innumeracy, which is probably the book that I learned the most math from in my life. And I use it in my classes, whether it's true or not. (Paulos might know. I'll point him here.)
posted by madcaptenor at 1:00 PM on May 26, 2012

Paulos writes on twitter, in response to my asking if he knew the history: "No, but I remember seeing a minimalist version of it in one of Daniel Dennett's early books."
posted by madcaptenor at 2:42 PM on May 26, 2012 [2 favorites]

No, this is how you start a hedge fund. Several accounts with different "trading strategies". Then you go with the one which was "successful" to investors and tell them "See, what we can do?"

I was going to mention this as being related also. Specifically you fold the fund with the poorer record into the fund with the superior record, but you advertise the better return instead of calculating the average of the combined return. In addition to scrubbing return rates it also resets the high water mark that factors into performance-based pay. It happens in mutual funds as well though not as much.
posted by michaelh at 2:44 PM on May 26, 2012

No, this is how you start a hedge fund. Several accounts with different "trading strategies". Then you go with the one which was "successful" to investors and tell them "See, what we can do?"

The academic paper I linked is about this precise strategy, which is the context in which it recounts the story.
posted by escabeche at 7:48 PM on May 26, 2012