Ooblog has a cool post about the mathematics of a conceptually clever, if practically difficult scam. I’ll let you go there for the details and his nifty diagrams, but the short version is that you make a bunch of simple predictions about (roughly) 50/50 events to a group of people in a way that at least a few of them heard (almost) entirely correct predictions. Once you’ve convinced that small subset of your predictive prowess, you tell them that they’ll have to pay handsomely for any further predications.
A comment on the piece argues that this wouldn’t really work because you’d have to contend with things like spam filters, short attention spans, and the fact that most interesting predications of this sort (e.g., stock price movement) aren’t really 50/50, as the amount of change is as important as the direction. He also feels that you’d need to have plausible justifications for your predictions so they wouldn’t just sound like guesses.
My sense, though, is that the commenter may be giving the suckers too much credit. Remember that there are people out there that fall for Nigerian oil scams, and those are about as transparent as it comes. People also presumably fall for some pretty lame phishing scams, to judge by some of the junk that ends up in my mail spool. Making up bogus but vaguely plausible justifications would probably just be a matter of stringing together a bunch of market buzz words, and making your “successes” sound significant is presumably just a matter of spin. I’m sure that over 10 predications you could probably make claim like “If you’d invested $X according to my suggestions over the past two months, you’d have a return of Y%!!!”.
I think the hard part (which is raised in the comment) would be holding their attention over the 10 or so e-mails and have them remember that you’ve been so brilliant. The spin mechanism above might, however, be a way to continue to hold their attention. On all but the first “predication” e-mail you could crow about how much money they would have made if they’d been following your predictions. For some that might be enough to get them to fish through their trash for the old e-mails, or at least start paying attention to the subsequent ones.
All this does, however, indicate a problem with the original system, where you whittle a pool of 1,024 people down to a single sucker, who’s received 10 consecutive “correct” predictions. If you’re going for a single sucker, you absolutely need that one guy to stay on the line right to the end. If you assume a certain failure rate due to spam filters, short attention spans, basic understanding of the market, generally suspicious natures, etc., then you really need the higher return approaches (like going for the people that got at least 7 correct “predictions” out of 10) for any sort of reasonable return. You might even do better with the 7 or 8 out of 10 folks, since I’m sure at least some people will be very doubtful of a perfect track record.
Regardless, a cool model and some nice little bits of math :-).
No tag for this post.
Related posts
