This is a true story, and not designed to make you hate me.
A couple of years a go I got an e-mail from a very bright young guy — a graduate student with the kind of clarity, gumption, technical skills and instinct that I might have had at his age if I had been a little brighter and had ever managed to master higher math.
(And also if, someplace deep down, I didn’t feel a little guilty making money too easily. Like anyone else, I want easy money; but on some deeply neurotic level, I feel as if I need to suffer for it.)
This young man, in short, had the tools and the passion, and — through the expenditure of a tremendous amount of hard work — had built a proprietary “model” that, with an impressive degree of confidence, predicts swings in the market.
We had actually corresponded on a couple of other topics prior to this — he had invented a device to keep plastic soda bottles from going flat after you opened them — and I was fairly blown away by the way he handled himself. This guy was 23? A 40-year-old should be so professional.
(The soda bottle device fizzled, but that’s neither here nor there.)
Anyway, over a period of a few e-mails, he described his market-beating model, which he had been putting to good use for friends and family, and on which he was basing a nascent market newsletter.
“Oh, please,” I wrote back. “You cannot predict swings in the market.”
“Can, too,” he wrote back.
I am giving you the executive summary. My e-mails were actually long and dismissive. His were longer, respectful, and brilliant — to the limited extent I could understand them.
And the net of it all is that I figured — never dreaming it could make me $1 million — I would give him a chance to “prove” it to me with real money.
I figured, in the first place, that, heck, he might even be able to do it. Second, that if he failed, I’d have the great satisfaction of being right. (The three sweetest words in the English language: “You were right.”) And third, this was a way to give a bright young kid a boost.
So I authorized him to lose as much of my money as he wanted trading puts and calls on the market, up to a strict limit of $20,000.
Part of me wanted to exploit this young man’s talents to make me rich. Part of me wanted him to fail dismally, to affirm my view of the impossibility of timing the market, no matter how smart you are. My ego exceeding even my avarice, I can’t say I wasn’t actually, on balance, rooting for myself to lose money.
I began to get e-mails from him and his wife like this one:
Date: 09/24/1999 2:23:09 PM Eastern Standard Time
Andy, in spite of our deep fear of this market, I placed a call trade for you near the close yesterday. I placed it using our model which evaluates calls assuming equal volatilities in all strike prices. It turns out that the volatilities are much lower right now on the out-of-the-money calls, so today we sold the first position for a $1400 loss and opened a very sizeable and risky position in the Oct 700 calls, which are a better deal when you combine the model’s historical results with the present situation in which in-the-money calls are disproportionately expensive. I don’t have a confirmation yet, but the trade was to buy 15 Oct 700 OEX calls @ 5 1/4 or better.
At first, I was smugly relieved to see, the results were anything but spectacular. Here is another actual e-mail a few weeks later — blanking out only the name of his newsletter lest I appear to be touting it:
Date: 11/05/1999 7:46:45 PM Eastern Standard Time
Andy, our [market newsletter] is doing very well, and our model is up 11.7% for the prospective period (since 7/15) vs. -2.8% for the S&P 500. I am disappointed and frustrated, however, with our performance in your account, in which we have lost greater than $1,800 trading options.
In spite of our best attempts not to do so, we have second-guessed ourselves when contemplating options trades for your account. We did not buy calls for you on our last buy signal, for example, because our signal’s waveform wasn’t exactly perfect, and we were shaken by losing money on our last trade in your account.
If it is acceptable to you, I would like to try something that requires a little less thought. Specifically, when the model is long, I’d like to be 200% long S&P 500 depositary receipts (i.e. 50% margin), and when the model is short, I’d like to be 100% short SPY (no margin). This will leave less room for any brains/fear to get in the way between our model and your account.
I hope that this does not degrade your confidence in our model. If the model’s current success continues, we will very likely be seeking financing to expand our business in the early part of 2000, and it would be very nice to have your support.
I allowed as to how I had never had much faith in the model to begin with, and that so long as I didn’t lose more than $20,000, he was free to do whatever he wanted.
Little did I dream that within a few short months he’d have run my stake up to nearly $1 million!
Little did I dream it, and little did he do it.
Date: 04/20/2000 11:18:57 PM Eastern Standard Time
Andy, things haven’t turned out as well as we had hoped, to say the least. We are very close to being down $20,000, close enough that any further trading would probably be a desperate act.
Our beloved model, which so soundly beat the S&P from July through December, has been mostly dormant this year, occasionally stirring to give us a piece or two of really bad advice and then going back to sleep. We’re down 17% YTD, and about 9.5% worse than the S&P; we have been stopped out on both of the last two signals in our futures trading. Our performance in your account has also been affected by my own unjustifiable optimism in the model, such that I have not consistently sold to cut our losses when I should have, for which I apologize.
At this point there are only a handful of possibilities concerning our model:
1. Our good returns were due to random chance, and the model never worked at all.
2. Our model worked for a while, but doesn’t anymore, because
a. the nature of one/both of our variables has changed
b. different psychological patterns are emerging
c. it was calibrated to a strong bull market / lookback error / etc.
d. Things Fall Apart.
3. Our model STILL works, and the long time between recent signals interrupted by completely WRONG ones indicates a period in which investors are not acting in a uniform fashion as they usually do, etc.
We are hoping, of course, that the answer is closest to #3, but only time will tell, and even if we again soundly beat the market for the year, I doubt many subscribers will be lining up to renew for such an unreliable system. My wife and I are probably going to turn our fledgling enterprise into a standard retirement planning/basic asset allocation sort of firm for at least long enough to recoup our losses.
We still believe that if it is possible to time the market, it MUST be by this kind of psychology-based approach. We recently did a thorough analysis of the interrelation and persistence of numerous short- to long-term trends since 1950, and found, in a nutshell, that while short trends persist, medium ones don’t, and long trends persist, but exactly which trend one should be paying attention to shifts from month to month and year to year such that any resulting system would require a great deal of daily interpretation and “intuition.” It’s no wonder the promise of a purely mathematical model was so enchanting.
I’m sorry that the pie in the sky turned out to be plastic sushi — very attractive until you try to bite into it, and then you’re out a couple of teeth.
In short, things could hardly have worked out better. I had heard those three sweet, sweet English words — “you were right” — and it hadn’t even cost me the full $20,000. There was about $5,000 left.
I view this $15,000 as one of the routine costs of running this site on your behalf, to help spare you the folly of trying for easy money with some brilliant market-timing scheme.
I began by saying this column was not designed to make you hate me. And, yes, I did that in part to build a little suspense into the tale . . . but also because I fear that many of you do hate someone who ends every story, like this one, with boring, slow-but-steady, no-free-lunch, live-beneath-your-means advice.
The good news is that slow but steady does win the race, and that this is why, having steadily contributed “the max” to my Keogh Plan in the 70s and 80s, there’s enough left over to keep this $15,000 hit from hurting.
Tomorrow: A Novel Mutual Fund Fee Structure