Shortly after September 11, I wanted to get information about how the market had performed after major events that drove prices down significantly. I signed up for a trial from a site called MarketHistory.com. I found what I was looking for, and used it as a basis for the article "Our Resilient Economy." My conclusion was that, though it's impossible to compare this event to others with any exactitude, the market had generally recovered in time after major catastrophes.

While I think it's useful to consider historical precedent, I recognize that it can only do so much. It can indicate tendencies in the past, but it cannot predict the future. Disasters cannot be separated from their contexts. The pre-disaster state of the economy, the concurrent valuation of the market, and future corporate earnings in non-affected businesses are as important, if not much more important, for subsequent price movements.

It's not clear to me that MarketHistory.com understands this. Since registering, I have received its daily newsletter. It contained this passage last week:

"As of Thursday's close the Nasdaq Composite has gained 6.6% in the last 5 trading days.

Q: What happens when the Nasdaq gains more than 5% on the last 5 sessions on any Thursday in the fourth quarter?

A: This sort of fourth-quarter strength has happened on 16 prior occasions in the history of the Composite index, and whenever it has happened, we've seen Friday gains in 13 of the 16 cases. The average gain of the 13 winners was 1.5%.... Two weeks later, the 13 gainers posted an average return of another 5.0%. By three weeks later, the average of the 13 gainers was 7.2%, and by four weeks, it was up in the 13 cases by 9.0%."

Basically, the assertion is that past charts with any tangential relationship to present circumstances can help predict future charts. This is not a formula for success, ladies and gentlemen. In fact, it's downright silly. The market is not an Escher sketch (that link is great, but it's a slow connection). It does not run in infinite, symmetrical loops. No one but the data miners knows or cares what the heck happened on the fifth Tuesday in the third quarter with the sun in Sagittarius and a bad moon a-risin'. You can't replicate a time or event later, so that you know to buy or sell this time around. The result is that the observation is meaningless.

The trouble with "event-based" strategies is that it's impossible to recreate exactly the same scenario at another time. That's why technical analysis, at least in its pure chart-reading manifestation, isn't a convincing strategy to me. There may be "support" or "resistance" at some price or other at some time, based on general investor psychology, as the market generally "feels" that a stock is cheap or expensive given recent history.

There is little permanence to those observations, however, so there's no reason to think that they are valid considerations at any particular time. When a company misses its estimates by $0.50 a share, the market doesn't give a rat's patooty about its "support at 30." Neither, it inevitably turns out, do the chartists, who will tell you ex post facto that you were the stupid one for reading the chart wrong, or at least for not having a stop-loss in place.

Bill Mann said it well in January: "Quite simply, I have never seen any evidence that shows that any replicable technical scheme works for an extended period of time. When I say "works," I do not mean that every technical analyst loses money, but simply that technical analysis does not provide returns long-term above the S&P 500 at a level to justify the tax, commission, and time outlays that its adherents must shell out to play the game that way."

How does this apply to the Rule Breaker strategy?
If I'm going to attack historical technical analysis as unsuccessful strategies, then fairness dictates that I apply the same attitude toward the Rule Breaker strategy. It has an historical element of sorts that calls for a relative strength of 90 or above. Relative strength rates a stock against the general market in terms of price performance. A stock with a relative strength of 90 has essentially outperformed 90% of stocks in the market (depending on the method of computation and the index used for comparison). The theory is that such a stock, if it satisfies the other criteria, has a good chance of continuing its upward momentum.

Statisticians could spend the rest of their lives trying to prove or disprove that assertion, but I myself don't care about the results. To me, it is nonessential information. Whatever the incidental relative strength, the important questions to ask about a business before buying it are:

  • Is it in an important, enduring industry, so that I can confidently project future earnings?
  • Does the company have a sustainable competitive advantage?
  • Is its management of unquestionable honesty and integrity?
  • Is it priced such that a conservative estimate of future cash flows provides market-beating returns?

This is, essentially, the investment strategy that seems to have worked best, judging from the record of the great investors of our day. In fact, I cribbed it from super-investor Warren Buffett's concise explanation at the press conference following Berkshire Hathaway's (NYSE: BRK.A) annual meeting this year. The first three points also are a rough summation of Philip Fisher's 15 points.

They also relate to the Rule Breaker strategy, with two distinctions: 1) the Breaker looks for unproven companies, where it is difficult to project with any precision future earnings; and 2) the Breaker pays little attention to price. Number two springs from number one, since there can hardly be any estimate of price if there is no confident estimate of future earnings. That's why it's very risky.

So is the Rule Breaker strategy broken?
I've dismissed relative strength and the other stock criteria, and noted that the business criteria cannot help determine the proper price. Does that mean I think that the Rule Breaker strategy doesn't work?

I wish life were so simple that I could answer yes or no, but it's not. Investment strategies themselves don't "work" or "not work." To the woe of chartists -- and all others -- who try to create a machine that spits out winning stocks, like those machines in the back of comic books that spit out money, they don't work. The best a strategy can hope to do is to give you a framework for thinking about investing in the proper way.

Buffett's strategy doesn't work on its own -- it needs good information put into it and sensible conclusions taken out. The Rule Breaker strategy makes things tougher, because it omits some of the information in Buffett's model, but still requires the same good conclusions. Without any way to estimate intrinsic value, the strategy is much more likely than Buffett to miss it one way or the other. This is not a way to preserve capital, but it is a way to win big. That kind of speculation can easily lead to total loss of investment, so use it very sparingly.

Investing takes thoughtful, thorough analysis, and even then it is uncertain. There are no money machines in the equity markets.

Brian Lund is still searching for the strategy that will carry him to victory in every chess match. He owns shares of Berkshire Hathaway, and wouldn't want it any other way. The Motley Fool is investors writing for investors.