<COMMUNITY>
Post of the Day
September 17, 1999

From our
Mechanical Investing Folder

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light.


Subject:

Re: "Clusters": Last Stand?

Author: sparfarkle

"It's like predicting the chances of picking a red marble out of a mixed bag, then saying the chances are the same when you switch bags of marbles in the future. I think the screens have a good track record of picking out good stocks to buy, but the arguments/discussions of whether one model will return 41%/year every year in the conceivable future vs another model delivering 68%/yr are not useful"

This harkens back to the many discussions we've had out here on the predictive power of statistics. In the past, the discussions we've had on this topic have typically resulted in an obvious differentiation between how reliable statistics are in certain kinds of endeavors as opposed to others. You mention that what we're doing out here is "like predicting the chances of picking a red marble out of a mixed bag, then saying the chances are the same when you switch bags of marbles." The issue I have with this is whether you see picking businesses, based on their known fundamentals, as EXACTLY the same as picking one marble from a large set of marbles.

"...I would say picking the businesses based on known fundamentals can be profoundly different--from the perspective of defying random chance--than picking from the bag of marbles..."

Here is an extreme example of why I would say picking the businesses based on known fundamentals can be profoundly different--from the perspective of defying random chance--than picking from the bag of marbles: Let's imagine that a genie popped out of a bottle with the power to promise us that it can, for five minutes during each of the next 250 trading days, send one of the Dow Jones Industrial stocks to a share price of 10 cents. Let's say you and I make a little stock screen based on this, the rules for which read like this:

A)Buy any stock on a trading day that sells for 10 cents a share.
B) Hold in perpetuity.

Do you think our screen will do well, beating the market averages and beating what chance would imply are our hopes of beating the market? I think our screen will do phenomenally well! We'll get Coke at 10 cents a share, Wal-Mart, many more. These companies have growing profits in the billions each year. We will repeatedly and mechanically capture an enormous market inefficiency in pricing, and we will wait for the market to close that inefficiency. I wouldn't be surprised if this "genie screen" posted a multi-thousand percent CAGR every day! Now, we have only one problem: There is no genie. However, if we postulate that the people in the stock market, through fear, greed, and other emotional reactions, may consistently produce miniature versions of the large discrepancy the genie promised us (pricing inefficiency), we could use historical databases to see what kinds of approaches to identifying those pricing inefficiencies show a very, very, very high level of repeatability. That's what our screens are, and I say that's a lot different from picking one marble from many. All the marbles have the same form factor. None is Coca Cola Corporation, and none has the management, brands, and universal love that Coca Cola Corporation has. The marbles don't have different shapes, and they don't have constantly changing shapes.

"One can never disprove that chance may be behind the performance of the best screens we have..."

One can never disprove that chance may be behind the performance of the best screens we have, but the high level of repeatability, and the frequently yawning outperformance of the screens as compared to the market creates a very good argument that for chance to keep causing all that oft-repeating outperformance is unlikely. Warren Buffett, in this same argument, likes to cite the fact that he and about 20 other value investors learned quantitative investment rules from Ben Graham at Columbia University. After graduation, several decades later, they ALL had trounced the market, and had done so picking different stocks. Buffett argues that the intellectual common approach they were taught was the key to how they defied chance (you can find their multi-decade records in the back of The Intelligent Investor).

He argues that if a disparate group of 20 investors trounced the market for decades, that might be chance, but if we suddenly discovered that all 20 lived in the same house, we might suddenly get very suspicious of chance being the explanation. And he argues that the more years they all keep doing it, the more suspicious of chance we must become.

Sparfarkle