Session 4.  What You Know Will Get You Into Trouble

This session is about the way the markets take a stream of partly predictable economic events and turn them into a string of almost entirely random price changes.  It thereby makes fools of us all when we try to use what we know in ways that would work well in other pursuits of life. 

Randomness and Statistics

What does random mean?  For our purpose we will take it as the kind of unpredictable event that you get when you flip a coin, or roll a pair of dice.  It doesn't mean that we know nothing about a collection of random events.  But we can't predict a specific outcome.  For example, when the faces of a die (a cube) are marked consecutively with the numbers 1 through 6, and we roll two such dice and add up the numbers on the topmost faces that result, we get a number ranging from 2 through 12.  In advance, we don't know which.  But if we repeat this trial many times, we begin to see a pattern, a distribution of random outcomes.  For example, the number 7 comes up much more often than does 2 or 12.

Take a look at the distribution of 1000 randomly generated outcomes of computerized dice.

We can describe the pattern with statistics.  Here are some of the most generally useful.  The mean (or average) is estimated by summing many outcomes and dividing by the number of trials.  The median is the outcome you discover if you sort trials from high to low outcomes, and pick the outcome in the middle.  Measures of dispersion are a bit more arcane.  The variance you estimate if you subtract from each outcome the overall mean, multiply the result by itself (square the differences from the mean), sum up all such products, and divide by the number of trials less one.  The standard deviation is the square root of the variance.  When multiplied by itself, it gives back the variance.  As we will see in later sessions, even though we can't predict each individual investment outcome, knowing these statistics about a random process that generates investment returns can help us develop better rules for dealing with it.

Generating Randomness From Human Interactions

How can we get randomness from people?  Do you remember the children's game of rock, paper and scissors?  Two children each hide a hand behind their back and form a fist, a flat hand, or two fingers sticking out.  On the count of three, they bring that hand to their front and see what the interaction says about who wins.  Rock smashes scissors, scissors cut paper, and paper covers rock.  A prize is exchanged (perhaps, if they are rowdy boys, the privilege of punching the other in the shoulder).  Then the game is repeated.  The overall winner is the one who can predict the other's choices without being predicted.  The outside observer sees a sequence of rocks, papers and scissors that becomes more and more random as the children become more expert.

A closer analog to the stock market is a similar game where one player simply tries to guess which hand of the other is clenched over a penny. If correct, he or she gets the penny.  If wrong, he or she pays a penny.  The sequence of lefts and rights quickly becomes highly random.  It will not alternate outcomes, because that would make it too easy to guess the next result.  Instead, there will be lots of two "lefts" in a row, and even a few runs of five or six or more "lefts" in a row.

Think of buying and selling a stock as like this game.  If you correctly guess a "right", you win because you bought before the price goes up.  If you correctly guess a "left", you win because you sold before the price goes down.  If you guess wrong, you lose money to the other players.

Stock Market Interactions and Return Statistics

Buying and selling real shares of ownership is fundamentally the same, but with some extra details.  It is not a zero-sum game, where if one player wins, the other must lose.  On average, the value of ownership goes up, because the underlying real economy grows.  Also, there are more than two possible outcomes each trial or time period.  Values may go up or down a lot or a little.  Also, in the real stock market one is facing not just one but many intelligent opponents all at once.  While not impossible, your chance of outwitting them collectively is much smaller.  Stock market movements are not dominated by the gradual working out of prices as various investors discover the news.  They are largely a series of jolts as surprises (news!) hit active traders who react nearly simultaneously and rather quickly.  To this extent, they are random, because a surprise is by definition unpredictable.

Keep in mind that this doesn't mean that there aren't possibly discoverable causes behind each price movement.  Their randomness simply means that you can't predict them individually.  The chart below shows a sample sequence of prices generated by the computerized equivalent of flipping a coin to see whether the price should move up or down by a fixed percentage for each of seven trading hours each day over a six month period.  Note how similar it is to a real stock price that you may have seen charted.

Let's pretend that we are reading the comments of simulated analysts over several months.  Initially, the price declines as bad news hits the market.  The sell-off is amplified by day-traders, but this peters out after the first month.  Then there is a period of consolidation, followed by a sharp rise as management announces a new contract for an important product.  The stock rises to a price of 100, where it encounters resistance, but continued buying sentiment breaks through, rewarding the faithful long-term investors.  This plausible interpretation is similar to what you can hear every day in the investment commentary.  It amplifies whatever is actually discovered as the cause to include extraneous rationalizations, the result of human beings looking for non-random patterns where causes are reliably apparent only in retrospect and often only in part.

How Your Knowledge Loses Value

Suppose you have discovered that a company has superior management, great products, and a lock on its well-defined and growing market. Will that help you earn a higher return by buying its stock?  Probably not.  Regrettably, unless you have nearly unique knowledge, others among the thousands or even millions of investors interested in the stock will have also figured it out, bought, and driven the price up to the point where further buying no longer offers an edge.  In a competitively-priced, liquid securities market, shared knowledge is worthless.

Most predictable change in price is predicted and acted on by many investors.  They move the price so quickly when easy-to-interpret news becomes public that only those with very fast reactions make any unusual profit.  There are exceptions.  Sometimes the market as a whole acts in concert.  Then it may be exploitable.  Sometimes there is a difficult concept that very few investors understand.  But most of the time, what we know doesn't help us because what we know is shared with too many others.

What remains useful is knowledge of the statistics of the random process.  Everyone can know something about means, medians, variance and standard deviation of results, and apply that knowledge to benefit their investment process, as we will see in later sessions.

The Role of Experts

Professional investors face the same problem you do, because they are competing to outwit a market whose prices are largely set by other professional investors.  This means that you are unlikely to have a financial advisor who on average has skill in predicting individual outcomes.  Even for the few that seem to exist, their skill level is likely to be sufficiently modest relative to the variance in the results of their efforts that you will probably not be able to identify them as true experts in any reasonable amount of time.  The cruel truth is that performance records that "experts" cite are mostly based on luck.  To believe otherwise is the "triumph of hope over experience."

Because of the inherent unreliability of past performance for predicting future performance, experts are nearly in the same position as the average investor with respect to picking the best stocks, or even knowing if they are in truth adding value or have just been lucky.  Brokers, financial planners and money managers can do many useful things for you, but you should not expect them to do the unlikely.

Owners may justifiably employ expert help that is based on statistical knowledge, or on important issues like how to keep taxes, fees and trading costs low.  If properly motivated, an expert can reliably help an educated owner with diversification, risk management, assessment of other experts, and customized financial planning. This can be of great value.

The key concepts for this session are not to trust your own predictions and not to expect the unlikely from experts.  Even if you ask a true expert to pick the best stocks, or to time the tops and bottoms of markets, you are asking him or her to outwit thousands of other experts operating collectively as an implicit team to keep that from happening.  This is a tall order, and you should not expect reliable results even under the best of circumstances.

[Next Session]