Wednesday, August 24, 2016

Sporadic/Intermittent Reinforcement

John Huber has an excellent post on the 1977 WSJ article on Warren Buffett I linked to a couple of days ago. I wanted to highlight a couple of paragraphs here about his mention of value traps because it ties in with a mental model from psychology (Sporadic/Intermittent Reinforcement) that I've seen in a couple of places. From John's post:
Whenever I review my own investment mistakes, they almost always come from situations where I was attracted much more to the valuation than to the business. These are the so-called value-traps—catnip to most value investors, but very often poor choices as investment candidates. I’ve learned to steer clear. 
The problem for many investors is that sometimes these so-called value traps work out. You’re able to buy them and sell them for a 50% gain. But a year or two later they are often trading at or below (often well below) your original purchase price. The investment looked smart based on the realized gain, but was it really being smart, or just fortunate timing?
The key thing I want to highlight above is John's mention of the problem being "that sometimes these so-called value traps work out." This is similar to a point made by Mark Yusko in the Morgan Creek Q2 Letter in a discussion about Seth Klarman and his thoughts on why the average mutual fund investor vastly underperforms the average mutual fund. The average investor in mutual funds changes funds too often, chases what is hot, etc. And the problem here, again, is that sometimes it works, at least for a while. As Yusko wrote: 
Klarman's response to this phenomenon is, "while no one wishes to incur losses, you couldn't prove it from an examination of the behavior of most investors. The speculative urge that lies within most of us is strong; the prospect of a free lunch can be compelling, especially when others have already seemingly partaken." That speculative urge is a psychological characteristic in all of us that we must fight in order to reach our full potential as great investors. B.F. Skinner did a great deal of work on trying to discern why human beings seemed hard-wired to want to speculate (gamble), and found that the behavior was linked to a concept called "sporadic reinforcement." In essence, by winning only occasionally, the desire to participate in that activity actually increases.
And to tie this idea back to some source material from B.F. Skinner himself, here are a couple of paragraphs from his book About Behaviorism:
All gambling systems are based on variable-ratio schedules, although their effects are usually attributed to feelings. It is frequently said, for example, that people gamble because of the excitement, but the excitement is clearly a collateral product. It is also sometimes said that people gamble "to satisfy their sense of mastery, to dominate, to win" — in spite of the fact that gamblers almost always eventually lose. The inconsistency is explained by calling the gambler who ruins himself and his family "compulsive" or "pathological," his "irrational" behavior thus being attributed to an illness. His behavior is "abnormal" in the sense that not everyone responds with similar dedication to the prevailing contingencies, but the fact is simply that not everyone has been exposed to a program through which a highly unfavorable ratio is made effective. The same variable ratio schedule affects those who explore, prospect, invent, conduct scientific research, and compose works of art, music, or literature, and in these fields a high level of activity is usually attributed to dedication rather than compulsion or irrationality.  
It is characteristic of intermittent reinforcement that behavior may be sustained over long periods of time with very little return. This has been explained by saying "Human beings are creatures of hope and are not genetically designed to resign themselves," but there is nothing essentially human about the effects, and it is not hope or resignation but the contingencies which are the conspicuous and accessible cause.