©2006 All content on site protected by copyright
Nobody Asked Me, But... from Victor Niederhoffer
Nobody asked me, but one of the commonest errors in trading is to assume that you have perfect knowledge of an event. Such an error is behind all the mumbo jumbo about the end of the Fed cycle being bearish. One does not know what and when the Fed is going to do. Often their moves are anticipated by big market rises. The dating of their announcements has always been problematic, and the tendency for anticipation has changed drastically. In the bad old days, the market would invariably rise on the morning of the day they eased. With about 15 changes in direction of easing and tightening over a 100 year period, the opportunity for multiple comparisons is 100%, and any hypothesis can be supported if you weasel enough with Fed Funds, Discount rates, time between announcements, etc.
Along this line, the book Meta-Analysis by the Confidence Profile Method by David Eddy, Vic Hasselblad and Ross Shacter gives methods and programs for the synthesis of hypotheses using general and easily understandable methods. It points out that there are problems that scientists face such as multiple pieces of evidence, different experimental designs, different types of outcomes, different measures of effect, biases to internal validity, biases in comparability , indirect evidence that must be considered, mixed comparisons, etc. Their methods are designed to interpret, adjust and combine all such problems in each experiments.
The problems that quant investors face are similar to those that the three Doctors face. There are conflicting patterns, different conflicting relationships, overriding considerations relating to drift and Fed moves, cycles relating to seasonal factors, drift relating to the difference between bonds and stocks et al. These factors must all be combined and the confidence profile method is a good way to do so.
Finally, I remember the chapter in the Monthly Illustrator of June 1883 on Tips and tipsters.
There are many men who haunt brokers officers without any apparent purpose. They are well dressed and appear to live well.. Some of these have confederates in various parts of the world to give early information. 'Are the employees of a railroad talking about striking?' 'Is Congress going to approve a committee to investigate some trust?' This information is conveyed at once to the tipster and by time it is made public he has arranged a deal. It is almost impossible to carry through any important deal in the Street in perfect secrecy.
It was always the custom to buy a tipster lunch in case his tip were to pan out, although such a deal was not enforceable by the courts, as "they have held that his contracts are mere gambling agreements."
The next time you receive a tip of impending doom, and it happens to be that one in three time that the market does spend a quarter down, remember that the lunch is not enforceable.