Daily Speculations

Survival Statistics

Posted 10:30 p.m. Friday, Jan. 23, 2004

One of the most helpful ways of looking at markets, and predictably one almost completely overlooked, is in terms of survival statistics. Survival time, or
lifetime, may be defined as the time to occurrence of a given event. It's usually used for things like time to death, relapse, failure or accidents. It can
be extended with ease to such things as the time it takes for declines of a certain size to occur, or the waiting time between market declines of, say, 2%, or market rises of 10%.

Many survival time distributions are best modeled as survivals with immunes. After a certain period, all the susceptible elements of the population have died, and the remaining are relatively immune. Such is often the case in the market.

We wrote in February 2003 that the market had not had a 10-point decline in the S&P in 42 trading days, since Nov. 18, 2002. You might think this is a relatively bearish event. But no, the facts show that when a long time has occurred without such a decline, the remaining " stock market moves" are relatively immune to declines.

The expected move until the next such decline when a waiting time of 40 or more days has occurred is relatively bullish. Such are the insights that
can come from looking at markets this way. And such is the way that technical analysis might be augmented.

Vic