|  NOV 1991

Merrill MW Waves by Arthur A. Merrill, C.M.T.

Merrill MW Waves by Arthur A. Merrill, C.M.T. Suppose prices have been moving in a certain way. What are the probabilities for future prices? You must look to the past for similar situations. How similar must the past situation be? If you ask your computer to search a historical database for an exact copy of a current price swing, it will have a difficult time. Suppose it succeeds in finding a duplicate situation, and you see that the market rose after the past price swing. Does this mean that now we can count on a sure rise, with 100% probability? Of course not. The evidence is insufficient; it's called anecdotal evidence. If a coin turns up heads one time, it isn't a sure probability for another heads next time. To assess the true probability of possible outcomes for the current price pattern, you can't ask the computer for an exact duplication of the pattern. The situation must be similar, but not exact. Your computer must be able to find enough occasions to indicate a bias in the situation, and it needs enough occasions to formulate a probability for the future. For example, if prices rose 40 times and declined five times after a similar price pattern in the past, the probabilities for the current price swing would be 40/45, or 89%. Of course, the past moves may have been the result of simple chance. To check this, you can use a statistical test called chi square. You'll find it described in any statistical textbook.

by Arthur A. Merrill, C.M.T.