Wednesday, March 21, 2012

Schulz, The Intelligent Chartist, part 3

Three themes for my final post on this book (and I will have dealt with only 43 pages in a 275-page book), the first one being “chart playing” vs. “chart reading.”

Writing about trading ranges, Schulz notes that “Chart trading ‘methods’ that dispense with critical analysis and therefore lean heavily on lagging indicators – that is, on ‘signals’ – are at a special disadvantage in these situations.”

. . .

“At this point, I should like to differentiate more clearly between what I have been referring to as ‘chart playing’ – that is, crude speculation responding uncritically to so-called chart signals – and what can more properly be described as ‘chart reading.’ … In my view, chart reading begins when automated response to signals is rejected as the only or predominant feature of the ‘method.’ From this rejection it necessarily moves toward recognition of earlier indicators and, in the process, is constrained to become more discriminating, if only because the leading indicators are inevitably more equivocal than the lagging indicators. The search for leading indicators and the need to interpret them involves a measure of critical and analytical reasoning. In its most fully developed form, chart reading searches for the earliest possible indications of trend changes; it can then be appropriately described as chart analysis. If the process is extended to include critical examination of other technical evidence, it can qualify for the broader classification of ‘technical analysis.’ With every step toward greater critical refinement, the process moves further away from being a method virtually devoid of reasoning and closer to the ideal of pure analytical reasoning, internally consistent and taking account of every available item of technical evidence, and designed to produce a fully integrated view of the price movement in all trend perspectives.”

Second, pattern recognition.

“[I]t is necessary to realize that pattern recognition lends itself very well to popularization because it calls for visual comparisons rather than for the mental effort of careful analytical reasoning. (Bar charts are frequently offered under the motto ‘a picture is worth a thousand words’; under many conditions, in my opinion, it can be worth less than silence.) Pattern recognition is also an especially tempting technique because its emphasis on numerous historical models leads the chart reader to assume, if but subconsciously, that there is something of the inevitable about the relationship between chart pattern and subsequent price performance.”

Finally, and critically important, historical models. Here I will quote more extensively. I call attention to Schulz’s claim that fundamental analysis relies more heavily on the notion of the past as prologue than does technical analysis—essentially a defense against one of the most frequently advanced criticisms of technical analysis.

“At bottom, reliance on historical models must assume that earlier cause-and-effect relationships are being and will again be repeated. But, certainly in the stock market world, it is as a rule very difficult – if it is at all possible – to assign specific effects to specific causes in the present with any kind of assurance that the connection being made is a correct one. The attempt to relate effects and causes is constantly being made by professional and amateur alike; for current conditions, it is called explanation; for future conditions, it is called forecasting; in either case, it meets with a not very remarkable lack of success.

“To establish the proper cause-and-effect relationship for a past stock market event, selected to serve as a model, should be more nearly feasible, provided an adequate labor of research is performed. But even if the relationship of the past is satisfactorily demonstrated, the much more difficult task of establishing the similarity with current conditions still remains. It should be apparent, therefore, that insistence on multiple precedents (and expectation of multiple repetitions) aggravates the problem instead of easing it. Most likely, the problem has no workable solution since the necessary amount of research may conceivably exceed in cost the tangible benefits of the result.

“It is, however, possible to argue that the problem is, in practice, more serious for fundamental analysis than it is for technical analysis. … [In technical analysis the problem] ought to be more readily coped with because it exists within a more limited focus. Technical analysis functions on the premise that cause-and-effect relationships in the stock market can be most practicably investigated at the level of supply-demand shifts as the most direct cause of price and trend behavior; and that these shifts can be best – if imperfectly – appraised through analysis of price and trend changes themselves. This part of technical rationale can and should be qualified and refined. But certainly the view of supply-demand changes as the least equivocal of cause-and-effect relationships is tenable.”

. . .

“From this viewpoint, the essential function of technical analysis is that of accounting for current and historical price movement; and, in my opinion, that is its proper and most fruitful function. The accounting is done primarily in terms of trend development, and its emphasis should rest much less on historical parallel than on the historical evolution of trends. In other words, it assumes that past supply-demand relationships have a bearing on subsequent supply-demand relationships, and that this bearing arises more clearly from the internal logic of the subject than from a process of repetition.

“The basic premise for this assumption is the view that the movement of stock prices is a coherent process; and that it is an orderly process in the sense that small price fluctuations link up to form short trends, that short trends link up to form trends of a larger order, that these in turn are the components of still larger trend categories, and so on. It follows logically from this assumption that the several trend categories are interdependent and interactive at every point in the price movement, and that technical analysis should strive to allow for this mutuality in accounting for price behavior, past and present.”

No comments:

Post a Comment