Monday, September 19, 2011

Carlson, George Lindsay and the Art of Technical Analysis

Investors often look to technicians for signs that a market is either about to roll over or that a current downtrend is not simply a pullback but the beginning of a longer-term bear market. Among the technical patterns that portend doom are the direly-named Hindenburg Omen and the Death Cross. The Three Peaks and a Domed House pattern may not sound as menacing as the other two, but it too signals a severe market decline—at least when it occurs in the chart of the Dow Jones Industrial Average.*

In George Lindsay and the Art of Technical Analysis (FT Press, 2011) Ed Carlson introduces the reader to the “seemingly bizarre” discoverer of this pattern, as Louis Rukeyser described Lindsay. Among other things, he wore a bright red toupee, and his last face-lift “left him a bit strange in appearance as it pushed up his eyebrows so he looked perpetually surprised….” (p. 14) Lindsay had an advisory/forecasting service and wrote a weekly investment letter; he did not trade for his own account. Apparently many of his forecasts were spot on.

Lindsay’s sole book was The Other History, which he self-published. It was an attempt to describe temporal patterns in international events, what he called technical history. His technical studies of the stock market appeared only in his newsletters.

Lindsay made a bold claim for his most famous reversal pattern—that it “could be found at 60% of bull market tops and at the peaks of rallies in bear markets (cyclical bull markets).” (p. 41) (Thomas Bulkowski didn’t include the pattern in his Encyclopedia of Chart Patterns because he said he couldn’t find enough samples.) Lindsay, who introduced the Three Peaks and a Domed House concept in 1968, said he found inspiration in two patterns which began in 1893 and 1910.

In its idealized form it looks like this:

Carlson spends a great deal of time describing this pattern and its variations, such as the domed house coming before the three peaks.

He also explains how Lindsay calculated how far markets might fall after this formation appeared.

In general, only three points are needed for the calculation: F, G, and N. If(N-G)/(F-G)—the so-called swingover ratio—is less than 2, the calculated number is used as a multiplier; if it is 2 or greater, the multiplier is simply 2 (with one exception, not worth going into here). The number of points that the market is expected to fall from point N is [(N-G) x multiplier] – (F-G). Lindsay himself admits that not all three peak patterns can be used for this calculation. “Market history shows that a formation can be discarded (1) when it is supplanted by another pattern which precedes or follows it, (2) when it is short or imperfectly formed, or (3) when it occurs at a very low level, historically, in the average.” (p. 89)

Lindsay also had a triangulation timing model that was comprised of three elements: the 107-day top-to-top interval, the low-to-low-to-high interval, and the convergence of these two intervals, which gives a targeted top or high. (p. 95)

Lindsay’s work is complicated, which may be a virtue or a vice. I have described only a few of the timing models and observations that Carlson carefully analyzes.

I will leave my readers with one last takeaway from Lindsay: “I am amazed because few technicians recognize that the length of time that market movements last has always been much more nearly uniform than the number of points which the averages gain or lose. Perhaps it is because, at least in my version of it, I sometimes start counting from secondary highs and lows. But such counts are made comparatively seldom.” (p. 174) Readers who are interested in cyclical market analysis should profit from reading about Lindsay’s variations on the theme.

*Lindsay wrote: “Averages composed of a small number of blue chips have always had crisper chart patterns than all-inclusive indexes. It is largely because unseasoned stocks are in a state of flux: new ones are being added, old ones are dropped, and the number of shares is constantly changing. The Dow Jones stocks are more stable in composition. Talk of the Dow Jones Average as being unrepresentative is beside the mark. If you want to know the true level of ‘the market,’ look at the broader averages. If you want to predict the future, go by the Dow or the New York Times Industrials. Indeed, some technicians get the most reliable results by using an index of only ten or twelve sensitive and influential stocks. The NYSE Index of all stocks is nearly worthless in forecasting.” (p. 40)

No comments:

Post a Comment