Lindsay's Theory of Middle Sections

I can think of no better way to begin this post than with the following 1974 quote from George Lindsay:

"The first original idea I ever had on the stock market remains the best. In 1950 I published a copyrighted pamphlet "An Aid to Timing" which introduced the concept of the "Middle Section". The pamphlet sold well and I received so many letters that I was encouraged to start a market letter of my own the following year on a capital of $6oo. Since I have never been a big advertiser, it is remarkable that I have lasted in the business for 23 years. I could never have done it without this method. [.....] In all the years since then, I have mentioned the principle [of the Middle Section] only once in my advisory letter. [.....]. Counts from the Middle Section [....] are my prize way of calculating time in the market. "

It is not my intention in this post to explain in detail Lindsay's theory of the Middle Section. You can find all the information you might want about it in his article "Counts from the Middle Section" which has been reprinted in the booklet "Selected Articles by the late George Lindsay". This booklet is available directly from Investors Intelligence (and NOT from their online book store).

The basic idea of the Middle Section is a kind of symmetry principle. The Middle Section is an interruption of a bull market advance. Generally the Middle Section shows up as a sequence of jagged rallies and declines during which the market makes only minimal upside progress. Such a pattern generally lasts 20 weeks or so and is called an ascending middle section by Lindsay.

The first chart above this post is a schematic which depicts an ideally shaped ascending middle section. The middle section itself is shown in red. The advance from point A to point J is a Lindsay basic advance which typically coincides with a bull market. The drop from point J to point AA is a Lindsay basic decline which typically coincides with a bear market. The advance from point AA to JJ is a second basic advance which would typically coincide with a new bull market. The whole schematic is meant to depict market action over a period of 5 years or so, sometimes less, sometimes more.

In an ascending middle section one begins a time count typically from red point E (although sometime point C can be used). Counting forward from E to point J, the bull market top and then counting forward the same number of days from point J should predict the date of point AA, the subsequent bear market low. Alternatively, counting forward from point E to point to point AA and then counting forward from point AA the same number of days should predict the date of point JJ, the subsequent bull market high.

This all sounds very simple (and hence implausible) but of course there are generally complications. First of all, a basic advance like A to J often has more than one plausible middle section. Secondly, it is unusual for the two counts described above from the same middle section to both make accurate predictions. It is more common for one middle section to predict the timing of point AA and a different middle section to predict the timing of point JJ. Finally, even within the same middle section it is sometimes difficult to determine which of starting points E or C to use.

There is no hard and fast rule for sorting out these complications. But Lindsay was trained as an artist, not as a statistician. And his method was to look for the beautiful picture that results when his several distinct methods (of which the Middle Section was one) agree on the timing of a high or low.


The second chart above this post shows two counts from the Middle Section which I think Lindsay would make himself. There was a basic advance in the Dow from point A on October 10, 2002 to point J on April 6, 2004. The ascending middle section encompasses the choppy action from point B on June 17, 2003 to point H on September 30, 2003. This 15 week period was a little less than the ideal 20 week length of a typical ascending middle section.

Point E according to my calculation occurred on September 19, 2003. Measuring forward from E to point J gives and interval of 200 calendar days. Counting 200 calendar days forward from point J on April 6, 2004 gives October 23, 2004 as the projected low of a basic decline. The actual low occurred two days later on October 25 at point AA.

To project point JJ I again start from point E but this time count forward to the alternative ending point of the basic decline from point J. This is point AAA on April 20, 2005. The time from E to AAA is 579 calenday days. Counting 579 days forward from AAA gives a projected top at point JJ on November 20, 2006.

This November 20 forcast should be compared with the November 27 forecast derived from the minor three peaks and a domed house formation, the November 18 and 29 dates derived from the theory of basic advances and basic declines , and with the November 17 and November 21 dates derived from mirror image calculations. All of these puzzle pieces fit together. It seems very likely that an important top in the Dow will occur during the second half of November 2006.