Custom Search

mathematical structure of the algorithm

The truly new insights provided by the Midas method are twofold. The first is the heretofore unrecognized hierarchal structure of   support/ resistance levels and their a priori prediction in terms of the volume- weighted average price taken over an interval   subsequent to a trend reversal point. The second, to be introduced in the present article, is that there frequently exists a remarkable   underlying structure which dominates or "guides" the bull or bear move as it develops from that point onwards.

This structure is revealed when actual prices are compared to a new type of theoretical   support/resistance curve generated by what   I call the TOPFINDER algorithm. (We expose our taurine – as opposed to ursine - leanings as "BOTTOMFINDER" would be   equally appropriate since the support/ resistance symmetry applies to the new algorithm as well as the S/R hierarchy). In   presenting TOPFINDER, we will follow the same pedagogical path used up to now. To whit, without initially revealing the   TOPFINDER equation, we will show what are hopefully compelling demonstrations of its power. The next steps in the program   would then be to outline the psychological principles underlying the mathematical structure of the algorithm, followed by the display of the algorithm itself.

In the present instance, however, the situation is reversed. I discovered TOPFINDER empirically, and while I have some ideas as   to the underlying principles and mechanisms giving rise to its applicability (which I will put forth in due course), it is still a subject   for conjecture and research. Indeed, perhaps one of you will come up with a useful approach!

To begin, then, let us turn to the figure wherein we revisit Union Carbide. Recall from the fourth article that following an extended   period of "foothill" behavior where UK found repeated support at S2, it abruptly took off in a doubling move during which support   was found at successively higher order levels - culminating with S5. In time, this dramatic bull move ended and the price returned   to the S3 level which was launched at the start of the move.

In the figure we show a new curve labelled "T3", which is launched concurrently with S3. As with the S/R levels, this   TOPFINDER curve is generated by a simple universal algorithm , now containing two parameters to be determined by fitting to   the trend reversal points. As before the first parameter is the launch point. In TOPFINDER, the second parameter represents the duration of the move as measured in cumulative volume.

In other words, TOPFINDER is predicting that the move which started with the launch of S3 will, if the move fulfils its "destiny",   terminate when the cumulative volume reaches the value indicated by the dashed line joining the diamonds, after which the price   should return to the more "normal" (i.e. unaccelerated) support S3. It is seen from the figure that if T3 is fitted to the consolidation   reversing at point "F", then the end of the entire bull move is for all intents and purposes predicted exactly!

Since technical analysis is often referred to these days as "rocket science", we can employ this metaphor by likening a move in a   stock to a rocket launch. Already we have referred to a trend reversal as the "launch point"; now we imagine that - as with a rocket   - the move's duration is pre- programmed by loading a given amount of "fuel" which in our case is a fixed amount of cumulative   volume. During the powered phase of the launch, the rocket's control mechanisms act to follow the nominal trajectory defined by   the TOPFINDER curve. When the fuel is completely burnt, the rocket returns to the Earth's surface represented by the S/R level.

In the market, as with rockets, there are abortive launches which do not go all the way to burnout, but return to Earth prematurely.   Hence the TOPFINDER curve (and especially the predicted burnout point) are to be regarded only as potentialities. So the new   viewpoint is that every time there is a bounce from an S/R level of order "n", we begin the computation of two new curves: the   next higher- order S/R level (order "n+1") and the corresponding TOPFINDER labelled with the same index. The move   originating with the bounce has the potential to "take off" (i.e. accelerate) in which case its trajectory is predicted by   TOPFINDER. Or, it can fail to "ignite" in which case it merely rides along the newly launched S/R level in a succession of less   dramatic "hops" before either attempting a new takeoff- or penetrating the level and dropping back down the hierarchy.