Quote from RoughTrader:
Hi Jack,
Thanks for piping in. The issue at hand is not of paradigm, but rather of quantitative implementation. An edge-based system, belonging to what you refer to as "conventional orthodoxy", can be described quantitatively under certain restrictions, as I am hoping SCT can.
SCT, as you have mentioned, focuses on the "here and now" of remaining in sync with unfolding price action. The channels are constructs which enable decision-making in the "here and now." Channels are formed, and channels expire. Seemlessly moving from one channel to another enables us to avoid losing "lock" on the market movements. Regardless of the "level" one is using exploit market action, the channels will always remain the universal construct to keep in sync with the market. I don't see how this cannot be captured with a model that updates each bar to reflect the "here and now" component in addition to projected channel lines, which are simply extended by knowing slope and price intercepts.
The restrictions I was referring to earlier pertain to soft decisions, i.e. those made by subjective interpretation of the trader. The very first issue I wish to address is the following: Does SCT rely on soft-decisions? If so, it is a method that contains stochastic decision points and cannot be captured by a numerical model. If the answer is no, then SCT can unequivocally be described by numerical models, no matter how complicated they are, simply because each decision is deterministic.
There is no answer in between yes and no regarding this issue. The decisions and rules are either hard, or they are soft. For a given "user level", or implementation complexity, there will be only ONE possible trade path of buy and sell points in a session. Soft decisions will yield multiple possible paths.
The question becomes: If we were to strip SCT down to the simplest beginner level, can it be captured by a quantitative model?
Burrowing further down, can one arrive at a singular solution for channel lines in a session using such a model? If the answer is yes, the result is promising. If the answer is no, parameter inputs to the system would determine the solution set, which would be one of many.
RoughTrader