If you don't take time to understand and manually practice this method, I'm convinced you'll fail. No amount of hand holding can help you. Don't waste your time! Also, this method is not a set of rules to be mechanically applied, it is a mode of seeing the market. So I doubt it can be fully automated. You may just eventually create a set of tools to help you.
I agree with you 100% cnms2, I commend roughtrader for his effort and enjoy reading his posts here but I really feel like this method isnt codeable, at least not unless you've been doing it for some years and can then account for many of the subtle variables that pop up. And even then, I feel like it would be incredibly difficult. Maybe you could code the most simple of all situations but the complexity past that, to me, seems like an almost impossible task.
cnms2, I did not say I would extract from the pool of knowledge here and then shut my doors once I get what I need. Code is the baseline by which the technology is implemented. It is not the single representation of core ideas and algorithms. My coding style is unique to myself and would be useful to others in only a limited extent. My desire is to organize the rules from the pool of knowledge. Once I have reached a satisfactory milestone, I am happy to discuss "methods" by which I have organized the Hershey method into quantitative modules. I suppose the closest thing you could compare this to is "pseudocode". Please remember, my journey is to "go beyond the hype" of this method and find out for myself whether this thing can be captured in backtest. That being said, I will quickly dismiss both cheerleaders and disgruntled pessimists. I am only interested in garnering facts and rules, and reporting back to everyone who participates in this thread. If you don't feel like chipping in, that's cool. But if / when I return with some demonstrated results, good or bad, I think people will raise more than an eyebrow or two and open the way for more discussion. This is not about protecting intellectual property. It is about organizing this system and reporting back to you how I have organized it. RoughTrader
Hi Steve, Thanks for chiming in. I'm still at the beginning stages of understanding the method. That means, drawing tapes and channels. For about 2 weeks I have spent about an hour each day drawing tapes and channels for historical charts. In addition, I have been taping and channeling charts and drawing the gaussians for live market action. I have not found one session of ES (and most likely the case for the remaining e-minis) where I am "stuck" in the flow of putting down tapes, then creating channels around them. I have realized that it all comes down to creating, extending, widening, creating, extending, widening, and so on and so on. There is most definitely a "seamlessness" in the progression through each session, rooted in the inherent overlap of channels. It is because of this seamlessness I can actually see the quantitative bar patterns that dictate the geometries of the tapes and channels. They can most definitely be cataloged in code, and auto-taping and auto-channeling can be done. It just takes time to sit down and think about how to organize what seems to be a sea of nuances that must be accounted for. As I continue this process, I find it to be more iterative. The more tape patterns I come across, the more diverse channels I come across. Some cases leave me questioning, "Is the tape I have drawn correct? The progression from the up-channel to the down-channel seems awkward. Or it seems inconsistent with the last channel transition." In these instances, I revert back to the tapes and see how I drew them manually around bar formations. I then question whether certain bar formations are valid for marking tapes or not. And as you probably know, there are MANY types of formations that the Hershey document does not describe explicitly. Most mechanical strategies will only issue orders at the close of the bar. Jack's method REQUIRES order entry intrabar to be profitable. Fortunately, the default fractal is the 5 minute bar, which can be decomposed into five 1 minute bars. My hope at this point is that 1 minute bars will provide sufficient approximation to the timing required for intrabar triggers. However, that is not my immediate goal at the moment. Regardless of whether this journey leaves me at a destination of realizing automated order execution, the auto-chart annotation engine is worth development in and of itself. It is here my focus will remain for the time being. RoughTrader
An example of a bar progression in which the tape is not immediately obvious. The examples go on and on.... RoughTrader
Hi nkhoi, on page 24, didn't you comment that a pennant would need more than 2 bars to form a tape? second, this is the first I've heard of "split bars". There's two issues here: 1) we don't know, without a tick-by-tick replay whether the outside bar actually decomposed the way you draw. It could have hit the low first, then the high, then closed below (H + L) / 2. In this case, your split bars would be descending first, instead of ascending. 2) How the heck do you tape that on a 5 minute chart? Thanks, RoughTrader
that is how I view it, I have no idea how Spyder views it. split the way it makes sense you. flat a bit, again this is just my view