SCT is programmable. http://www.elitetrader.com/vb/attachment.php?s=&postid=576715 http://www.elitetrader.com/vb/attachment.php?s=&postid=576831
Hi redduke, Thanks for your offer, I appreciate it. I'm left with the impression that Jack's SCT is like an onion, in that as you peel off one layer of complexity a deeper one awaits. As such, the logical choice is to start from the beginning, with channels. I am working at this very moment on building a numerical construct for tapes, pt3 channels, and longer time frame channels. This is all done on historical data. A real time feed with real-time generation of channels is the natural extension, so I'll worry about platforms / tools at that time. I'm most comfortable with TS easylanguage, so I like to prototype with that. It will be a giant milestone to auto-construct the trendlines. From that point, consideration is given to trading, starting with the elementary levels Jack defines. Wish me luck. RoughTrader
ERLURKER, Where did you find those snapshots? I'd be interested in following the discussion.... RoughTrader
automated channels and annotations here: http://www.elitetrader.com/vb/showthread.php?s=&postid=756183#post756183 http://www.elitetrader.com/vb/attachment.php?s=&postid=756202
You stated that SCT is programmable.I clicked on those links expecting to see something that supported your statement. All those links showed were hindsight charts. Where is the coding?
This is off to a good start it seems RoughTrader, good work. I tried automating the channels simply as a chart indicator and to use in some backtesting, but I had an awful time with the length to draw them. There is no fixed length in SCT, and to capture the events you need three seperate layers of data (obviously not the hard part) but it seemed like you need to create events for each dataset or maybe just a state for each dataset to know where you are at all times. One approach I was going to look into was skipping to the smallest level of channels. Jack, if you're reading this I know I'm "inventing" but something you just wrote seems to support this. If an FTT on a L3 channel is an FTT on a L2 channel which is an FTT on the L1 channel, why not just draw the channels for the tapes and act on them only? One reason that comes to mind is that you take "extra" actions on the bigger channels, when they all line up, however, that is not my understanding. Anyway, it's worth considering doing the tapes and creating a higher frequency system, because if my understanding of SCT is correct (which it probably isn't), then it seems that this one level of channels would capture all the action points. Anyway, good luck, and if there's anything I can do to help... TNG
Jack, spyder, or other experts on SCT: I'm making good progress in coding channels, starting with tapes. There are several bar formations I have questions about. They are illustrated in the attachment. At the start of a new session, as suggested by bundlemaker's video, we wish to use at least the first two bars to make a tape. My concern is with the formation of large "outside" bars: 1) If Bar 1 is "inside" of bar 2, these two bars cannot be explicitly used for a tape because H2 > H1 and L2 < L1. Do we ignore bar 1 and start taping from bar 2? 2) If Bar 1 is a wide-range bar, and we have a succession of bars that are inside bar 1, do we also ignore bar 1 and start taping from bar 2? Thanks, RoughTrader
RT, I admire your ambition, but you are about to find out that Jack does not give out hard-and-fast rules. To do so would subject the method to the kind of rigorous testing you seek to accomplish. Let me simply say that others have gone before you. My comment is meant in no way to denigrate SCT. Quite the contrary, I have found the study of it to be quite thought provoking. If nothing else, you will exit your own study with greatly enhanced algorithm development skills, as did I.