Raise a cup with handle to Brooke. We now know the lengths the ATC folks go to avoid going ga-ga looking at those tubes and watching the idiots up there flying funny.
I was amazed by the speed and interactivity of the group in doing the implementation. So much so, that I forgot to reread the TASC article.
Which leads to some observations on making the ultimate QP2 C&H scan:
1.) Perfection is nice, but if a scan misses some opportunities, as long as it hits enough candidates to make the user rich, who cares if some get missed...... So it may be possible to do a sweet spot scan suboptimizes things (confirming the glory of what has already be done and so happily received by all).
2.) You likely want to segment the scan. Just from looking at Figure 3 in TASC, you see formation of a C&H takes a long time. So it is not necessary to do all of the scan at once - and you save time by doing some segments less often. The worst problem here would be a fast handle former (handle to Buy transition) might not be detected in time to do adequate research. So you miss some!
As a sidebar, doing this also lets you look at the "failures and dropouts", those stocks that are forming up fine and then fail - we might learn more looking at them, too.
3.) (For real programmers, which I'm not) Gary provides for external processing by giving us links, I think - and some of this might be done best externally "in a real programming language". So an internal, external, internal multiscan sequence might work.
4.) Order counts. If we had experience, reordering the scan segments so that the "fast killer criteria" run first can speed up things.
And so, the secret is out. Some of those ATC consoles have hidden built in user programmability, so that all of the patterns on the screen aren't just aircraft. |
