Any scan still depends on your particular unambiguous objective definition of a Pipe Bottom.
Saying that something has an average price variation of 0.24 and an average price overlap of 66% isn't a definition. It describes aggregate characteristics, not the pattern we are actually attempting to identify.
We could check for the two most recent bars to have more than a 66% overlap using the following:
MINH2 - MAXL2 > .66 * (MAXH2 - MINL2)
If you want to use Weekly bars in TC2000 version 12.3, then set the Time Frame of the Condition Formula to Weekly using the drop-down menu in the Condition Formula. Weekly formulas in TC2000 version 7 need to be created manually.
The Weekly formula for the above for TC2000 version 7 would be:
MINH10 - MAXL10 > .66 * (MAXH10 - MINL10)
What else can we find out about Pipe Bottoms by searching the internet? The bars need to be longer than average. We can add that both bars need to be greater than the average over the prior 52-bars as:
MINH2 - MAXL2 > .66 * (MAXH2 - MINL2) AND H - L > AVGH52.2 - AVGL52.2 AND H1 - L1 > AVGH52.2 - AVGL52.2
The Weekly formula for the above for TC2000 version 7 would be quite long because we would need to manually expand the moving averages.
The pipes need to extend below the surrounding price action. Assuming we are attempting to identify the pipes when they form and aren't waiting for some sort of confirmation, then we can only compare them to some arbitrary number of previous bar. The following adds that the lows of both bars at more than 50% of their individual ranges below the lowest low of the previous 5-bars:
MINH2 - MAXL2 > .66 * (MAXH2 - MINL2) AND H - L > AVGH52.2 - AVGL52.2 AND H1 - L1 > AVGH52.2 - AVGL52.2 AND MINL5.2 - L > .5 * (H - L) AND MINL5.2 - L1 > .5 * (H1 - L1)
Excluding the manual expansion of the moving averages, a Weekly version of the above for TC2000 version 7 would be:
MINH10 - MAXL10 > .66 * (MAXH10 - MINL10) AND MINL25.10 - MINL5 > .5 * (MAXH5 - MINL5) AND MINL25.10 - MINL5.5 > .5 * (MAXH5.5 - MINL5.5)
If you wanted this to have happend 3-bars ago with both pipes extending more than 50% of their individual ranges below the lows of the most recent 3-bars as well, we would get:
MINH2.3 - MAXL2.3 > .66 * (MAXH2.3 - MINL2.3) AND H3 - L3 > AVGH52.5 - AVGL52.5 AND H4 - L4 > AVGH52.5 - AVGL52.5 AND MINL5.5 - L3 > .5 * (H3 - L3) AND MINL5.5 - L4 > .5 * (H4 - L4) AND MINL3 - L3 > .5 * (H3 - L3) AND MINL3 - L4 > .5 * (H4 - L4)
If we again exclude the moving averages, the Weekly TC2000 version 7 of the above would be:
MINH10.15 - MAXL10.15 > .66 * (MAXH10.15 - MINL10.15) AND MINL25.25 - MINL5.15 > .5 * (MAXH5.15 - MINL5.15) AND MINL25.25 - MINL5.20 > .5 * (MAXH5.20 - MINL5.20) AND MINL15 - MINL15 > .5 * (MAXH5.15 - MINL5.15) AND MINL15 - MINL15 > .5 * (MAXH5.20 - MINL5.20)
I'm not saying this is going to identify a Pipe Bottom (it might, but it probably won't). I'm just trying to take you through the sorts of steps that are required in actually coming up with a definition and implementing that definition in the Personal Criteria Formula Language.
You need to come up with the definition. Once you have a definition, I can help you implement it as a Condition Formula.
-Bruce Personal Criteria Formulas TC2000 Support Articles
|