Joined: 17 Oct 2005
Location: Portland, OR
|Posted: Thu Apr 27, 2006 7:52 pm Post subject: A J-Hook PCF
|Here is a J-Hook PCF Iíve been working on; it is still a work in progress and I can not guarantee the results you will get using it, but it does find some J-hooks.
Before getting into the PCF it is probably best to review what a J-hook is. Simply stated it is a short pull back after an uptrend. The pull back is attributed to a few days of profit taking consolidation. Like many moves in the market there is some symmetry; with a J-hook the uptrend before the pull back should predict the size of the uptrend after the pullback, provided the recent high is exceeded.
I will explain what I intended for each section so you can modify it to suit your needs or tune it for better results
((L1 = MINL4) OR (L2 = MINL4) OR (L3 = MINL4) )
This first line is simply ensuring that the low in the last 3 days is the minimum low in the last 4 day; some pullbacks are shorter in duration some are longer; this variable time is part of what makes writing a robust J-Hook PCF so difficult.
( (MAXC3 < MAXC4.3)) This is simply ensuring that the max close over the last 3 days is lower than the max close over a 4 day period beginning 3 days ago. Once again the 3 day period is arbitrary, the pull back could take longer, if it does there is still a good chance that the max close of the recent 3 days will still be less than that for the prior 4 days. (Part of the test for a recent peak a few days back)
( (H3 = MAXH15.4) OR (H4 = MAXH15.4) OR (H5 = MAXH15.4) OR (H6 = MAXH15.4) OR (H7 = MAXH15.4) ) This section is making sure that the highest high in days 3 Ė 7 is the highest high in days 4 Ė 19 (for 15 days beginning 4 days ago) (Part of the test for a recent peak)
(((MAXH4.3 - MINL4) / (MAXH4.3 - MINL21.3) > .23) AND ((MAXH4.3 - MINL4) / (MAXH4.3 - MINL21.3) < .62) ) This section is determining the percent of retracement from a high in the 4 day range three days back to the recent low compared with the range of the lowest low in a 21 day period beginning 3 days back; and then ensuring the retracement is at least 23% and not more than 62% (this is also arbitrary and set to the common Fib retracement points. The 21 day period to gauge the retracement is also arbitrary, it could be longer or shorter)
((AVGH3.5) > (AVGH3.8 ) AND (AVGH3.8 ) > (AVGH3.13) AND (AVGH3.13) > (AVGH3.18 )) With this section I compare three 3-day samples to determine the uptrend, if each the average high in each section is higher than the average high is the previous section, it must be an uptrend. There are probably more elegant ways to determine an uptrend but itís what I ended up using. (And yes the 5, 8,13 and 18 days back numbers are also arbitrary, the uptrend could be longer or shorter, but you have to start somewhere.
So, there you have it. It may miss the very short/quick J-hooks and it may miss the long ones or long uptrends. I tried to hit a happy medium that would catch the largest percentage of them. You will still have to eyeball the results and choose the oneís to your liking but itís better than wading through your universe looking for them.
If I make any significant changes to this PCF I will post the changes to this thread. Good trades to you.
Just copy and paste into a new PCF, it does work as it is:
((L1 = MINL4) OR (L2 = MINL4) OR (L3 = MINL4) ) AND
( (MAXC3 < MAXC4.3)) AND
( (H3 = MAXH15.4) OR (H4 = MAXH15.4) OR (H5 = MAXH15.4) OR (H6 = MAXH15.4) OR (H7 = MAXH15.4) ) AND (((MAXH4.3 - MINL4) / (MAXH4.3 - MINL21.3) > .23) AND ((MAXH4.3 - MINL4) / (MAXH4.3 - MINL21.3) < .62) ) AND
((AVGH3.5) > (AVGH3.8 ) AND (AVGH3.8 ) > (AVGH3.13) AND (AVGH3.13) > (AVGH3.18 ))