Hi Scott, as I'm sure you know, few people construct trendlines consistently. Ten people looking at the same data will draw them differently, perhaps sharing some areas of commonality. Even the same person, given the same data 10 days later will likely draw the trendline differently. Some people refuse to let the trendline pierce any other data, while some allow this, and some will allow it sometimes. People will also unconsciously draw the trendlines to fit their desired scenario, even if it's inconsistent with the way they typically draw them. Also, some require a minimum number of "touches" of the line, or base the strength/validity of the line on the number of touches.
What constitutes "penetration" of the line varies from person to person also. Such a seemingly simple indicator, yet I'd venture to guess that it is actually one of the most complex of technical indicators to get a handle on, and the fact that few professionals use it consistently AND accurately (which don't necessarily go hand-in-hand) is testament to its underlying complexity. You have to develop the consistency part first, then work on the accuracy. This means you need to start with several consistent, rigidly defined ways of drawing them (no subjectivity allowed). Once you have these, you determine which is the most accurate for the way you're using it.
A couple of things to think about: Regarding use of close versus hi/lo for trendline boundaries, I think to some extent it depends on how you are trading the stock. If you have a stop set (which can trip intraday), then using only the highs and lows makes the most sense because the intraday swings WILL affect your position. You WANT the trendline to represent the extremes of sentiment through the day. If you close a position out only when it CLOSES below/above a certain point, then using the closing prices to construct the trendline makes the most sense (to me, anyway) because the intraday extremes play on a different sentiment than the closing price changes. To see what I mean, look at the volatility of the daily range (high minus low) versus the volatility of the closing prices (close minus close yesterday). These can be 2 extremely different looking charts. Just some food for thought.
Also, you may experiment with creating your trendlines from right to left, rather than the conventional left to right. That way your initial focus is on the more recent data, which carries more significance. Just draw 1 short term line starting on the right. Then draw an intermediate line and a long-term line, also starting from the right. Some charts it doesn't matter, but in general I bet this will give you some more "consistency" candidates to play with if you thought you ran out of ways to draw them from using the "left to right" method. I came across this when trying to program a purely mechanical trend line drawer. I found I always had to look beyond the trendline breakpoint to validate that it is in fact a breakpoint, so starting on the right just made the most sense.
dh |
