Charting 101: What is a Trendline?
What rules do you use with respect to trendlines?
I believe, as does Minyanville Prof. Bennet Sedacca and Brian McAuley (my partner at SitkaPacific), that the long term trendline on the US long bond is now broken. Bennet posted a chart and Brian on his own came up with a similar looking chart that I posted on my blog. Here is a comment I received from "AJ" about that chart.
To draw correct trendlines, you need to follow the same rules all the time. The rules are, connect two consecutive highs or two consecutive lows at the price extreme.
See a chart here.
That trendline of Brian's started at the top of the 1981 highs and magically sliced through the upper part of the 1984 highs so he could make the line appear closer to the 1994, 2000, and 2004 highs.
He did something different with the line through the 1980 and 1986 lows, making it appear closer to the 1993 and 2003 lows. He actually started below the 1980 lows, sliced through the 1986 lows, but somehow magically made the line tag the 1998 and 2003 lows. Well that's completely biased. The ends justify the means, I guess.
There are no consistent rules he is following to draw those lines other than making them illustrate a bias. The bias is a 26-year trendline and range has been broken, and that supported your case why bonds went up last week after remaining inside a range that the chart illustrated had held for 26-years.
As I have mentioned and illustrated before, that bias is incorrect. The 26-year downtrend remains quite intact, by a pretty long ways.
The rules for trendlines are simple. Connect two consecutive highs or two consecutive lows. Draw at the extremes, not the bodies, as ignoring part of the candle brings bias into the trendline.
Now, I would not be trumpeting that trendline break unless I believed it (given my position that another treasury rally is in store based on increasing domestic demand), but that is beside the point.
Count me in the group that thinks a trendline is not even valid unless it has three touches. Given that we are now waiting 23 years for that third touch as shown in the link from AJ, I think that is a bit too much "dogma" about methodology. Then again, AJ does not require three touches whether I do or not. Also I am willing to throw out the absolute high or low on the belief that it is often an invalid starting point. I read somewhere that some always throw out that first point on the absolute top or bottom but I cannot find that reference just now.
Brian had this to say in response to AJ...
Well, the best way to draw a trendline is to let the market show you where it is - not depend on some theoretical 'trendline rule'. On the TYX chart, the market struggled with the trendline - as drawn on my chart (as shown in The Bond King's Capitulation) - for a good time last summer before turning down and staying in the downtrend. Now, a year later we have a decisive breakout just where it should have been if the line would be if extended down from last summer's struggle. The only high this trendline misses is the '94 high, but other than that it connects 5 highs before this year's break.
Connecting the first two points on the a trendline isn't always where the market decides the trendline is. Often the first point (at which the chart makes a final low/high in the previous trend) isn't a valid point anyway, and real trend line only shows itself from the second point on.
People can deny that we have a break up in yields if they want, but both TYX and TNX showed decisive breaks this week IMHO. The stock market, with its sell off this past week, realizes this too - as does the dollar. Since the market clearly seems to think yields have broken up, I side with the market (and the chart).
Looking at all these charts one more time, I lean toward the following chart (slightly different than Brian's but nothing that a crayon can't take care of). Besides, and this is important: just because a trendline breaks does not necessarily mean the long term trend has changed. Trendlines reform all the time.
In this case however, my opinion is that it is likely but not 100% guaranteed that the secular move towards lower yields is over. Whether or not it is over, there is nothing preventing a retest of the 4.5-4.0 area (and that is something I suspect will happen as housing weakens further and the strength in the economy is shown to be a mirage).
Anyway, I think AJ's questions are worth exploring and would ask other Minyans and Minyanville Professors, what rules do you use with respect to trendlines, and how hard or "etched in stone" are those rules?
I sense this could be a useful discussion for many Minyans. Any Professors care to chime in?
P.S. A Public Service Announcement for Minyan Bloggers...
Anyone using sitemeter on their blogs or going to blogs where sitemeter is used might be infested with up to nine tacking cookies that may degrade your computer performance of if you are writing blogs, your user's performance.
Details can be found in Selling Out To Spyware.
Copyright 2011 Minyanville Media, Inc. All Rights Reserved.