Tuesday, February 7, 2012
TSI Trend Line Break Mathematics - XGLD
I have updated the chart with the 20 short/SELL ticker symbols with today's closing prices. Today 13 of the 20 stocks (65%) closed lower than their close of last Friday. Also, today was a fairly positive day with only 42% of NYSE closing in the red.
Some readers have sought some clarification on how I define a trend line break and I hope this post will provide the relevant details. I will show a couple of charts - one demonstrating the trend line break BUY signal, the other demonstrating the SELL signal - then conclude with a chart that is in serious jeopardy of yielding a Negative Divergence SELL signal (XGLD).
This past fall I took quite a few weeks off from writing this blog to work intensively on my skill to program the True Strength Index (TSI) indicator BUY/SELL signals using the Think or Swim platform. One of the techniques I successfully mastered was an algorithm for having the computer accurately identify the trend line breaks of the TSI indicator.
Our first chart is Alexco Resources (AXU) and demonstrates the mathematical calculations I used to get the computer to tell me when there was a trend line break BUY signal.
The technique I used was to have the computer recognize when the TSI had made a reversal in direction that resulted in a 'high point' - such as we can observe on Days 1, 16 and 21.
The next step was to calculate how far the TSI had fallen in the interim periods then divide that calculation by the number of periods. That gave me the slope of the trend line.
Finally, the computer kept track of the TSI reading for each day and compared that with the reading expected that day if the TSI continued at or below the established slope. Once the computer got a reading (such as occurred on Day 25) that was ABOVE the expected reading, a trend line break had finally occurred.
Here is a second look at this technique - this time the SELL signal. The algorithm is the same, of course. Find at least two points that define a slope then wait until the mouse crosses down through the radar zone to generate a trend line break SELL signal.
I hope this post makes clear that I need to have at least two points to define the slope of a trend line break. It really cannot be defined without at least two points.
But I have also written earlier my belief that the trend line break trumps the divergence signal. If there are two points from which I can define a slope and the TSI moves such that the slope is violated - I recognize that as a trend line break and do not look at the situation as possibly capable of generating a divergence BUY/SELL signal.
We'll conclude this discussion with an example of a stock that - at the moment - does not have a trend line break anywhere in sight...... and therefore is likely to yield a negative divergence SELL signal very soon.
This is a daily chart of our good friend XGLD. Should price continue higher, the TSI will go higher. But I suspect that if price makes a new high the TSI will NOT make a new high.
Also, the closer the TSI gravitates towards ZERO, the more likely a SELL signal will have a powerful effect. That is because of the absolute TSI certainty that if the TSI is falling below ZERO, price is also falling. The same cannot be said when the TSI is above ZERO.
And, notice that XGLD's TSI reading is about as low to ZERO as when it's rally began over a month ago. Every hot air balloon lands sooner or later - and so will this one.