One of the True Strength Index (TSI) indicator techniques I have found quite reliable for yielding profitable SELL signals is called the negative divergence. It has concerned me lately that the Market Vectors Gold Miner ETF (GDX) appears headed for generating this sell signal and in the past couple of trading sessions I have lightened up on my mining positions considerably with this very thought in mind.
Then I got to wondering, what are the true odds that a TSI negative divergence sell signal on something like GDX would actually foretell a drop in the price of the miners ETF? How often does the signal fail? When it works, how well does it work? And so on.
Before I tell you the two rules I use to define a negative divergence, I should tell you that what follows are 6 calendar year charts of GDX (2007 - 2012) with the TSI indicator. Each chart shows the price and TSI readings that triggered the trade, as well as a possible near-term closure of the trade and the outcome in terms of gain or loss.
And at the end of this article I will recap our findings which, as it turns out, may rather surprise you.
OK - so what is a negative divergence sell signal? I have two rules. First, on an intra-day basis, price makes a higher high and the TSI that day closes lower than the TSI reading of the previous price high. In short, price makes a higher high while comparatively, the TSI makes a lower high.
Second, a TSI trend line break BUY signal trumps the negative divergence SELL signal. In this very unusual situation, the SELL signal from any active negative divergence is closed and it is BUY time. That's it.
I enjoy doing research on stock market related stuff because it helps me be a better trader. This article, and the hundreds of articles I have written before this evening, are first of all for me, and secondly, for anyone who is interested in my research. I spend eons more hours making the charts and crunching the numbers for the readers than for myself. I usually figure out what I want to learn fairly quickly.
Anyway, I had a challenge to figure out how to present this data so that it would effectively communicate the possibilities of the TSI negative divergence SELL signal. To this end I made a couple of rules to guide myself and I would like to share them with you now.
First, the GDX price to the penny that initiates the negative divergence condition is the price I used to begin my imaginary short trade. Second, I closed each trade on the most favorable terms possible within a reasonably short period of time.
Now if your thought is, 'wow - this isn't realistic at all', I quite agree with you. It is not realistic to think a trader could cover the position on exactly the perfect day at the absolute lowest penny of the day. I know that.
Still, I would like to have a feel for how often this trade works, does not work, the general size of the gains and losses possible, and so on. This is research, not reality, but my findings will benefit me and I hope they benefit you as well.
So let's get started with the daily chart of GDX - 2007.
Click on any chart to ENLARGE
2007 had four instances of the negative divergence sell signal. The gray shaded vertical rectangles on the price chart locate the exact bars that apply. The first trade could have been closed about a week later at roughly break even - but I documented the occasion that the trade was closed 3 weeks later for a 2.1% gain or closed 4 weeks later for a 8.4% gain.
The second trade (see second gray vertical rectangle from the left) closed within days for a 6.0% gain or the following week for a 12.1% gain.
The third trade closed 8 or so trading sessions later for a 6.7% gain and the fourth trade could have been closed in 8 days for a 3.4% gain or the following week for a 9.6% gain.
The 2008 GDX had a couple of negative divergence trades that worked out well, then at the end of the year it got rather complicated. Let's take a look at that chart now.
The first trade was worth a maximum of 5.9% and the second could have yielded 13.6% in a matter of days and with a couple weeks of patience reached a sweet 32.4% gain.
Then in the month of December, as the D-wave completed the infamous 2008 bottom and price began it's meteoric ascent, the plane hit a whole lot of volatility. Price was jumping around like a guy who just found out he has flees. Anyway, the first trade was a dud. You would have been lucky to close the trade a couple days later with a 1.6% loss. If you didn't take that out you got one last chance about four weeks later (on the next chart - 2009). This too would have resulted in a 1.6% loss if you were lucky. After that, price has never again been as low as $27.....to this very day.
The other trade was ignited on December 30 or so. On the following chart (2009) we'll see that that one was worth as much as 19.8% in a matter of a couple weeks or less.
2009 produced a lot of negative divergence signals as the miners powered out of their terribly oversold condition. By the way, a negative divergence, at least the way I think about it, is a condition where price simply gets ahead of its own momentum. It is a condition that says that price is out of whack and needs to be corrected in order for the rally to continue.
The February trade could have yielded between 12.4% and 22.0%, depending on your level of patience. The March trade was worth 10.4% in a matter of days, or as much as 19.4% if you held in there for 3 weeks. June yielded a possible 17.0% gain. Interestingly, this is the rally out of the B-wave bottom of our previous ABCD repetitive gold pattern. And July tacked on another 8.2%.
The November time frame was tricky. It included a single trend line break and what appear to be several negative divergences. But on very close examination, using the rules I defined earlier (and always use for myself) there were no negative divergences. If you get out your microscope and think it through bar after bar, I am confident you will come to the same conclusion.
2010 sported 7 more negative divergence setups.....all potentially profitable. 8.6%, 11.0%, 10.6%, 1.8%, 3.8%, 5.1% and 7.8%. The October time frame was a mess. It was one of those 'he said, she said, he said, she said' and so on. Though price managed to inch higher within these pink shaded vertical rectangular days, I have seen this enough to somewhat instinctively know that something is very wrong here - just hang on because their is going to be a shake-out and then it will be time to go.
The 2011 GDX chart above gave us just 4 potential winners. The February/March period was messy - we saw that once earlier - and I shaded it pink to remind myself about these things. There were actually 3 negative divergences nestled in that mess.....if you played the last one you could have taken a ride from 61.25 down to 54.68 (10.7%). I did not put this one on the chart - I just looked at the mess and figured if you traded the first one and got out with lots of luck, you would have made 4.8%.
The 3 following trades were garden variety - potentially worth 11.8%, 4.0% and 3.4%.
Well, here comes the 2012 GDX chart. Of course it's just a tad shorter than 6 months to date, but only 1 single trade that was worth 10.5%.
The other thing to point out on this 2012 chart is our current situation. The previous TSI high was 63.56 and the price of GDX on that occasion, intra-day, reached a high of 48.72. Looks like I made this chart before the close today...which should read TSI 43.18 and high price for today of 47.89. In other words, if price reaches 48.72 ($1.04 higher than today's close) and the TSI on that bar does not close higher than 63.56 that will yield the second negative divergence for 2012.
By the way, the workaround is for GDX to drop just enough to cause the TSI to drop. That would create a new high in which price is not out of whack with the TSI. Then a trend line break could be drawn as price causes the TSI to rise again, and that would be a BUY signal.
OK, let's recap what we found. We had no fewer than 24 trades. 23 of 24 could have been closed for a profit. The definite losing trade was coming out of the 2008 D-wave bottom. The loss could have been significant but there were two opportunities to escape with a loss of 1.6%.
The potential gains for all the winning trades were not ingredients of a get-rich scheme but were healthy enough to consider playing.
If you have any questions about the TSI negative divergence SELL signal technique, or something else stock related, please feel free to write me an email and I'll see what I can figure out - even better I will probably learn something from you!
Keep in touch,