I'm somewhere between (2) and (3), but am more open to (1) after tracking Adam Hewison's predictions on spot gold for a couple of months. I would like for (1) to be true, as would most people, for obvious reasons.
Certain chart and volume patterns are supposed to foretell a stock's (commodity's, ETF's, etc) future price movement. Occasionally, if I come across such chart patterns I'll post about them. Then, I'll check back after a while to see how the predictions turn out. With enough such experiments, it'll hopefully be possible to either confirm or deny (3).
Since a stock can go either up or down from any given point, there's a 50% chance for each prediction to turn out correct (I'm ignoring the possibility that it'll stay the same). If technical analysis doesn't work, predictions should be right around 50% of the time. If it works, predictions will be right or wrong the significant majority of the time. What constitutes a significant majority of the time will be determined by the number of predictions made. If it seems wrong to say that technical analysis works if it's wrong almost all the time, remember that it is just as hard to be wrong most of the time as it is to be right most of the time when you have a 50% chance of being right or wrong. Put another way, if it's wrong most of the time, technical analysis can be very useful, as we can do the opposite of what it predicts.
A major thing that can skew results is that I can easily misread charts. For this reason, I'll stick to the so called archetypal patterns: double top, double bottom, head and shoulders, cup and handle, etc, that are easy enough to identify with software.
Today, I'll look at two chart patterns, the "double top" and the "double bottom."
The double top chart has two peaks that are roughly around the same level, with a moderate dip in between. It is supposed to be a bearish signal when the stock goes below the level of the dip between the two peaks.
Here's one stock that has a double top chart: American Capital Agency (AGNC). The pattern manifested over the last seven or so days, in the right corner of the image below. AGNC closed at 19.13 on 12/10/08. Let's see where it will be at the end of next week.
A double bottom is essentially the opposite of a double top. It is two about equal dips with a moderate peak in between. The chart is supposed to be a bullish indicator when the stock's price goes above the peak.
Drugstore.com (DSCM) closed at $1.19 on 12/10/08. The double bottom for DSCM started out in mid November:
So, let's see how these stocks do in the near future. We're looking for DSCM to go up and AGNC to go down.
Disclosure: I hold no positions in the securities mentioned above.
More information is always better than less. Click here for analysis on any stock, commodity, currency, or ETF.
5 comments:
Awesome post--I'm eager to see how the two stocks turn out vs their technical-analysis predictions. I, too, lay somewhere between 2 and 3, but probably closer than 3. I bought a few cent's worth that I had leftover during a free trades promotion. It had been predicted by technical analysis that it was going to plateau. Shortly after, it went up a lot, and has since hovered way above its "plateau".
One also has to account for the fact that if you are trading using technical analysis you might incur more trading fee's, thus narrowing your margin.
The technical trader's toolbox includes so many tools that its really hard to dismiss the trade as a whole. I tend to lean your way (between options 2 and 3), but recently have also opened up to the possibility that there exists some validity.
One of the problems is the huge survivorship bias in the industry as a whole. the losers don't brag about how much they lost and the remaining technical analysts passionately defend their method.
I experimented briefly with moving average crossovers and did a bit of brief analysis that makes me believe that some forms of tech analysis may have more validity than others.
Check it out: http://fattyfatfat.com/2008/12/moving-average-crossover-theory/
After seeing this post I was a little disappointed that more responses weren’t provided. I to have tried what you are suggesting. Currently I am between 2 and 3, but leaning a little closer to 3. Most of my experimenting (and real loses) have occurred since December 2007. I can draw at least one BIG conclusion: that outside factors will easily outweigh predictions from these methods (which is a disclaimer most state). As in courses (presentations) such as Better Trades, it is easy business to look back at selected chart patterns and say, see what happens. And everyone thinks, WOW. One thing that I am very sure of is when we are near or at a market bottom, such as now; the next move over the follow couple of years is UP. Then somewhere between 5 and 10 years following the bottom there will be another down turn.
Lastly, I believe myself to have a good grasp of statistics and probability. And in classes such as Better Trades you will have loser, winners and those that break even…that’s just chance. And the group then focuses on the folks that get it right (lucky) and wonder why they can’t do the same, when it is really is just chance/randomness.
Thanks for the comments, guys. Sorry for the super late response.
Carl, you're absolutely right about trading fees. But if technical analysis works, I wouldn't mind paying the extra fees.
Fatty, what do you think of support and resistance lines?
Technical analysis is a lot more than just a set of patterns recognized in a chart. A good technicalist relies on both chart reading/trend following and volume analysis. The goal of which is to determine the supply and demand characteristics in the market. Can't say I believe it as a science, but I do believe that it can work for one reason or another.
Double tops and double bottoms are easy to spot in hindsight, but difficult to use in practice especially if you're trying to spot them without any other context. One of the reasons is that to confirm a double top or a double bottom you need to wait until the bottom of the top or the top of the bottom (the dip or peak between) is breached. Otherwise, you're likely going to spot many false positives.
On a final note, technical analysis is highly dependent on time frame. You may see different trend confirmations when looking at an hourly, daily, or weekly chart. So to run this experiment as scientifically as possible, you ought also to determine a time period in which you expect the chart pattern you predict to fully manifest itself. Is the double top confirmation effective if it correctly predicts a down week but not a subsequent reversal? Or, is it only effective if the stock falls to its 52-week lows?
Post a Comment