How to Decide When to Buy, Sell, or Wait
12/29/2011 2:00 am EST
Starc bands stand out among other overbought/oversold indicators for their ability to generate highly reliable signals for traders in all markets and time frames.
As we approach year-end, I often advise traders (and investors also) to use some of their holiday time to review the past year's trades. There are many reasons to do this, and I generally focus more on my losers and those with small gains than I do on big winners.
Once you have your list of trades, there are probably some you can immediately identify that were a bad idea from the start.
Others were likely the result of poor risk/reward analysis whereby a stop was used that was too wide. I am sure there are others for which you will conclude the trade was justified, but the result would have been much different if you had just waited for a better entry point.
Buying too high or selling too low is a problem most of us have, and while there are countless overbought/oversold indicators to help you, the best tool that I have found are the Starc bands.
Starc bands were developed in the mid-1980s by the late Manning Stoller, with whom I had the pleasure of teaching technical analysis in many cities around the world. Starc stands for the "Stoller Average Range Channel," and by far, these are my favorite banding or channel techniques.
The same formula is used on all markets and for any time frame and goes as follows:
- Starc+ = 6-period moving average + (2 x 15-period Average True Range) (ATR)
- Starc- = 6-period moving average - (2 x 15-period Average True Range) (ATR)
The Average True Range (ATR) was developed by Welles Wilder and was discussed along with the formula in an earlier trading lesson.
The beauty of the Starc bands is that unlike most indicators or methods, they tell you when it is a high- or low-risk time to buy or sell. Using two times the ATR, Stoller estimated that 90% of the price activity should stay within the bands.
As for interpretation, if prices are near the Starc+ bands, it is a high-risk time to buy and a low-risk time to sell. Conversely, when prices are near the Starc- band, it is a low-risk time to buy and a high-risk time to sell.
NEXT: The Most Memorable Starc Band Signal from 2011
On August 17, GLD traded above the monthly Starc+ band that was at $173.30. The August high was $184.82, and then in early September, GLD made a higher high of $185.82, which was also above the monthly Starc+ band (point 3). Since 2006, this had only happened twice.
In April and May 2006, GLD also traded above the monthly starc+ band (point 1) before dropping 23.8% from the May highs. GLD also traded above the monthly starc+ band for three months during the period from January through March 2008 (point 2), when it rallied 20%. This was followed by an eight-month correction that took GLD 34.3% lower, from $100.44 to $66.
GLD is currently just 15.9% below its September highs, as the decline so far has not been too severe. The monthly on-balance volume (OBV) through the end of November was acting stronger than prices, which is consistent with a positive long-term trend.
In August and September, GLD stayed above its weekly Starc+ band for three weeks, which was also a strong warning signal. This last occurred in GLD in April 2006 when prices exceeded the Starc+ bands for five weeks.
Once prices turned lower (point 1), the drop was quite dramatic, as it took only five weeks to drop below the Starc- band, point 2. During the week ending November 10, 2007 (point 3), gold's high at $848 slightly exceeded the Starc+ band at $837, and for the next six weeks, gold moved sideways, testing the six-period MA before resuming its uptrend. This is a very typical pattern.
Gold had reached a low of $773, which was $75 below the high. By the latter part of March 2008, gold reached $1033 (point 4), peaking just below the Starc+ band at $1043. For the rest of the year, gold moved back and forth between the Starc bands and tested the Starc- band three times, culminating in the October 2008 drop (point 8).
Gold rallied steadily from these lows and by late-February 2009, it approached the Starc+ band (point 9), which turned out to be the beginning of a triangle formation. Frequently, a test of either the Starc+ or Starc- band will be followed by the formation of a continuation pattern before the trend resumes.
NEXT: Study Another Past Example in Gold
Another interesting example developed in late November and early December 2009 when for two weeks, gold popped above the Starc+ band at $1207.95 and hit $1227.50, point 10. This correction was discussed in detail in this recent Trading Lesson, and as noted, gold traced out a nice triangle formation by February 2010 and came within $40 of the Starc- band (point 11) before the correction was over. When either band is reached on both a weekly and daily basis, the signals are even stronger.
For example, on May 13, 2010 (point 1), the December 2010 gold contract moved above the daily Starc+ band, and if you look back at Figure 1, this corresponds to point 12, as gold also came within $16 of the weekly Starc+ band.
Just one week later, gold was back to the daily Starc- band and stabilized for a few days before again turning higher. This sideways period lasted into August, and in early July, the Starc- band, point 3, was again tested, setting the stage for a brief bounce before gold declined further.
On an hourly chart, this bounce was clearly a continuation pattern. In October, the daily Starc+ bands were tested twice in less than a week (points 4 and 5), and this coincided with gold just matching the weekly Starc+ band, as noted at point 13 on Figure 1. Therefore, this was clearly a high-risk time to be buying.
After the ensuing correction, gold dropped $70 from its highs and reached the lower daily Starc band (point 6), presenting a lower-risk time to buy. Just two weeks later, gold was again making new highs near point 7.
The same Starc band formula can be used on any market, no matter how volatile, and as an example, let's look at Silver Wheaton (SLW), a stock with a beta of 1.56, which means that it will move just over one-and-a-half times as much as the overall market.
SLW bottomed ahead of the overall market, as it turned higher in late 2008 and reached the Starc+ band at point 1. Then, on the week ending July 11, SLW approached the Starc- band (point 2) and held above the major 50% retracement support level. Therefore, this was not the time to sell and actually turned out to be a more favorable time to buy, as SLW turned higher the following week.
There were two other instances in 2009 when the upper and lower weekly Starc bands were almost reached, points 3 and 4, and in both cases, prices reversed. In May, the weekly Starc+ band was tested (point 5) before SLW formed a trading range. As I mentioned earlier, this frequently occurs when prices test either of the Starc bands before tracing out a flag or triangle formation, which can take up to ten or 20 bars before the trend resumes. I have highlighted this period with a square labeled "a."
After SLW broke out of this formation, it tested but did not exceed the Starc+ band for five consecutive weeks before consolidating and moving back to the six-period moving average (circle b). Within four weeks, SLW had moved above the weekly Starc+ band.
NEXT: Using Starc Bands to Time Short-Term Entries
For short-term traders, the entry price is even more important because generally, tight stops are used and a common problem is either buying too high or selling too low. Therefore, I wanted to include this hourly chart of the cash EUR/USD to give you an idea of how the Starc bands work on a shorter time frame.
On October 28, 2010, the euro tested the Starc+ band, point 1, and then moved sideways for eight hours before making a brief, marginal new high (by six pips) before reversing to the downside. Clearly, too tight a stop above the prior high would not have worked.
One method that Stoller used was to also plot a second version of the Starc bands that used three times the ATR instead of two times. In this instance (point 1), the wider (3 x ATR) Starc+ band was at 1.3970 versus 1.3940 on the regular (2 x ATR) plot, so a 30-pip stop could be justified and would not have been hit. By the next day, the euro had dropped to the Starc- band at 1.3810.
On November 1, the Starc+ band was again reached. I want to emphasize that similar bar or chart formations are observed regardless of whether you are using weekly or five-minute data. A good example of this developed between November 2 and 3 (see box), as after the second touch of the Starc+ band, the euro moved sideways for 16 hours before resuming its uptrend, point 4. After peaking two days after the election, point 5, the euro started to trend lower, but each time the Starc- band was tested, prices stabilized or moved a bit higher for a few bars before the decline resumed.
Since the formula for the Starc bands is quite straightforward, it is not difficult to add it to most charting packages. I use these bands in both my investing and trading, and I even run it on the mutual funds in my 401(k). I also use the bands to help determine a limit price to enter or exit a position. Like with any good indicator, I suggest that you work with it first, study it on several markets, and convince yourself that it can help you before using it on a real-time basis.
Tom Aspray, professional trader and analyst, serves as senior editor for MoneyShow.com. The views expressed here are his own. Readers can post questions or feedback in the comments area below or send to TomAspray@MoneyShow.com.