HUI Bollinger Bands
Adam Hamilton December 19, 2003 3315 Words
Last week in “Standard-Deviation Technicals”, I discussed the logical and philosophical foundations behind the idea of applying the statistical concepts of normal distributions and standard deviations to technical analysis of financial prices.
While it all sounds complicated, the underlying ideas are actually anything but. Markets tend to abhor extremes, and markets always tend to revert back to their means over time… certainly very easy to understand! The more statistically extreme that prices become, the higher the odds that they will soon mean revert.
As speculators, when we find ourselves observing current conditions where prices are running two or three standard deviations above or below their key moving averages, we can be almost certain that the probability is increasing that these very prices are due to reverse over the short-term and swing back in the opposite direction.
Since speculation is ultimately little more than putting our scarce and valuable capital on the firing line when probabilities seem to be in our favor, standard-deviation analysis provides a valuable tool for analyzing these ethereal odds. Standard deviations grant us another crucial perspective on whether a particular price happens to be relatively high or relatively low at any given moment in time.
In last week’s foundational essay, we used standard deviations built off of really long-term 200-day moving averages to build our charts. This week I would like to expand on our standard-deviation technical analysis, migrating in two important directions.
First, we will limit our focus to the fascinating HUI unhedged gold-stock index, ground zero for the accelerating bull market in gold and gold stocks. Fortunes have already been won in this gold-stock bull for the daring early contrarians who were stalking this thing way back when it began. From a long-term perspective though, the really staggering Great Bull gains are never earned until the final bubble years, so the best almost certainly still lies ahead.
Second, as we are moving from the realm of standard-deviation theoretical into practical, our HUI standard-deviation-band charts this week use John Bollinger’s own SD-band conventions. There are a couple key benefits to using what are considered classic Bollinger Bands while analyzing the HUI.
Conventions become conventions because they have been observed to work rather effectively in the past. Rather than starting with an unconventional approach and moving back towards conventional, it makes much more sense to begin with the accepted and expected and tweak our analysis from there if necessary. In addition, most charting software easily renders standard Bollinger Bands so following convention will make it much easier for folks to replicate these charts on their own computers if they wish.
Bollinger Bands, while not explicitly defined, through standard usage have come to generally mean price bands drawn two standard deviations above and below a 20-day moving average. To compute them you just create a 20dma of a data series, calculate a rolling 20-day standard deviation of the raw price data itself (not an SD of its 20dma!), and add and subtract two times this SD number from the original 20dma line. The resulting graph shows raw prices within an expanding and contracting SD pipeline centered on their 20dma.
These 20dma-based +/-2 SD Bollinger Bands are recommended for intermediate-term trading, or positions designed to be held for weeks to months. By statistical definition, 95.4% of prices ought to lie between these two lines, so a price outside these +/-2 SD bands should only happen in about 1 in 22 trading days on average.
Longer-term SD-band analysis uses 50-day SD bands of raw price data graphed above and below a price’s 50dma. Technical analysts generally use +/-2.5 SD bands for longer-term trading, for positions expected to be held for months to over a year. Statistically, market prices ought to be within two and a half standard deviations about 99% of the time on average.
Outlying price data should only happen about once in every 100 trading days on average, but as I discussed last week market-price bell curves have fattened tails and are not true normal distributions. Thus extreme outliers happen more often than statistics predict. Nevertheless though, even with fat-tailed bell curves standard-deviation analysis is a very valuable addition to any technical-analysis toolbox.
In this essay, we are actually looking at both sets of SD bands for the HUI, the standard 20-day +/-2 SD Bollinger Bands as well as the longer-term 50-day +/-2.5 SD Bollinger Bands. We will start with the longer-term 50-day bands and then zoom into the present from a short-term 20-day SD perspective.
It is really amazing to behold just how many major short-term turning points within the long-term secular gold-stock bull uptrend were carved exactly on key standard-deviation lines! The HUI conforms rather nicely to the conventional standard-deviation parameters of classical Bollinger Band analysis.
While the red Bollinger Bands in this chart use rolling standard deviations of 50 days of raw HUI price data, the blue line, these +/-2.5 SD bands are keyed off of the 50dma itself, the white line. Thus these Bollinger Bands form an expanding pipe around the 50dma when volatility is rising and a contracting pipe centered on the 50dma when volatility is waning. The oscillation of actual HUI index levels within these long-term Bollinger Bands is quite fascinating.
These HUI Bollinger Bands enable us to quickly process a huge pile of complex statistical data visually without even taxing our brains! Statistically the HUI ought to be meandering between these two two-and-a-half-standard-deviation bands about 99% of the time, so when it manages to sprint outside this SD-band pipe we immediately know that a fairly rare and noteworthy event is transpiring. Speculators can immediately see whether the HUI is relatively high or relatively low at any given moment by checking its position within its Bollinger Band trend pipe.
As this chart clearly reveals, major short-term turning points for the HUI generally coincided with the kissing or piercing of one of its own Bollinger Bands. The red arrows highlight some of the most important short-term trend changes within the gold-stock bull to date, key moments in time when speculators who made the right decisions were richly rewarded by the markets.
While Bollinger Bands alone are not designed to trigger primary trading signals, the relative position of the HUI within its Bollinger Bands helps speculators game the probability of a short-term turn in the key gold-stock index. The farther that the HUI manages to migrate away from its 50dma in standard-deviation terms, the higher the probability that this particular extreme anomaly will not last long. All of the major turning points above were marked by moments where SD-band probabilities were extremely favorable for a tradable short-term trend change.
Interestingly however, there seems to be great asymmetry in the relative value to speculators of knowing that the HUI happens to be approaching its lower Bollinger Band as compared to its upper one. While the HUI has hugged its upper Bollinger Band for months at a time when it is really running hard, obscuring a specific SD sell signal, whenever it managed to bounce off of its bottom SD band a fantastic buying opportunity existed.
As the lower red arrows in the chart above show, the greatest rallies within our entire gold-stock bull to date have all launched from -2.5 SD or so levels on the HUI. When gold stocks become so beaten down on short-term pessimism and fear that they trade two-and-a-half standard deviations or more under their 50-day moving average, odds are that this anomalously dark situation won’t last long. When speculators witness the HUI hitting its lower -2.5 SD band again in the future, it probably will mark another fantastic time to deploy fresh new gold-stock speculative positions. Watch for this event!
On the other hand, the HUI has managed to hug its upper +2.5 SD line for considerable periods of time during its major rallies. These episodes are shaded in red above, and they help to illustrate just how difficult it is to call a short-term gold-stock interim top using Bollinger Bands alone. Yes, the HUI did tend to top at SD extremes, but each interim top happened after a long series of challenges of the upper Bollinger Band by the HUI.
In a strong bull market, speculators have no way of knowing in real-time whether a particular challenge of the upper Bollinger Band is the one that will mark an interim top or whether the rally will still keep powering higher. As such, it is not a good idea to liquidate short-term gold-stock speculations on upper SD-band extremes alone. Bull markets always tend to surprise and amaze to the upside, and they can run strong for quite awhile at 2.5 standard deviations above their 50dmas.
From a practical real-world speculation standpoint, these observations lead to some important rules for gold-stock speculators to consider when incorporating Bollinger Bands into their own technical analyses.
First, the HUI’s relative position within its Bollinger Bands tells us for certain whether the HUI is relatively cheap or relatively dear, but not much more. Your odds of launching successful gold-stock speculations are highest when you deploy new capital when the HUI is near its lower SD band and lowest if you buy when the HUI is approaching the top of its SD trend pipe. If you want a high-probability-for-success gold-stock trade, patiently wait to deploy until the HUI is challenging its lower Bollinger Band.
Second, in secular bull markets, lower Bollinger Band kisses or breakouts are far more useful and tradable than upper Bollinger Band ones. Most of the greatest rallies to date in the HUI erupted soon after the index traded more than two-and-half standard deviations under its 50dma. Next time that you witness the HUI 2.5+ SDs below its 50dma, which is about 180 today, odds are that it will be an excellent time to deploy fresh capital in gold stocks, especially if key other technical indicators confirm.
Third, if you are looking for a specific short-term interim topping signal within a secular bull market, Bollinger Bands are not the right tool for the job. The HUI has hugged its upper Bollinger Band with great and long-lasting affection for much of its bull market to date. While you don’t want to deploy new capital when the HUI is challenging its upper SD band, neither do you want to sell outright though.
I have long recommended that my speculator clients, when chasing a bull like this, rely on increasingly tight stop losses to back them out of trading positions sans emotions. If you have speculative gold stocks that you are holding for trading purposes, and the HUI gets up to two-and-half standard deviations or higher above its 50dma, then you ought to consider ratcheting up your trailing stop losses.
Prudent stop losses enable you to ride individual gold uplegs as long as possible, while vastly limiting the intense emotional challenges involved in selling. If your stops are not yet hit, you can stay long even when the HUI and gold stocks look technically overvalued. But once your stops are hit and the markets close out your positions for you, remain on the sidelines and patiently wait for another high-probability buying opportunity. There is no hurry!
One of the greatest things about the financial markets is that stellar speculation opportunities are always coming and going. Never fear missing a given deployment window to launch short-term speculations, since more are always coming in the future. Being in a “hurry” to speculate often leads to losses, so just be patient and wait for a great opportunity like a HUI hovering over two-and-a-half standard deviations below its 50-day moving average.
Before we move on, there is one more important point to note in the chart above. So far in our gold-stock bull to date, each major HUI rally was officially over and a downleg upon us when the HUI broke decisively below its white 50dma after each major upleg. White arrows above mark these episodes in the past couple years.
Today, the HUI is hovering right above its 50dma again, which is around 225 at the moment. If the HUI breaks significantly below its 50dma, odds are that it will then head all the way back down to its 200dma just under 175. If you want to look for a signal that this awesome gold-stock upleg of 2003 is finally embarking upon its inevitable healthy bull-market pullback, then watch the HUI relative to its own 50dma very closely in the weeks ahead.
Our next graph takes a shorter-term perspective, showing the HUI within its 20dma +/-2 SD Bollinger Bands. Once again though, highlighting the endlessly fascinating fractal nature of the financial markets, the “rules” at this short-term scale are virtually identical to what we observed above.
Naturally this much shorter 20dma Bollinger Band trend pipe is far more volatile and chaotic than the original 50dma one above, but it underlines the very same lessons. While great buying opportunities in the HUI universally happened when it challenged its lower 20dma Bollinger Band, times to sell were ambiguous since the HUI often hugged its upper +2 SD band.
A speculator employing 20dma SD bands in his HUI trading has an excellent opportunity for big long profits if he deploys gold-stock speculative capital only when the HUI is beaten down to two or more standard deviations under its 20dma. Conversely, when the HUI rebounds and launches back up towards its upper band, it is a good time to raise your stops and let the markets close out your speculative positions for you.
While odds are that any important HUI pullback will trigger off of a +2 SD extreme anomaly, it is always hard to tell whether it will be today’s +2 SD high or next week’s or next month’s. The yellow-shaded zones above show the large areas where the galloping HUI traded in lockstep with its upper Bollinger Band. As such, just as I noted above, the HUI Bollinger Bands are far more useful for discerning great moments to buy rather than attempting to divine the right time to sell into short-term interim tops.
Now zooming in and combining our first two graphs for the 2003-to-date period alone, the interactions of both long-term and intermediate term Bollinger Bands and the HUI offer even more insights into the practical utility of these standard-deviation bands. While not primary trading indicators like the Relative HUI, the Bollinger Bands are able to do a fantastic job of illuminating the balance of probabilities in favor of or against an imminent and tradable short-term trend change.
Tying it all together, watching the HUI relative to its long and intermediate-term Bollinger Bands in 2003 would have proven quite profitable for gold-stock speculators. Let’s run through a 2003 play-by-play analysis…
In January, the HUI hit its short-term 20dma +2 SD Bollinger Band, letting speculators know that it was relatively richly valued from an intermediate-term perspective. Speculators know that an upper hit is not necessarily a strong sell signal in a powerful bull market, so they would have merely raised their trailing stop losses and upped their vigilance in January.
And if they weren’t stopped out earlier, when the HUI broke decisively below its own 50dma in February that was a hard technical signal that the previous upleg was most probably over. With the HUI drifting lower and lower towards both of its lower Bollinger Bands, prudent speculators could have bought when it traded under its 200dma, as our clients and we did, or they could have waited until the HUI bounced off of its lower Bollinger Bands. This mid-March time frame proved to be close to the ideal moment to buy gold-stock speculations for the HUI’s fantastic 2003 bull-market upleg!
Once probabilities favored the long side again by mid-March, the HUI began to gradually climb higher, delighting speculators. By June it had pierced its 20dma +2 SD line, but couldn’t quite manage to break above its 50dma +2.5 SD line. Either way, this was a caution signal especially over the shorter-term, to make sure that one’s stop losses were properly set and ready for a pullback.
The pullback did indeed come, but it was fairly minor. It bounced off of its lower 20dma Bollinger Band in the middle of July. Even more importantly however, the HUI’s 50dma held. In technical analysis, it is useful to think in terms of The 2% Rule. If you are watching a key support level, like the HUI’s 50dma, don’t consider it “decisively” broken until the HUI closes more than 2% under it. This 2% rule helps filter out meaningless short-term market noise and enabled speculators to hold in July even while the HUI’s 50dma was being actively challenged.
From July to September, the HUI soared! It was one heck of a ride and everyone long gold stocks was blessed with some fantastic profits. In late September the HUI bounced off of its 20dma +2 SD band and headed south, but its pullback ended at its key 50dma which also happened to be its lower 20dma -2 SD Bollinger Band at the time. This pullback was sharp, and did stop us out of some of our own positions, but the 50dma held so this major gold-stock upleg was still technically alive.
From its early October rebound, the HUI soared until late November. It has pulled back sharply again in recent weeks since then. While it is probably not wise to use the upper Bollinger Bands alone as primary sell signals, our other primary technical indicators like the Gold 50/200 MACD and Relative HUI did and do suggest that a short-term pullback in gold stocks back down towards their 200dmas is highly probable. But the real acid test for the longevity of our current upleg is yet to come.
If the HUI does not bounce at its 50dma, around 225 today, and it trades decisively under this white line, then odds are that our current gold-stock upleg is over. Every single time in our awesome young gold-stock bull to date when the HUI’s 50dma decisively fell, it was on a major pullback back down towards the index’s 200dma.
Only time will tell whether the HUI’s 50dma will hold again as support in the weeks ahead or not, but if it doesn’t then the good news is that a fantastic new buying opportunity is rapidly approaching, when the HUI careens back down towards its 200dma and its lower 50dma -2.5 SD Bollinger Band. We haven’t witnessed a similar magnificent buying opportunity since March so I am certainly looking forward to seeing these glorious technical gold-stock tidings once again.
If speculating in gold stocks is your game and you are interested in which potential stocks to consider buying for the next major gold-stock upleg after the pullback, you may wish to check out the current December issue of our acclaimed Zeal Intelligence monthly newsletter. In this issue, titled “Blue-Chip Golds”, we ran all of the elite gold stocks of the HUI and XAU through technical screens to find the best-performing and most-promising candidates for serious fundamental analysis.
In the upcoming January issue for our subscribers, I intend to continue these studies and farther advance our technical and fundamental preparation for the next major gold-stock upleg. As always, new subscribers will receive a complimentary electronic copy of the current December issue of ZI and your subscription will start with January’s dispatch. Please join us today!
Bollinger Bands, even though they are not necessarily designed to be primary trading signals, are very valuable to gold-stock speculators. They help us to discern in real-time when the HUI happens to be priced relatively high or low, excellent information to possess.
The HUI Bollinger Bands also help us to understand when stellar buying opportunities exist when the HUI is crowding its lower standard-deviation bands. Since buying low is a crucial half of this whole speculation ballgame, gold-stock speculators really ought to pay attention to the HUI Bollinger Bands.
Adam Hamilton, CPA December 19, 2003 Subscribe at www.zealllc.com/subscribe.htm