What happens when one mixes a great timing indicator like the Dow Theory for the 21st Century (DT21C) with a relative strength system? Answer: Even more outperformance and a better risk-reward profile v. Buy and Hold (B&H).
Relative Strength (RS) is based on the proven fact that assets (stocks, indexes, sectors, commodities, etc.) that have displayed strength relative to other peers should continue to be relatively stronger in the future. RS is not trend-following, as one relatively stronger asset may be falling, albeit less than its coequals. This is the weak point of RS investing. RS tends to outperform in Bull markets but gets decimated in Bear markets. For many RS systems, the risk-adjusted profile is not better than Buy & Hold due to large drawdowns during Bear markets.
What if we could get the upside of RS while getting rid of the downside? If we used RS only in Bull markets and exiting stocks during downturns, we would be getting the best of both worlds: Catching the RS’s outperformance in good times while avoiding the bad times.
We know that the DT21C is one of the best trend-following devices. This Dow Theory blog provided ample evidence of the net superiority of the DT21C over another trend following systems.
Thus, if we applied an RS strategy to equities only when the DT21C is in a BUY mode, we would have greater odds of achieving significant outperformance, as we’d have two sources of outperformance: the DT21C, and additionally, the one derived from the RS strategy.
However, there are many RS strategies.
Some are good and robust. Some are not. After carefully researching the issue
since 2015 and testing the waters with my own accounts, I came to what I
consider one of the “best” and most robust RS systems. The test starts on 1/06/01 and finishes on 12/31/22, ca. 21 ½ years. Take
a look at the key charts and figures:
Below there is the performance breakdown for each year from 2001 to 2021:
Key performance figures:
I calculated the profit factor (PF, total profits made divided by total losses, see row “K” above) by computing the yearly percentage changes for each strategy and adding the total percentages of points won (row “H”) or lost (row “I”). We see that the DT21C almost doubled the PF of B&H (6.12 v. 3.63). The DT21C+RS scored an astounding PF of 22.33, which means that we made much more than we lost, and, more importantly, a high degree of accuracy at spotting trends. We can also observe that by using a good trend filter (the DT21C), we get rid of the drawback plaguing RS systems: huge drawdowns during bear markets. If you look at the “Total Loss” cells (row “I” above), you’ll see that the DT21C has a smaller total loss (as expected in good trend-following) but the DT21C+RS had an even more minuscule total loss of -12.09% percentage points over the last 21 years. In other words, by mixing RS with the DT21C, we increased performance, but we did not increase drawdowns. Given its propensity to significant drawdowns during bear markets, this is quite a feat when dealing with an RS strategy.
One approximation to a risk-adjusted measure of performance is to divide the average annual performance (row “A” above) by the annual standard deviation of such performance (row “B”). B&H scored a decent 0.51 (it’s been a secular bull market after all), the DT21C had a better reading of 0.66 (it was better on both counts: more performance and less volatility). The DT21C+RS system scored an exceptional 0.88 (row “C”), which implies that we are increasing performance and decreasing risk. Another way to measure whether we are doing an excellent job at keeping drawdowns at bay is to subtract one standard deviation of the annual returns from the average annual returns themselves (row “D”): The less negative the number, the lower the odds of enduring a big drawdown in the future.
Row “E” shows the Compounded Annual Growth Rate (CAGR). As expected, it is higher for the DT21C and much higher for the DT21C+RS. Please remember that CACR gives us a more accurate performance measure than the average annual return (row “A”). However, measuring the average yearly return is helpful in order to relate it to its standard deviation to appraise risk.
The DT21C+RS system only had two modest negative years (row “F”). B&H had four negative years. The worst annual performance (“Max Loser,” row “G”) was -38.48% for B&H, a very decent -13.54% for the DT21C, and an outstanding -8.30% for the DT21+RS. So despite being an RS strategy, we managed to reduce drawdowns significantly thanks to the use of the DT21C trend filter.
The correlation coefficient of the annual performance of B&H v. The DT21C is 0.789. Given that the DT21C was tested on the S&P500, we see that the DT21C is not so correlated to B&H, which makes itself evident by slightly underperforming when the market is strong and overperforming when the market falls. The correlation between the DT21C+RS and B&H is more interesting, dropping to a more modest 0.618. Thus, adding the RS element to the DT21C provides an additional layer of diversification. We are achieving more outperformance and decoupling ourselves from the vagaries of B&H.
Watch out! There is no holy grail in the markets. We are not going to have the wind at our backs all the time. While over the long term, DT21C+RS beats the pants of B&H, there are rough patches that we have to overcome psychologically. Out of 21 years tested, DT21C+RS outperformed B&H for a total of eleven years. However, B&H defeated the DT21C+RS ten years. So there is an almost 50% chance that in any given year, we will be underperforming. Furthermore, there have been two episodes of 3-years-in-a-row of underperformance. The good news is that underperformance is not a lack of performance, and our DT21C-RS system fared well during the underperforming years but less than B&H.
The chart below shows the cumulated outperformance of the DT21C+RS versus B&H:
As you can see, the uptrend of outperformance is clear. Furthermore, you may observe that the periods of underperformance (blue line going down) are short-lived and moderate. In other words, the DT21C+RS will spend less time underperforming B&H than the DT21C or RS alone. This insight is vital as time underperforming B&H wears out the investor’s patience. The shorter we keep the time underperforming B&H, the more likely we will stick to our trend-following system in real life.
Well, now it's time to open the black box and discuss the rules of our Relative Strength + DT21C strategy, namely:
1. Selecting our universe (find the right assets to trade).
2. How to select the highest relative strength assets.
3. How to rank.
4. Buy, Sell, and rotation rules.
5. How do we avoid the risk of a crowded trade.
6. Suggested alternatives to improve performance further.
.................................... ............................ .......................................
Do you want to continue reading?
Do you want to know the rules of our top outperforming DT21C+Relative Strength system, and learn to trade it?
Subscriber, and you’ll have access to the full Special Report. Furthermore,
we will be informing our Subscribers of the trading candidates we will be selecting according to
our new strategy.
Additionally, your subscription will give you a wealth of information (i.e., access to our Letters since 1962 and their accompanying trade recommendations, the power of the consumer confidence report as a timing device, the special report about the yield curve, how to calculate profit objectives that work, trend assessment for gold, silver and their ETF miners -GDX & SIL- and much more). More importantly, you’ll be punctually updated through our email service of any change in the trends for US stock indexes. Not accidentally, our Newsletter has consistently been ranked among the top investments Letters.
What are you waiting for? Our new strategy scored 622.45 additional percentage points than Buy & Hold in 21 years. It can be rightly named Dow Theory on steroids.
Editor of thedowtheory.com