Monday, August 29, 2016

Dow Theory Special Issue: Putting the Dow Theory under Stress-Test (IV)

What happens to the Dow Theory when the market enters into “fibrillation” (Narrow ranges)?

We continue our saga of looking at the Dow Theory from all possible angles (specially the stressful ones).

Narrow ranges are the plague to trend followers. While the Dow Theory by integrating the time and extent element is less prone than breakout systems or moving averages to false signals, small losses are likely to occur when ranges occur. It the lack of any meaningful trend (be it up or down) persists for a long time (i.e. 4 years) it is likely (not a certainty, though) that a succession of small losing trades will occur. Would this unusual market situation result in the proverbial death by one thousand cuts? Would a succession of small (or not so small) losing trades result in a drawdown exceeding 50% (as buy and hold is used to). In 110 years of Dow Theory history this situation has never occurred, and the occasional periods of narrow ranging have not lasted enough to decimate Dow Theorists. In other words, as far as I have been able to observe, the draw downs caused by the cumulative effect of losing trades have never been worse than those of buy and hold. More specifically, it is far from even 50%. If my observations don’t betray me, the largest cumulative loss underwent by Schannep’s Dow Theory amounted to ca.18.5% due to two consecutive losing trades on June 20, 2008 and January 30, 2009. It is worth noting that the largest cumulative loss was caused not by “narrow ranging” environments but by a powerful cyclical (or even secular) primary bear market. It is also worth stressing that under Schannep’s DowTheory, the largest losing trade amounted to -10.45% (Jan 30, 2009), which is far from being lethal to the investor. Once again, it bears repeating that the largest losing trade was not caused by a trading range.

Thus, it is unlikely that a persistent narrow range condition would result in deep drawdowns. Historically seen narrow ranges have not lasted enough to cause a high number of losing trades resulting in deep drawdowns. However, the market might not be so lenient for us in the future, and hence, we've to imagine scenarios of protracted trading ranges.

How "narrow" is a narrow range to be harmful? Answer: At least a 3% range from closing lows to closing highs confirmed by at least two indices. Less than this does not generate signals, as the minimum threshold of 3% is not reached. On the other hand, swings exceeding 15- 16% between confirmed closing highs and lows are likely to result in profits or, at least, very well contained and sparse losses.

Thus, we can put some tentative figures to the market environment which is detrimental to the Dow Theory: Narrow ranges between 3% and 15-16%. Less than 3% results in no change of trend (and hence no signal, no trade), and more than 16% is a swing of sufficient magnitude to lock in profits in most instances. We have to add the time element to the extent (percentage) element. Narrow ranges of very short duration will not develop secondary reactions, and hence will not lead to the set up for bull/bear market signals. Our likely loss is not linear, and is not proportional to the magnitude of the narrow range. This is a vital point. In other words, a narrow range of 3% per definition is not likely result in losses exceeding that range (unless a sudden event results in a macro down day with a close well below the lows of the secondary reaction). The same applies to a range of 4 or 5% where our likely loss is contained by the confines of the ranges. However, things change as we approach the 7-8% range level. Ranges exceeding 7-8% are likely to result in smaller losses. Thus, a range of 11% is likely to result in a smaller loss than a range of 6%. This is good news, since, even under general adverse market conditions (ranges from 3% to ca. 15-16%) most of the ranges are marginally harmful. If they were too narrow because the loss would be small in direct proportion to its “narrowness”, if they were broader, because we'd begin to have sufficient magnitude for the Dow Theory to extract some profits (or have very small losses). Please now follow me and take your time to digest.

We know  (see here why) that ca. 7-8% is lost when buying and a similar figure when exiting. The Dow Theory (no trend following system for this matter) is not able to catch the absolute top or bottom. Thus, from a given “swing”, we “lose” ca.- 15-16%. Accordingly, our minimum swing from closing high to low, containing a buy and a sell signal, just to break even, should be ca. 16%. A larger price advance between bottom and top is likely to result in profits and a smaller one is likely to cause losses. Why I say likely? Because, we know that our entries and exits are at ca. (ca. “circa” means approximately, around, from this Latin word comes the word “circle”), and hence “ca.”, “around”, imply no certainty. We have seen instances where the buy or sell signal were signaled very close to the top or bottom (i.e. at 5%) whereas in others it has been signaled at 11%. Thus, I am not talking about mathematical certainties. The market will do what it wants to do. However, I try to provide myself and my readers with a framework to properly appraise what is a “worst case” scenario under listless markets. 

Thus, very narrow ranges of 3-4% which manage (a very rare occurrence) to flash buy and sell signals, will result in losses in the vicinity of 3-4%. If our range approaches 6-7% our likely loss will be nearing 6-7%. 

Look at the table below.

Or look at the chart below. The horizontal axis displays the extent of the swing from bottom to top. If the range does not exceed 3%, there is no secondary reaction, no signal, and hence no loss (as not trades are taken). Once we exceed the 3% threshold, you can see that the larger the range the larger the likely loss until the range reaches enough magnitude to first reduce the amount of the likely loss and finally result in profits. 

Profit and Loss profile of Dow Theory trades according to the extent of the range (swing)

While reality should not match my model, “on average” it is clear that the extent of each range tends to favor a determined risk and loss profile.  The specific profit or loss of each trade will depend on the extent of the secondary reaction which led to the setup of the buy/sell signal.

On the other hand, once the ranges exceed 7-8% and given that we know than on average  the buy and sell signals tend to “eat up” 7-8% when entering and exiting. We know that it is likely that broader ranges will result in smaller losses. Thus, a 12% range with an average entry and exit implies that ca. 15% is going to be lost when entering and exiting the trade. However, 15%-12% equals 3% which is the loss likely to be made on the trade, which is less than the loss made with a range of just 6%.

When we talk about narrow ranges, we should clarify what we mean by “narrow range”. Thus the last 1 ½ year can be characterized as a narrow range, and it certainly affects the outcomes of our Dow Theory trades. However, we should be more specific. We should zero in on the specific range pertaining to each “buy/sell” pair. On any given Dow Theory trade (buy/sell signal) the success or failure of our trade will be largely influenced by the amplitude of the swing which contains the bottom of the preceding bear market and the top of the bull market. By “swing” I mean a movement not interrupted by a secondary reaction. The larger the swing, the more likely we will lock in profits. The narrower the swing (trading range), the more likely our trades will result in losses.

As an aside, Classical/Rhea Dow Theory is less prone to such whipsaws as it is less predisposed to signaling secondary reactions, and hence less prone to setups and signals (but then, it has a larger average losing trade. Everything comes at a price). However, in past experience, (see here) the slightly higher number of trades resulting from Schannep’S Dow Theory did not result in degrading performance; not even the number of losing trades. Just on the contrary: ca. +2% p.a. outperformance versus the classical Dow Theory and drastically reducing average and largest losing trades. However, the price structure of the past (aka: the shape of the charts) may be different in the future.

(to be continued in next post where we will draw specific conclusions as to what to expect under narrow ranges, and more specifically “fibrillation” –repeated narrow ranges in succession-)

The Dow Theorist

Saturday, August 20, 2016

Dow Theory Update for August 20: Recent declines in Gold and Silver (and their ETF miners) do not qualify as a secondary reaction yet

Future important post

When I find time (oh time!!!), I’d like to write about the market conditions where the Dow Theory (and specially Schannep’s Dow Theory) might “suffer” and underperform buy and hold. In previous studies (see here), I have shown that the Dow Theory tends to underperform buy and hold when markets display nice up trends (of course, if the market only goes up, nothing is better than buy and hold) whereas most of the outperformance is predicated on declining markets (where buy and hold losses big, and the Dow Theory cut losses short, and opportunistically buys at a much lower "post crash" price). However, I'd like to further deepen this issue, as there is a third market condition which (while not fatal) may result in temporary underperformance for the Dow Theory and demoralize does not really acquainted with the “guts” and inner workings of the Dow Theory.

Of course this “third market condition” afflicts all Trend following strategies. Furthermore, the Dow Theory is very well endowed to ride through the adverse condition quite successfully, as “underperformance” does not mean losing one’s shirt (as does buy and hold when the market crashes).


The primary and secondary trend is bullish as explained here and here


The primary trend is bullish (Dow Theory signal of March 17th, 2016), as reported here and here.

The secondary trend is also bullish as explained here

Recent declines in SLV and GLD do not qualify as a secondary reaction. The time extent has not been met (GLD just declined 7 trading days and currently stands above the last minor recorded closing low July 20th).

Find below an updated chart which displays all price action since the primary bull market signal of March 17th, 2016 (left side of the chart). As you can see there was a secondary (bearish) reaction (red rectangle in the middle of the chart) against the primary bull market which was resolved by confirmed higher highs. Thus, the primary trend is bullish. The current decline (small blue rectangles on the right side of the chart) are a mere minor decline. 

The current decline does not qualify as a secondary reaction


The primary and secondary trend is bullish as explained here, and more recently here

SIL and GDX have been relentlessly making higher highs. Recent declines do not qualify as a secondary reaction. The chart below displays SIL and GDX price action since the primary bull market signal of March 3rd, 2016. As you can see hitherto no secondary reaction has occurred yet. The small blue rectangles on the right side of the chart display the ongoing decline, which, I have said, does not qualify as a secondary reaction against the primary bear market. 

Powerful primary bull market swing: A secondary reaction has not occured yet

The Dow Theorist

Monday, August 15, 2016

Dow Theory Update for August 15: More on the primary bull market signal of August 11, 2016

Trends in precious metals universe unchanged

As it was reported here, last August 11th a primary bull market was signaled according to Schannep’s Dow Theory.

Here you have un updated chart.

Schannep's Dow Theory signaled primary bull market on August 11

It is important to note that the “Rhea/Classical” Dow Theory has not signaled a primary bull market yet. Why? Because the Transports remain below their July 14th secondary reaction closing highs. Since the “classical” Dow Theory only uses the Industrials and Transports, the failure of the Transports to better its secondary reaction highs implies no signal. Of course, it could be argued that Schannep’s Dow Theory is “too reactive”.  If past performance is to serve us as a guide, the “over reactiveness” of Schannep’s Dow Theory clearly outperformed the “Rhea/classical” Dow Theory. Both in terms of average performance (ca. +2% p.a.) and drawdown reduction. More about the comparison between both Dow Theory flavors here.

Thus, like in short term trading. We don’t know in advance the outcome of any given trade. This specific one might be a false signal whereas the classical Dow Theory by not signaling a primary bull market would look "smarter" this time. However, we have to look at the long term record, which clearly favors Schannep’s Dow Theory. Of course, it could be argued that past performance is no guarantee of future performance, and, hence, that the “Rhea/classical” Dow Theory is likely to outperform Schannep’s Dow Theory in the future. This could happen, but, as I wrote here, I find it very unlikely:

“I’d say that the very structure of Schannep’s Dow Theory is likely to continue outperforming the “classical” Dow Theory in the future. By design Schannep’s Dow Theory is more responsive (i.e. detects earlier the onset of a new trend) because:
1)     It uses three indices (it includes the S&P), instead of just two, whereas it requires just two indices confirming. 
2)  The definition of secondary reaction (which is vital to determine breakout points, which in turn, define primary and bear market signals) is “shortened” as, 10 days, or even less, is enough to qualify a movement as a secondary reaction. 
3)     In the same vein, by doing away with the 1/3 to 2/3 retracement rule (which is one of the requirements under classical Dow Theory for a secondary reaction to exist) and merely requiring a 3% move, many movements which under classical Dow Theory escape the definition of secondary reaction are labeled as such under Schannep’s Dow Theory. 
Thus, the very make up of Schannep’s Dow Theory makes it foreseeable that in the future it will continue to cut losses short (that is detecting earlier changes of the primary trend) because its very rules are designed to spot secondary reactions in an early fashion.
Of course, critics could say that everything comes at a price, and that premature labeling of secondary reactions makes Schannep’s Dow Theory prone to false signals or worse yet, reduced profits. 
However, such objections have been exhaustively debunked during this “face off” saga. Schannep’s rules, while being “early” both in signaling entries and exits, manages to increase profits, not reduce them.”

Therefore, it is perfectly possible for this specific trade to be a loser which would make look the classical Dow Theory “wiser” for having avoided this specific trade. However, I am focused on the long term record.

Of course, if the current signal is a valid one (by "valid" I mean that the trade is closed at a profit), then the “Rhea/classical” Dow Theory would look “dumb” as the primary bull signal would come later, and hence signaling an entry at a higher price.

The current signal offers a moderately bad risk reward ratio. The entry price for the SPY was 218.65 (close of August 11). Since the primary bear market closing lows of June 27th (199.60) the SPY has rallied 9.54%. Hence, our initial stop loss (more about the Dow Theory stop loss here) lies at -0.087% (This is the percentage decline from 218.65 to reach 199.60). Thus, the entry comes at a slighter higher level than the average entry which tends to be at ca. 7.4% of the primary bear market lows. On the other hand, and while nobody can predict the extent or duration of a primary bull market (Rhea dixit), I feel quite leery as to the “old-age” of current bull market cycle, which seems to reduce the “reward” (extent) side of the risk/reward ratio. Furthermore, the primary trend, as determined by weekly bars, remains bearish. Thus, the misgivings I expressed concerning the preceding primary bear market signal of April 13 remain fully applicable.

However, as I wrote on April 14th, 2016, the fact that I am leery about the signal does not imply that I will override it. Nonetheless, what works for me does not necessarily work for others. Each investor should do his own homeworks.

The primary trend is bullish (Dow Theory signal of March 17th, 2016), as reported here and here.
The secondary trend is also bullish as explained here

The primary and secondary trend is bullish as explained here, and more recently here


The Dow Theorist