Friday, August 9, 2013

Face off: Schannep versus “classical” Dow Theory



Part II. Overall performance figures

Today we will continue our comparison of the “Rhea/classical” Dow Theory versus the Schannep’s version thereof. Our previous post, which you can find here, set out the premises of our study, as it is important to do a real “apple to apples” comparison.

Today, we will begin to evaluate the transactions taken in pursuance of Schannep’s Dow Theory, and we will compare them with the transactions taken in accordance with the “Rhea/classical” Dow Theory.

Our study starts in 1954 and finishes in 2013. We took 1954 as our starting date, as Schannep’s flavor goes back to this date. You can find Schannep’s Dow Theory record here.
 
With no more preambles let’s get started.

Schannep versus Traditional Dow Theory vital statistics

Total number of transactions (trades) taken:

Schannep:32
Classical:24

Comment:

We observe that Schannep’s Dow Theory resulted in ca. 1/3 more transactions than the classical Dow Theory.

This higher figure is in itself neutral. It can be good if such more frequent trading results in decreasing risk (drawdowns) and increasing profits, or it can be bad if it fails to achieve that goal.

However, I can make two observations:

1.     In spite of being ca. 1/3 more reactive to market conditions, Schannep’s version has nothing to do with short term trading. 32 transactions in almost 60 years bears no resemblance to short term trading, and hence we can infer that commissions and slippage should not plague Schannep’s performance.

2.     My investor and short term trader experience has proven me beyond any shade of doubt that the only way to avoid deep drawdowns when a bear market hits is by increasing the number of trades (i.e. by getting in and out quickly). Of course, overtrading can decimate your trading account. Thus, it is necessary to strike a balance between overtrading and overstaying a falling market. I find that the Schannep’s version of the Dow Theory strikes a very good balance between inaction and frantic activity.

Average duration of each transaction:

Schannep: 479 days.
Traditional: 629 days.

Let’s take a look at the median duration:

Schannep: 357 days.
Traditional: 403 days.

So we can see that on average transactions taken in pursuance of Schannep’s rules last ca. 23.8% less than those taken according to the traditional Dow Theory.

Once again, a shorter life-span for each transaction isn’t necessary good or evil, in itself. Thus, even though each transaction has a shorter life-span, we also know that there are more transactions when following Schannep.

So we must be patient and wait for more data, such as

Average gain in each transaction

Schannep: 22.07%
Traditional: 26.36%

Here I can see Schannep’s detractors shouting loudly: “You see Schannep’s Dow Theory severely underperforms the classical Dow Theory."

I’d say to them: “not so fast."

A more accurate measure is average gain per annum. Thus, I’d rather prefer to make 10% per annum rather than making 30% in ten years. Thus, we have to normalize returns with time.

Since Schannep’s transactions last significantly less (almost 24% less) than those taken according to the traditional Dow Theory, annualized performance would be:

Schannep: 16.8%
Traditional: 15.29%

So things now start to look much better for Schannep. This statistic is telling us that Schannep’s signals “extract” more profits from the market in less time, or, in other words, given equal time fully invested in the market, Schannep’s rules manage to make more money. This is a clear measure of efficiency. Or in plain English: Schannep’s rules do a better job at separating a true signal from noise.

Inquisitive readers should be thinking right now:

“This statistic is a little bit misguiding, as we really don’t know if there have been enough signals so that overall performance in the last 59 years has really been superior to that of the classical Dow Theory. What good is a more “efficient” signal if we don’t get enough of them or, worse yet, its life-span is so meager than in real life, there is not enough time to build up meaningful profits.”

Well, I am going to address this criticism.

If we take base=100 in 1954 for both Dow Theory “flavors”, such 100 would have grown into:

Growth of 100 USD since 1954

Schannep:14,973
Traditional: 8,756

So we can see that, while having shorter trades, Schannep’s rules managed to generate enough trades to make meaningful profits. Thus, Schannep’s claim that his “flavor” outperforms the classical Dow Theory by ca. 2% annual, is a correct one.

Total time in the market

Schannep:70.6
Traditional: 70.8%

Now it gets even more interesting. Even though Schannep’s transactions last less than those taken as per the classical Dow Theory rules, the total amount of time spent on the market is almost the same. This implies:

a)    That Schannep’s rules compensate shorter duration for each trade with more trades.

b)     Since total profit is much higher, the time spent on the market is best used, which implies it is better at “timing” the market.


Winning transactions (trades) versus losing trades

Schannep: 23 winners/9 losers (71.8% winners)
Traditional: 17 winners/7 losers (70.8% winners)

So we see that both Dow Theory flavors have a very similar percentage of winning trades. A high batting average, while not necessary for high profits (as low percentage systems can score good profits provided the average winning trade greatly exceeds the average losing trade), is important in real life because: (a) it increases confidence in the system; (b) drawdowns tend to be reduced as there is less likelihood of a long string of losses. Any short term trader worth his salt, knows what I am meaning right now. Here both “flavors” excel.

….

Well, little by little, we are deepening our understanding of Schannep’s Dow Theory, and, more importantly, why it is objectively better than the “classical” Dow Theory (which, in itself, is also an excellent timing system). Furthermore, the figures we are examining do not do full justice to Schannep’s Dow Theory. We will further explore this assertion when we close this saga of posts.

We will deal with more important figures and facts in the coming post, which I hope to publish soon.

Have a nice weekend.

Sincerely,

The Dow Theorist

No comments:

Post a Comment