SUGGESTED READING --> | Profiting from Advance/Declines | The Incredible McClellan Oscillator |

So far you have seen many of our highly successful long term market timing strategies at work. But many subscribers and investors were also seeking an accurate and reliable market timing system for shorter investment periods, say from 4-6 months onwards.

PowerStocks Research have developed such a system, which we have named "JSE SwissClock" due to its uncanny accuracy in the timing of JSE medium term cycles. It is based on a proprietary set of mechanical trading rules and calculations we have developed around the McClellan Oscillator and the McClellan Summation Index, which of course are derived from daily market advances and declines which we calculate on a daily basis.

SwissClock is a classic "Buy Low and Sell High" timing strategy that was vested for 58% of the time in the JSE over the last 11 years. During periods it was not in the JSE, its funds were sitting in a bank or money market account earning prime less 4%.

SwissClock is about twice as active as our other timing strategies, with a trade every 6 months on average as opposed to every 12-14 months.

JSE SwissClock has been back-tested 11 years to 1998 (since that is the oldest data we have for which we could calculate advance/decline data) and exhibited startling accuracy and performance:

  1. It made 43 trades in the last 10 years (22 buys and 21 sells)
  2. Only 2 trade resulted in losses of -7.2% and -11.3% respectively.
  3. 90.9% of the trades resulted in a profit.
  4. Only 4 trades delivered less than 10% growth
  5. Average holding period was 100 days (3 months).
  6. Over the 137 months it showed 75% positive months versus the ALSH’s 56% positive months.
  7. Its worst month was -12.67% versus the ALSH’s worst month of -29.87%.
  8. The worst monthly draw-down (ie when you have consecutive months of negative growth) was -13% versus the ALSH’s -42%.
  9. The ALSH had 13 periods of more than 1 monthly decline in succession, whereas the SwissClock strategy only had 3.
  10. The longest period of successive declines for the ALSH was 5 months, with the SwissClock strategy it was 2.
  11. Over the 44 quarters in the test period, SwissClock showed 86% positive quarters versus the ALSH’s 63%.
  12. SwissClocks' worst quarter was -10% versus the ALSH’s -39%
  13. Over the 11 years period SwissClock showed only 1 negative year versus the ALSH showing 3 negative years.
Our best performing Market Timing Strategy on the ALSH index so far this decade is our Econometric Timing Model (which did only 3 trades over the 11 years.) As we noted however, this strategy was fortunate to make excellent gains in the last bull run right up to before the 2008 crash, as it was the last strategy to hold onto the JSE before running for cover to the bank accounts.

SwissClock traded actively right through the JSE crash, and as such is the only strategy we have that is not scared to trade in bear markets. The Total Return Indices of SwissClock, versus the Econometric model and a ALSH buy-and-hold strategy are shown below:

The SwissClock (SCLOX) strategies' monthly, quarterly and yearly performance are shown below versus the Econometric (ECOMX) model and the ALSH buy and hold strategy:

KEY : AVG = period average growth, DEV = period growth standard deviation, SHARPE = Reward-to-Risk ratio (AVG/DEV), MIN = worst growth, NEG PER = negative growth periods, NEG% = % of periods negative, POS% = % of periods positive.

The chart below shows the 11 year time-line with the JSE ALSH Index, the McClellan Summation Index and the vested periods of the strategy shown in shaded boxes. Blue boxes were good to excellent trades, orange boxes were so-so trades (growth less than 5%) and red boxes were wrong-way (negative) trades. You can click on the image to download a 1200x800 version.

The transaction log for the backtest appears below, with the test period broken into two halves to ensure we can check for strategy consistency over separate 5 year periods.

Note there were only two negative periods (wrong way trades) and that only two positive trades delivered less than 10% compound annual growth rate (assuming when we invest we at least want to beat inflation). Amazingly, both periods averaged identical trade lengths of 107.4 days. The first 5 years showed better profit growth per annum per trade (54.2%) than the 2nd 5 years (30.9%) no doubt due to the losing trade on 29/07/2008 in the height of the crash. Nevertheless the individual CAGR's from the trades are highly impressive.

Whilst an active SwissClock timed trading strategy since 1998 delivered an astounding 1,299% growth, the earnings power of the strategy was effected by investment climate. Since the SUPERModel Composite Timing Signal we maintain tracks various investment climate components such as momentum, economic cycle and monetary policy environment, we back-tested a SwissClock strategy during various periods on the JSE defined by how many SuperModel sub-components were presenting a BUY signal:

We see that the SwissClock delivered its most powerful performance (90% CAGR) during the cumulative 531 days over the 2945-day test period (18% of the time) when all three SUPERModel sub-components (Repo Rate Model, Econometric Model and Trendex) were firing a BUY signal.

Now even though we know that JSE periods when no SUPERModel signals are present are notoriously weak (normally when we advise our subscribers to get out the markets), the SwissClock strategy still managed to scape up 32% CAGR (see "NO SIGNAL" above.)

The 2nd bar on the right in the above chart represents a "vanilla" SwissClock strategy that plays the markets regardless of what SUPERModel signals are present. As you can see it was in the JSE for 1,678 days (57% of the time over the last 11 years) and would have grown your funds 697% during those specific days, or 57% compounded per annum (assuming you stuck your money under a mattress when you were not in the markets.)

Now if we only played the SwissClock strategy when AT LEAST ONE of Repo Rate, Econometric or Trendex were firing a BUY signal (">=1 SIGNAL") we see we would only have been invested in the JSE for 1,313 days (44.6% of the test period) but would have received better compound growth of 67.2% during these fewer days. Similarly for a SwissClock strategy that only invested in the JSE when at LEAST TWO (">=2 SIGNALS") were firing.

When all three of Repo Rate Model, Econometric Model and Trendex Model are presenting BUY signals and the JSE market breadth assumes a posture that results in the SwissClock triggering a BUY signal, then history has shown that this occurs only 18% of the time and returns from the JSE are likely to be of a turbo-charged nature.

Note how aggressively the SwissClock strategy traded right through the 2008 market crash. In fact from the peak of 22 May till 18th June 2009 the strategy traded 3 times making 11.3% loss, 8% profit and 11% profit respectively, to still be up by 6% overall!

Most of our long-term timing strategies had bailed by June 2007 into the safety of their bank accounts, with the Econometric model being the very last to chicken out the market (hence its excellent 10 year performance)

Strictly speaking, a more conservative version of SwissClock would have bailed after the 5th last transaction on 20/11/2007 since by then severe divergence was being displayed (see chart below)between the Net Advance/Decline Line and the ALSH, telling us that a massive reversal was afoot. If you recall our research on this topic, when the Net A/D Line fails to make new highs together with the JSE then a reversal is afoot. Divergence went away at the market trough on 3 March and the strategy would have picked up from their again. This would have suited more risk-averse investors, as opposed to those brave souls that followed SwissClocks signals right through the storm.

Alternatively, by only playing SwissClock when at least one SuperModel sub-component was presenting a BUY would also have allowed you to say "Thanks but no thanks" to its fearless nature since all of Repo Rate, Trendex and Econometric timing models are rather shy of bear markets and stock market crashes. As shown in the previous section, this strategy would have been vested slightly less overall in the JSE over the 11 years (44.6% of the time versus 57% of the time) and it would have made 17 trades as opposed to 22. This more cautious SwissClock strategy would have yielded 1268% growth over the 11 years versus the vanillas' 1299%, not very much difference and certainly exposed to less stock market risk (being vested 12% less of the time). The Total Return Index (TRI) chart below shows the conservative SwissClock versus the more aggressive vanilla version.

You have seen from our ground-breaking research titled: "Using Sector Rotation and the right ETF's to Boost Returns" that by investing in certain cyclicals or ETF's with your market timing strategy you can double or even triple your returns. We suspect the same will apply with SwissClock, we see no reason why it shouldn't. So just imagine what sector rotation or the selection of the right JSE ETF would do for your returns with a SwissClock strategy! Nevertheless we take nothing for granted and have scheduled the appropriate sector backtests which we will publish for our paid subscribers next month (these multi-sector tests are complex and time consuming.)

SwissClock weekly timing signals will be shown on the Weekly JSE Pulse (WJP) from 3 August 2009 (paid subscribers only, go here for subscription details). A sample is shown below.


For existing subscribers that want real-time signals emailed to them to allow them to get in early on trades, an additional cost of R200.00 per annum will be required. Email with SWISSCLOCK in the subject to top-up your subscription to receive the real time alerts.
Make a Free Website with Yola.