The SP500 Market Timing project

Currently, RecessionALERT subscribers deploy the Recession Forecasting Ensemble (RFE) to match  their equity allocation to recession risk. But the assumption in our RFE research note was that you would simply “buy and hold” the stock market during periods of low recession risk. This provided for far superior risk-adjusted returns across many business cycles, but you were unlikely to beat the buy & hold strategy during bull markets.  The aim of the SP500 Market Timing Project was to figure if there was a way to comprehensively beat the SP500 buy & hold strategy during bull markets, without the use of leverage, using an end-of-day system.

The research project deployed the “BUY POINT” Market Timing Signals in the Weekly Sentiment Report to assess favourable times to be getting into the market (or “buying the dip”). The model would then construct a trailing stop to place a floor under the trade and minimise possible losses. The model from time to time determines that stock market momentum is faltering and will issue an instruction to offload 50% of the trade (if we are in profit) or abandon the trade (if we are in a loss.) In addition, critically over-bought conditions will also result in the model closing the entire trade even if we are above the trailing stop. Finally, if the SP-500 should close below the trailing stop, we close all positions. It is that simple.

The trading history since this bull market commenced in 2009 appears below (click for larger image)

It is not possible to examine the above chart “by eye” to determine if this is a good strategy or not. For this we need a robust statistical analysis. The most important is the general equity curve profile of the timing model versus that of the “buy & hold” strategy. This is displayed below and shows the timing model returned 203% versus the “buy & hold” that returned 147% (up to 21 May 2013). The timing model delivered 1.38 times the return of the “buy & hold.”

Greater returns are not always the objective of all timing models. Sometimes one is seeking out better “risk adjusted returns”. This is highlighted by the draw-down chart below which traces out 100-day daily draw-downs for the “buy & hold” and the timing model equity curves. What was not obvious from the first chart is very obvious below – the timing model suffers dramatically less draw-drowns than the “buy & hold”. Not only does the timing model equity curve spend far less time in draw-down territory, but the maximum draw-down is 7% versus a gut-wrenching 19% for the “buy & hold.” Market timers often refer to this chart as the “Ulcer Index” and it is clear there are far less and far smaller ulcers likely to be developed with the timing model than the “buy and hold.”

We now examine the specific trade statistics of the model. We have 19 trades (roughly 4 per year or one per quarter) of which 63% were winners. The winning trades averaged 11.1% and the losing trades averaged losses of 1.72% Total percentage points accumulated by the model was 133.33 versus 12.03 points that were lost in the losing trades, giving us a gain-to-loss ratio of 11.09. This is an exceptionally high number. The industry metric for a half-decent trading system is that it should have a risk/reward ratio of at least 3, 4 or 5 if it is a very good system. We have designed and developed and run hundreds of index trading models and can say this rates in the top 10 we have ever seen for a medium frequency trading model.


The final interesting piece of information above is that the model is only exposed to stock market risk 69.8% of the time. So it delivered 1.38 times the return of the “buy & hold” but was exposed to far less stock market risk. If we divide the return of 203% by the 69.8% timing model vesting time we get a Return/Risk ratio of 2.91. This is another way of saying the model performs 2.9 times better than the “buy & hold” when taking into account how much it was exposed to stock market risk. It is a way of equating the “buy & hold” returns of 147.7% which were obtained by 100% risk exposure to the timing model. If the timing model was exposed to the same risk (100% stock market exposure) as the “buy & hold” then theoretically its performance will be 2.91 times the “buy & hold” or 2.91 x 146.7 or 426.9%

Let’s examine some charts displaying the individual trades themselves. The first is the chronological sequence of trades:

The next chart sorts the trades low to high so we can get a feel for the trade distribution. This is a very good way to see visually if the model has a “positive expectancy”. We see there are a lot more green bars than red bars and we see that the red bars are all small – exactly what we like to see in a trading model.

Two Options
There are two options available for trading this model. The first, which is the recommend one discussed above, is to sell 50% of your positions when you see the profit taking signal. This is called “Strategy-B.” The second option is to sell EVERYTHING when you see the profit-taking signal, which is called “Strategy-A“.  This second option actually performs much better on an absolute and risk-adjusted basis than the recommended option, as shown below

The second “aggressive” option is obviously in the market for far less time – 61.6% versus the first options 69.8%. It also returns more – 223% versus 203%. Higher returns for less risk means a much higher Return/Risk ratio namely 3.62 versus 2.91. The gain/loss ratio is also higher, namely 11.64 versus 11.09. Maximum draw-down reduces from 7.1% to 5.6%. It seems a “no brainer” to opt for the 2nd option, but look at the red highlighted area in the above chart where the 2nd model incorrectly assumed the market was weakening and issued a profit-taking signal. If you sold out completely on the profit-taking signal (happy with your 6% gain) you would have prematurely ended your trade – since the markets continued a nice long run (see how the green line was rising when the red line was flat). You “lost out” on a further 6% gains. This is always a risk for the market timer and so we feel the first model that offloads only 50% on signs of trouble will actually be more robust in the long run and cater to a wider variety of market conditions into the future.

End-of-Day System
All actions of the model are based on end-of-day closing prices of the Sp500 Index. You can follow the model using any financial instrument that tracks the SP500, namely SP-500 ETF’s or tracker funds, SP-500 futures contracts etc. If there is a trading action decision,  it is issued using the value of the close of the SP500 meaning you have to be around 20-30minutes before the close of the market to catch the action early. Alternatively you can perform the trading action the first thing the next morning. Placing your actions just before the close of the market gives slightly better performance than executing your trade the following day.

Trading Model HUD and usage instructions
There is a live “Heads up display” (HUD) of the trading model that will give 15-min updates to trailing stops, profit targets and index levels so subscribers can monitor progress of the model and even monitor progress of metrics we use to issue BUY, PROFIT and SELL signals in real-time, allowing you to anticipate the odds of a signal coming in the next few days. You will be able to access the HUD from the MEMBERS section after you have logged in with your credentials. Clicking on “A” reveals an information box giving a short description of the chart you are viewing. The information box can be collapsed/expanded at any time by clicking on the “+”, “-” buttons.


You will be able to browse all 5 years history of the model with grab-handles “B” or by clicking on the pre-configured time spans shown at “H“. The default view is 3 months. Blue shaded areas are periods we are long with Strategy-A and green numbers “G” on the top of the shaded blue areas depict trade performance (gains.) The red dotted line “D” is the trailing stop “STOP/LOSS”. The black line “C” is the equity account of the trading model, which started in March 2009 with $10,000. The black equity curve is set to the right-hand axis on the chart.

When we open a new trade, you will receive an email alert and the chart will show a dark blue vertical line on the close of the day the trade commenced. Once a new trade is opened you will see a dotted orange line “E” called ABORT which is a maximum loss limiter for the trade of around 2.5% If the SP500 index should close below this ABORT line, we abandon the trade immediately. This threshold ensures the strategy sufferers small losses and because the maximum loss is capped we can use it for effective position sizing. Let us illustrate by way of example. Assume your portfolio is $100,000 and we are about to embark on a new long trade. Assume you want to minimise losses of any trade on your account to 2% of your capital (which is the prudent measure recommended for most traders.) This means we want to limit a potential loss to $2,000. We will incur this loss if the SP500 closes below the abort line (after falling 2.5%). We also need to budget for say 0.5% slippage (we are not guaranteed to close at the exact threshold) and maybe another 0.5% for brokerage meaning 2.5% + 0.5% + 0.5% = 3.5% losses plus costs/slippage will invoke the $2,000 loss. This means we can put $2,000/3.5%=$57,142 ungeared exposure into the market with our trade. If you wish to deploy the strategy with gearing (leverage) our tests show gearing above 2 times to be too volatile.

After the trade has run for a while without closing below the ABORT trigger “E” , the red dotted line “D” will eventually overtake the orange dotted line and take over the function of exit trigger for the trade. At this point the orange dotted ABORT trigger will no longer be displayed. It is important to note that regardless if we are following the orange dotted ABORT trigger or the red dotted STOP/LOSS to govern our trade exit, that the exits are END OF DAY decisions, meaning once you get a sense toward the end of a trading day that the market is going to close below any of the exits, you only then invoke your sell command to close the trade. If you set automatic intra-day stop orders based on these thresholds there is high likelihood you will be prematurely taken out of our trade due to intra-day volatility. If you have to use automatic stops, ensure they are well below the indicated triggers on the chart to ensure you do not get taken out of your trade prematurely.

Ratchet Stop : You will see an orange dashed line “F” on the chart when a new trade commences. It is normally quite high up from the entry point of the trade. This is called a RATCHET stop. Its use is very simple – when the SP-500 index closes above this level, then this level becomes the NEW STOP LEVEL, taking over from the red dotted line. We have effectively performed an aggressive “ratcheting up” of our stop from the red dotted line to the much higher dashed orange line. Provided the SP500 index remains with closing values above the RATCHET STOP we remain in the trade. If the SP500 stays above the ratchet stop for long enough, the red trailing stop will eventually catch up with it and resume its function of the exit trigger for the trade. If the SP500 should close below the RATCHET STOP however then we offload 50% of our positions if we are following the recommended Strategy-B or 100% of our positions (go all into cash) if we are following the aggressive Strategy-A.

Example walk-through trade : The below chart refers from a real trade recently concluded:

Point “1” signals the new trade. We then use the thin dotted line above the thick dashed trailing stop as our ABORT threshold. Fortunately the SP500 remains above this threshold all the way to point “2” by which time the dashed trailing stop “STOP/LOSS” has risen above the ABORT stop and takes over the function of governing the trade exit (note how the abandon stop ceases to display after this.) The SP500 remains closing above the red dashed trailing stop until point “3” when we close above the orange RATCHET stop. The dotted orange line is now our new stop level, locking in the bulk of our profit so far (locking in far more than the red dotted line which is much lower down.) The Sp500 index continues to close above the orange dotted line and after 2 weeks, the red dotted trailing stop eventually overtakes the orange line at point “4” resuming its function of exit trigger for this trade. Eventually the SP500 closes below the red dotted trailing stop at point “5” and we finally get out the trade in its entirety, moving 100% into cash. If at any point during the trade between points 3 and 4, the SP500 closed below the orange dotted ratchet stop we would have the option of offloading 50% of positions or 100% of positions depending if we are following Strategy A or B.

Please note that the next trade set-up can only occur once the SP500 has closed below the red dotted trailing stop “STOP/LOSS” At this point, Strategy-B gets out the market and we are ready for another trade. So if you are executing Strategy-A and got out the market but Strategy-B is still in effect, you will have to wait for Strategy-B to complete her trade before we issue another BUY signal. For this reason it is recommended you execute Strategy B (offload 50% when the SP500 closes below the ratchet-stop threshold.)

Hovering your mouse over any area on the chart will bring up a highlight box showing the exact values of all stops and ratchet targets as shown in the example below. All values are expressed in SP500 index terms. The equity value is the amount of money in the bank account of trading Strategy-A at the indicated point.


The service will be offered for free to RecessionALERT subscribers under a multi-month “BETA Program” until further notice, after which it will be available as an additional $249 per year option on top of the existing RecessionALERT subscription. It will not be offered on its own, but as an addition to the RecessionALERT subscription since the model uses the RFE and data from the Weekly Sentiment Report to determine bull/bear market status.

About Dwaine Van Vuuren

Dwaine has a Bachelor of Science honors degree and is a full-time trader and investor. His passion for numbers and keen research & analytic ability has helped grow RecessionALERT into a company used by hundreds of hedge funds, brokerage firms and financial advisers around the world.