All Ords vs S&P500 Market Strength Comparison.
These 2 indices generally perform pretty much in tandem so any exception to this is a matter of some interest.
On 11/2/10 the SP500 turned green. Interestingly, on the very same day, the All Ords bottomed out and began to rise - although it was another 19 days before it too turned green.
Anyone buying back into the Aus market of the basis of the SP500 strength would have timed their re-entry to absolute perfection.
http://i602.photobucket.com/albums/t...llOrdsSP36.gif
A Matter for further Research and Development vs Ratkin's Rancour!
Quote:
Originally Posted by
ratkin
You consider it a good call that people were buying australian stocks when your system was deep in the red?
It was a good call because it worked and because the buy decision was quite logically based on the fact that the MSI system had the US indices clearly green. History tells us that the SP500 and the AllOrds indices generally turn at the same time, so Jess9 was quite reasonably counting on this relationship holding good. Which it did. It was a good call, based on a logical hypothesis. It was not a call that I would have made, but it was a good one nevertheless, providing a nice illustration of yet another way of utilising MSI signals to advantage.
Quote:
Originally Posted by
ratkin
If the market had gone down in the last few days instead of up, then you wouldnt of been saying it was a good call, you would of been pointing out to them where they went wrong
Dead right! I would have been saying that it was a bad call and that the AllOrds was always the best index to apply to the Aus market. I would be saying that this was evidence that the US indices demonstrably did NOT lead the way for the Aus market and that the 2 markets were essentially unrelated.
But that is not what happened is it?
As per usual, the AllOrds was led by the US indices. In this instance, the SP500 "buy" signal was much earlier. This finding identifies grounds for further study. Is it best to act on whichever signal comes first? Would using the SP500 across the board provide superior signals for the Aus market? Could the SP500 perhaps give the better entry signals but the AllOrds give the better exit signals? Is it in fact best, overall, to totally ignore the US indices?
Backtesting will give the answers to all of these questions.
You learn something new every day!
The Australian AllOrds index and the S&P 500 USA index move closely in tandem. I have always maintained that because they are so close, no significant advantage can be gained by applying conventional technical analysis to the SP500 and using those signals in the Australian market. I argued that whatever effect the US market had on the Aus market was already "built in" to the AllOrds index. I found no useable, tradeable difference between the US and Aus indices - signals generally fired together.
Jess9 bought back into the market on an SP500 MSI signal that was not (at that time) showing on the AllOrds MSI plot. The AllOrds signal did not follow until 19 days later. This worked out very well for Jess9, but raises some interesting questions.
Was this simply an aberrant "one-off" fluke that may never be repeated?
Would SP500 MSI signals applied to the AllOrds perform any better than AllOrds based MSI signals?
It can be seen from backtesting 7 years of comparitive data, that SP500 MSI signals precede AllOrds MSI signals surprisingly often (light blue lines). The reverse is quite rare. Since SP500 signals are generally earlier than AllOrds MSI signals, they should therefore be more profitable, on average.
Backtesting shows us that the Market Strength Indicator functions as a type of trend indicator. This means that if ever the market becomes trendless (goes into a trading range) it will generate a string of alternating signals, as marked within the magenta elipses. These will tend to be losing trades, albeit with very small losses incurred. If needbe, this could be addressed by adding a filter such as the ADX or ADXR to warn those that can't see from the chart that the index is crabbing sideways.
BACKTEST RESULTS
Standalone AllOrds MSI signals gave an average annual gain of 13.1%.
Generating both AllOrds and SP500 signals and acting on whichever one fired first :- Average annual gain 18.4%.
Using only SP500 derived signals and applying them to the AllOrds :- Average annual gain 20%.
By way of comparison, Average Buy/Hold annual gain over the same period was 7.8%.
Keep in mind that all of these MSI backtests were performed using arbitrary Buy/Sell default values of >0.4 and <-0.4. Optimisation will find the "best" values to use and will accordingly give even better results.
http://i602.photobucket.com/albums/t...llOrdscfSP.gif
Mk XIII. Market Strength Index now optimised.
The process of optimisation involves backtesting indicators using historical data, with a view to finding the settings that best suit your aims. You can optimise for the greatest capital gain, the fewest number of trades, the best win/loss ratio, the lowest drawdown, the smallest number of losing trades etc etc. The MSI plot itself has no parameters, but we can vary the levels at which the plot is deemed to be "strong" or "weak".
If we optimise for the greatest gains, an entry trigger level of 0.1 is best. This is the "light green" entry signal. If we want a more conservative entry signal, one with fewer losing trades, an entry threshold of 0.5 (dark green) would be better. Of course, overall returns will be lower (see table). The MSI system, backtested over the last 7 years, has given returns approximately double those achieved by "buying and holding".
Similarly, the MSI trigger levels can be separately optimised for the S&P500 Index. Optimised MSI signals generated from the SP500 can then be applied to the AllOrds to see whether, on average, this would provide superior returns. In other words, we can investigate whether the SP500 leads the AllOrds or not. If it does, then this is very useful information. Results as per chart - more than triple the Buy/Hold profits and more than 50% higher profit than the "standalone" optimised MSI AllOrds results.
http://i602.photobucket.com/albums/t...PB/MSIss-1.gif
Table shows compounded gains from an initial investment of $100,000. 7 year period. Brokerage 0.12%/transaction. Figures rounded.
To me, the conclusion that, over the last 7 years, the MSI AllOrds signals have generally followed the SP500 MSI signals is inescapable. Furthermore, this fact can be utilised to give higher profits. The SP500 derived signals are not always better than the direct AllOrds signals, but over time, the difference really adds up. A couple of recent examples are shown here. You can see that the firing order was SP500 then AllOrds light green then AllOrds dark green. The March signals were identical. The trend ribbon along the bottom of the chart gives the SP500 derived MSI status.
http://i602.photobucket.com/albums/t...AllOrds315.gif