First look at Step Two: single-market pattern recognition, multi-market money management require a separate optimization effort.

User Rating: / 0
PoorBest 
Written by Forex Automaton   
Thursday, 13 August 2009 15:53

This is the first report on the multi-market trading system performance in the "isolated" mode. The "isolation" means that each market is being predicted independently of the rest, with no attempt to discover and learn inter-market patterns such as those we observed and reported. The money-management component of the algorithm is aware of the concurrent trading ideas form the different markets and weighs them against one another, selecting the most promising. Initially seen as a "boring" step (so-called Step Two) towards a smarter algorithm that does take full advantage of those patterns, the isolated analysis also provides a benchmark against which the value of the inter-market analysis will be measured. One might expect that for the same values of the parameters the profitability of the Step Two algorithm would be comparable to that of Step One, while the volatility of the performance (investment risk) would be reduced thanks to portfolio diversification.

This preliminary study reveals that, all the algo parameters being the same, the competition between trading ideas boosts the profitability considerably, but also increases the risk. The fact of the "interaction" between trade ideas questions the original assumption that the best money-management parameters for the isolated markets are the best in the multi-market scenario, as already noted in the preliminary Step Three report.

The summary report on the single-market trading system optimization, posted on July 21, 2009, ended with the words:

"For the immediate future, the plan is to switch to the optimization of the multi-market version of the system. This will proceed in two distinct steps, Step 2 and Step 3 respectively (this one being Step 1).

In Step 2, the money management component of the algorithm will become aware of the concurrent trade ideas in the different markets and will learn to treat them properly. The markets themselves at this point will be forecast independently, with the "consensus" parameter settings just obtained. The thing to check is the reduced "risk" or fluctuation in returns. It is expected to be reduced since the concurrent trades are not 100% correlated.

In Step 3, the forecasting algorithm itself will begin to handle the different markets jointly, effectively learning to utilize cross-market patterns. This is a step into uncharted territory -- not because nobody has done it, but because those who have done it are understandably reluctant to share the charts. The result of the Step 2 alone may already constitute a useful product."

The basic framework remains the same: a run of the program included simulations of trading histories of 24 independent "virtual traders" (forex robots), each of them being an incarnation of the same algorithm, differing by the setting of the adjustable knobs. These 24 are the winners selected on the basis of simulated performance in the Step One simulations. As in all other published studies so far, the trading is performed on one-decision-a-day time scale, with 1:100 leverage, risking no more than 10% of the trading capital at any point in time. This report uses the EUR/USD, USD/JPY, GBP/USD, USD/CAD, USD/CHF and AUD/USD day scale data covering the time interval from August 20, 2002 to March 23, 2009, with the actual trading starting in March 2006 (when the initial "training" of the system was completed). The trading system control parameters remain as previously defined.

Table 1. Comparing trading system performance in Step One and Step Two running modes. The sets of parameters are the same in both cases, and the same as selected to be the best in the summary report for Step One. Fred: 71, 74, 77, 80; stop-loss: 1.13; enter-the-trade threshold: 0.010, 0.011; exit-the-trade threshold parameter: 0.009, 0.011, 0.013. Thus, there are 4 × 1 × 2 × 3 = 24 traders. With 6 markets independently traded in Step One that results in 144 independent trading histories. In Step Two, there are only as many trading histories as there are traders, 24.

Mode of analysis Step One Step Two
Number of independent trading histories 144 24
Annualized return mean, 1=100% 0.43 0.81
Annualized return RMS 0.22 0.37

Table 1 compares trading system performance results obtained for a the same values of the algo parameters in two modes of operation: Step One and Step Two, both defined above. In Table 1, annualized return mean and RMS refer to the group mean and RMS of the group of traders, selected by cutting on the adjustable parameters that control the trading system.

The conclusion is that the parameter optimization problem needs to be re-addressed in Step Two mode. The limited parameter window shown in Table 1, however good it is for Step One, may very well be not the optimum for Step Two. Step Three will be different in even more ways and there is little hope of understanding that before Step One and Step Two, and the differences between them are understood.

Bookmark with:

Deli.cio.us    Digg    reddit    Facebook    StumbleUpon    Newsvine
Last Updated ( Monday, 04 January 2010 12:30 )