In trading systems, once you have the defined strategy and created the trading system, follow continually evaluating their performance and behavior in the market, to find the combination of parameters that best results shed in a given period weather. It is not the same as having a system in “demo” than real account, despite having applied the penalties for slippages and broker commissions.
The historical serves as a reference, but it is not an unchanging pattern, because the markets are changing and dynamic, spewing daily data to consider. The optimization has to be made from time to time, looking for the most favorable movements of the system with the new data.
The developer or trader has to study possible changes accurately and concisely, not merely seeking higher profits, but stability and consistency of the strategy. Otherwise, you can produce a sobreoptimización, which eventually will truncate the expectation of success.
All result of optimization has to be considered a new system, which starts counting from zero, however small it may seem the change made, otherwise the results would be entirely hypothetical.
Highlighting these qualities, it is concluded that greater historical and actual data, the strategy will be better because more stable behavior is achieved.
Systems based on the DAX index; If creation had data from 2000 to 2007, more than their patient information, it would give the result of a bullish trend system completely with gentle movements, from 2000 to 2003 bearish trend, and from 2003 to 2007, trend.
The benefits of any trend system in this period would be amazing and would work perfectly with entries long and short, but what would happen from 2007 ?, in this case, these movements are not as clear, and therefore without suitable filter, begin to produce significant declines in yields.
Since 2007, the market returns to be trend but more sudden movements, the system should be optimized to take advantage of this change in volatility.
Two trend systems can be very different from each other by the intensity or strength of the movements.
Optimizations must perform the more recent past. It is no use that the greatest benefits from the system has achieved in the distant past and now are media or even low results.
The data is the closer to the present, will be more stable to replicate the behavior of markets, trends, movements and volumes.
When choosing a system must be considered that evolution is always better a system with guidelines increasing returns, even if negative interim periods.
Optimization keep the updated system, but a system has its own guidelines, trend, antitendenciales, the underlying, etc.
For many filters to be incorporated, there is no system that always wins in all markets or movements. The solution is a combination of systems that correct these deficiencies and can obtain stable profits to be uncorrelated with each other, curvy positive and lasting benefits over time.
Automated trading systems allow us to very easily update the database so manually would be impossible.