Start The System On A Chart

 
Run the system on the historic graph. Strategy tests will begin as soon as you apply the system to the graph. As a general rule, you should try out your strategy on as much data as possible. For example, if you want to make transactions based on real time data or delayed data, you should test your system on data from different periods, e.g. different series of contracts. Or, if the system is based on daytrading data, you should test it on data from different years. As a general rule to verify whether the system generates proper signals you should generate at last 30 transactions. Only then will your sample be statistically reliable.

Another system performance test recommended by traders is trying your system out on other instruments in the same sector (probably presenting the same market type) and then on different data periods. The system should yield decent result on other, similar instruments, although this is not a rule. It should also do well in different periods, because after all it will be dealing with same instrument all the time. However, when you test a different period for the same instrument, you should focus more on the recent data. The key thing is for you to be comfortable with the system's results and behavior.


Back