1 - Optimization is not "forever" and re-optimization is necessary from time to time.
2 - There is a right and wrong way to optimize. Simply optimizing can lead to great looking results that then fail in practice. To overcome this the optimization process should include optimization on "in sample" data and then testing with "out of sample" data.
3 - Software for proper optimization need not be very expensive and AmiBroker is a good example.
Thx for the amibroker mention. Can you elaborate on the out of sample data?