I can only speak for myself but I have come to realize that I am the weakest part in the trading chain when it comes to speed and consistency but in another area I am stronger than computers: developing ideas, making assumptions that break the mold. I might be able to see what might work, but in the long run I have more chances if I let the computer calculate through decades in seconds.
So be honest, finding a trading system first and then automizing it is not how it works. You should pick some trending and reversal strategies you want to combine(there are some articles about it), set up a few EAs (automate) and then test the EAs in optimization mode for robust and reliable data. Can you imagine if you have let's say seven parameters which can have thousands of values? Even if you go for ten values for every parameter you still have 10million passes to go to find the best ones. Chances are that you see some of the average ones right when you take a sharp look at the curves though, but nonetheless you will waste a lot of time and effort which the computer can do for you.
In the beginning you have to do it manually a few times to pick up the experience.
Then test the best results even further. Look at the capital diagram. How do they behave? Do they rise consistently or do they fluctuate a lot? Do they work better in times of trend or sideways motion? How can you create a reliable equilibrium between EAs of different types and calibers?
I highly advise you to get aquainted with strategy tester. Might have to let your computer run for a few hours or even a day until you have robust results though.
Then you will have to think about 3D plotting to extract correlation data more easily, it can be done in excel but it is tedious.
Please enable the necessary setting in your browser, otherwise you will not be able to log in.