Basically you can use any data to test or study systems.

But you will notice that:

1- The data that you use for developing on FSB , lets say the demo data , when you applied to the actual broker data that you use it will render a different result, that difference could be 100%.

2- The developing of any system then must be based in the data that will be applied.

3.- The shorter the time frame selected the larger the difference in the results form different data sources.

4.- At the beginning I thought like the data is the data , but is not like that . The data that is presented in real time ( or collected from history) will provide for an EA different results . The reason is the brokers accommodate bid and ask in a different manner and that alone will make the mathematical computations to be at times completely different.

5.- Only in rare instances the data form the same period form two different sources will provide the same results with the same EA been applied. ( Some variables such as time of the server and a larger time frame will bring the gap of the results closer.

6.- The recommendation is always to use the same data in the EA from the source that is intended to be use. I have run  hundreds of test that confirm the rule , so it is clear 100% the why it better this way.

7.- Follow the threads in the Forum under the search : Data or best data source , and you will find more references to the topic. I have just giving you a summary or it .