I've only tried one and that is ProRealTime through IG (but can be IB or Saxo I think). Not sure how pro it is because it's in house code which is a version of Basic, so make what you will of that but it does what it says on the box. I'll be going down the Python road but I needed something that was fast to experiment with because I wanted feedback now not a multi year experiment. Which it is anyway but I learn fastest through direct errors unfortunately.
Funny, I initially asked my step father to work with me to build an RMT (risk management tool like in Prop firms) using the IG API. Once I set-up the software I could see it had most of this built in so I could just get into my ideas.
As for back testing data, for me to test a 1min TF I get 5 months. That's dangerville on my TF so I had to come up with a way to track performance which is just watching win rate and consecutive losses relative to the starting period. I save the 5 months charts and match it up to current period to get a larger sample. I do have 5min/1hr/4hr etc.. Think the higher TF's go back 16yrs and help produce more robust systems.
Even if I don't stay profitable on my first go at this, the knowledge obtained and programmes I've created are wonderful tools to compliment manual trading and keep you away from less probable outcomes. But that said there are so many ways to use algos and get the benefits of the execution being done for you that I see plenty constantly optimize to keep their trades valid. I chose a raw approach of leaving it untouched except (4/32 that had to go immediately). I am just using it to work out what works/doesn't and what direction to head but it ran up 170% first 3 months put in some 25% profit DD's and now sitting around 80% as I sized up a bit quickly hehe. So that's not a great outcome for future performance unless the trades with size bounce again as I size up during statistical draw down. But now I have a valid starting point.
Limitless really. But you're creative, knowledgeable in price action and productive....and that's what's needed.
Oh and I have no doubt Calvo pushed this exploration to the limit just look what he got out of it and can do! For me, what I've seen more than shows me this can be done on a retail level but it's not set and forget, you just get better hours and less stress. It's still very much trading.
I might as well put this in because we are talking about it on an old thread which is appropriate and it's about the approach.
My approach was to create both Long and Short MR and breakouts across 4 indices in their opening drives at different times of day to use the same pool of cash and pay no overnight fees. I have some of the same problems as anyone but with this method I can turn off and pull out pieces without upsetting the whole meta system.
(20min delay)