Thanks for any advice in advance.
I did a degree in math (bachelor's) with the intention of learning anything and everything that could be useful for trading. Managed to graduate magna since I read math at a high level. Personally, I like futures more than options.
Basically, error functions are good. A lot of algos are using simple error function stuff. You
could call that machine learning, I guess. I'm pretty sure this relates to GARCH or EGARCH and all that. But, I would also say that knowing market making, financial modeling, and market structure/microstructure are just as important as knowing that stuff.
For example, in something like an index future, there is the basis trade, the index differential market, and the cross/spread against bonds going at all times on top of the vol trade.
Knowing where to aim the machine learning stuff and why is just as important as knowing the techniques.
Signal processing, error/residual analysis, and probability theory are very good. If you combine that with real quant finance knowledge, you will have the insight to write some nice algorithms.
If you use IBKR you can code excel macros using VBA and the free API. With TD/Schwab, you can use thinkScript which is amazing for algorithms and research -- still, I'm envious of the guys who have everything in their own custom environment. Data feeds and software licenses can get expensive.
A lot of traders think you have to use X_Trader or CQG and the like. Not sure about that.