Matlab vs Python for real-time (market making) strategies

[isotope1]

I've built an elaborate trading system in Python that trades once per day, and that works well for me.

I'd like to start developing market-making and arbitrage strategies that are much quicker. When I think about what is involved in building such a thing in Python, it's a lot more work than a once-a-day trading system.

I want to know, what are the advantages of MATLAB and all the relevant toolboxes for this kind of work? Has anyone tried using Simulink and treated it as an optimal control problem?

Thanks!

 

[Robert Morse]

I'd like to start developing market-making and arbitrage strategies

May I ask what assets classes?

 

[isotope1]

Cryptocurrencies & their derivatives.

 

[Robert Morse]

Cryptocurrencies & their derivatives.

thx

 

[2rosy]

I've built an elaborate trading system in Python that trades once per day, and that works well for me.

I'd like to start developing market-making and arbitrage strategies that are much quicker. When I think about what is involved in building such a thing in Python, it's a lot more work than a once-a-day trading system.

I want to know, what are the advantages of MATLAB and all the relevant toolboxes for this kind of work? Has anyone tried using Simulink and treated it as an optimal control problem?

Thanks!

right tool for the job. it's easier to use java,c#,c++ or similar for something like that. both python and matlab don't really handle threading well

 

[Equalizer]

Many places use Matlab/Python/R/some custom DSL/combination of these for research/prototyping.

Now let me put it this way, there isn't a single execution engine in any serious shop (read: makes money) that I know of, written in any of the above.

 

[isotope1]

It's more in the strategy design and prototyping. For execution we can build something separate.