Back to the OP plus one for python. My last real job was at a multibillion dollar systematic HF that had previously run every combination of language setups under the sun. Then we finally took the decision to migrate everything to Python. The result was a huge improvement. In terms of large scale trading application, it doesn't get much larger than this. In the last couple of years I've been trading my own money, and I built my own fully automated python platform. With a little bit of embedded C and a sprinkling of cython you can deal with most performance issues to the point when it will be the latency of your communication with the broker and exchange rather than your technology stack that is slowing you down. By the way yes the IB API system is hard to deal with, but for those that haven't seen it I put a series of posts on my blog to explain how to set things up click here. GAT
agreed python does make things easier if you dont need the speed.Also, its nice to see that your posts have a good signal/noise and you walk the walk by showing code. Here's some links that may help others. http://j.mp/audgists http://bit.ly/17h05Rm
I like the second link, using python notebook. Python notebook is yet another very cool thing about python. I should also mention the IBpy library, for which quantstart is a good resource (https://www.quantstart.com/articles/Event-Driven-Backtesting-with-Python-Part-i). GAT
For a full scale backend trading framework could you explain where latency and speed is not relevant? If speed in the slightest matters then basically this would negate any Python implementation. And maybe that is why not one single hedge fund nor sell side firm would ever consider running it's trading platform off Python.
right tool for the job. what about products that are not latency or speed sensitive and barely trade? The OTC space dwarfs exchange traded products. I get RFQs all day and use python to store them and respond to them. I agree I wouldn't use python to trade exchange products but its used on research platforms at a number of hedge funds and in risk/pricing at banks (athena ,quartz)
Why did you decide to use Python. Is this the strength of Python libraries and the community of researchers and programmers that supports it or did you create your own algorithms? On a side how are you dealing with the issue (it is an issue for me) of significant white spaces?
On the corporate side libraries and availability of good programmers were definitely important. Python is a 'nice' language in the sense that non programmers from a technical background like the syntax and programmers like it also. "significant white space" can you elaborate? Are you referring to the compulsory indentation? To be honest after a week or so you no longer notice you are doing it. And it means code is always readable. GAT
Incidentally this is the hedge fund I worked for until a couple of years ago. Currently at $18bn and still using python https://www.ahl.com/coderprize https://github.com/manahl https://twitter.com/ManAHLTech GAT