Sub: Algo App - any recommendations? Looking for Flexible strategy build options
please suggest a best Algo app - with flexible strategy build options.
There should be a provision to mention, what to do, when an event occurs - when Target reached/ SL hit - there should be a provision to perform 2 tasks at least.
You should code the Algos yourself using a programming language you are comfortable with. If you want best speed then Java, dot net and C++ are superior. Python is simple to code, but speed is worst. If your algo setup is around creating straddles then using Python shouldn’t be a problem. C++ has best speed, but is toughest to code and so there is a higher chance to write poor quality and defective code.
Java gives a good balance between simplicity and speed. I build all my Algos myself using Java and never resort to any external Algo trading apps.
Yes, Kite has Rest API with which you can build your own Algos using your preferred programming lanuage. Zerodha team have even built SDKs for each programming language like C++, Java, C#, Python.
If you know coding, then starting to write Algo code should be simpler. Else, you need to learn coding first. Good Algo code should have very good exception handling mechanisms so that you don’t get unexpected results.
I know coding in Python, so will it be sufficient to read this documentation for real life trading, or do I need to do some course like Quantinsti does for algo trading
It all depends on what kind of strategies you want to develop and your coding skills. You can start building simple strategies first and eventually build up the complexity as you get more familiar with Kite Connect API. Don’t think any course is really needed.
@rahulsony77 Adding to what @Sujit_C said, Python is more than sufficient for real trading and backtesting, most of the number crunching and machine learning libraries in Python are written in C or Cython and Python is just a pretty interface to those low level stuff. There is no practical difference in speed viz a viz any other language, especially considering the most amount of time would be spent on I/O tasks such as interfacing with a database or making a network request.
In trading, risk management is everything so writing, testing, and deploying reliable code that you’re 100% certain does exactly what you want it do, even in unexpected conditions is quite a tedious task. Probably why even with programming experience @Ranganath.Bangalore created this thread.