Trade Nivesh | Created your algo? Here are three ways to start trading in a live environment
Similar to how an airline pilot will go through hours of simulated training, our algo needs to go through what it will be like in the real world.
Aditya Verma
@tradenivesh.com
In the last installment of this 4-part series, we saw aspects which will show the likelihood of algorithms (algos) working in the future, the same way they have worked in the past.
We will also see another few simple but essential steps to complete the journey. Do we start trading the system now with capital? Or do we go in a more measured way?
Feed handler and order routing:
Now, the algo has to be connected to live market data and orders need to be generated from it. If we are going for manual execution, we may not need to build the order router.
But the feed handler is essential, which will read live market data and generate the signals. We can say now, here is where we separate the signal from the noise.
The algo program has to be tweaked somewhat where it can now read data from the live market instead of the historical data which it was designed for.
This stage requires one to use application program interface (APIs) provided by vendors. We need to start and stop the algo multiple times and run it for a few days to make sure that the algo is fine-tuned and behaves well.
The important parts to check are: a) Not generating incorrect signals – like a sell order coming instead of buy b) No infinite loop – where orders keep getting pumped one after the other.
Paper trading:
This stage is one of the most important when it comes to making sure that the plane you are going to fly will fly well. Paper trading means trading an account with virtual money and real market data in a simulated environment which will give you “as close as possible” to live market experience.
Similar to how an airline pilot will go through hours of simulated training, our algo needs to go through what it will be like in the real world. Interactive Brokers India has an excellent paper trading account facility.
After generating a few trades, we can see whether the paper trading statistics are similar to backtest or not, or do we get any alarming signs.
Risk management and go live: Now we are ready to trade the strategy
We need to configure which instruments we are going to trade and in what quantities. We can have additional checks around quantity not exceeding a certain number or daily loss not exceeding a certain value (especially in case of intraday systems).
Then we start the algo with a smaller capital than what we can trade it on. After running it on some minimal capital with a few stocks, we can then monitor the trades and if everything is going according to historical lines, we know that the algo is not misbehaving.
Then, as we get used to the algo and vice versa, we start scaling up! We are now, slowly and steadily, moving from the base camp towards our destination – Mount Everest!!
As a trekker, one knows that the journey will be full of ups and downs – with snow storms (drawdowns) and avalanches (maximum drawdowns).
Same goes with a system trader. At the end of the day, it is the journey (trek) which can be as enjoyable as reaching the peak (returns).
The author is the Co-Founder of Dravyaniti Consulting.
Similar to how an airline pilot will go through hours of simulated training, our algo needs to go through what it will be like in the real world.
Aditya Verma
@tradenivesh.com
In the last installment of this 4-part series, we saw aspects which will show the likelihood of algorithms (algos) working in the future, the same way they have worked in the past.
We will also see another few simple but essential steps to complete the journey. Do we start trading the system now with capital? Or do we go in a more measured way?
Feed handler and order routing:
Now, the algo has to be connected to live market data and orders need to be generated from it. If we are going for manual execution, we may not need to build the order router.
But the feed handler is essential, which will read live market data and generate the signals. We can say now, here is where we separate the signal from the noise.
The algo program has to be tweaked somewhat where it can now read data from the live market instead of the historical data which it was designed for.
This stage requires one to use application program interface (APIs) provided by vendors. We need to start and stop the algo multiple times and run it for a few days to make sure that the algo is fine-tuned and behaves well.
The important parts to check are: a) Not generating incorrect signals – like a sell order coming instead of buy b) No infinite loop – where orders keep getting pumped one after the other.
Paper trading:
This stage is one of the most important when it comes to making sure that the plane you are going to fly will fly well. Paper trading means trading an account with virtual money and real market data in a simulated environment which will give you “as close as possible” to live market experience.
Similar to how an airline pilot will go through hours of simulated training, our algo needs to go through what it will be like in the real world. Interactive Brokers India has an excellent paper trading account facility.
After generating a few trades, we can see whether the paper trading statistics are similar to backtest or not, or do we get any alarming signs.
Risk management and go live: Now we are ready to trade the strategy
We need to configure which instruments we are going to trade and in what quantities. We can have additional checks around quantity not exceeding a certain number or daily loss not exceeding a certain value (especially in case of intraday systems).
Then we start the algo with a smaller capital than what we can trade it on. After running it on some minimal capital with a few stocks, we can then monitor the trades and if everything is going according to historical lines, we know that the algo is not misbehaving.
Then, as we get used to the algo and vice versa, we start scaling up! We are now, slowly and steadily, moving from the base camp towards our destination – Mount Everest!!
As a trekker, one knows that the journey will be full of ups and downs – with snow storms (drawdowns) and avalanches (maximum drawdowns).
Same goes with a system trader. At the end of the day, it is the journey (trek) which can be as enjoyable as reaching the peak (returns).
The author is the Co-Founder of Dravyaniti Consulting.
Comments
Post a Comment