Understanding strategy workflow

Before we start discussing the nuts and bolts of building strategies and sharpely, we will look at a typical “quant” strategy workflow or pipeline (very broadly) – steps involved in building a successful strategy.

We will also see how sharpely has got you covered through all the steps.

Step 1: Research and idea generation

The typical strategy creation process starts with research and idea generation. This process might involve fundamental research or iterative stock screening or metric (factor/feature) evaluation/engineering.

You must have the ability to test your hypothesis, keep the ones that look promising and discard the ones that don’t.

On sharpely, you can back test any of your screens, factor models and even individual factors.

Step 2: Build and backtest

Once you have an idea, you need to “codify” that idea into an executable strategy by setting up your model and parameters.

And then you back-test to see how your strategy would have done historically. On sharpely, we have a sophisticated complex event processing backtesting engine that throws up the accurate result as close to reality as possible.

However, always take your back-tested result with a pinch of salt and do not solely rely on back-tested data.

There are many pitfalls of blindly using back-tested results in live trading. We have discussed it here.

Step 3: Paper trade your strategy

The next step is to start paper trading your strategy to see how it performs in real life. While this is still not live trading, the results of paper trading will give you a much better idea of the effectiveness of your strategy as compared to back-tested results. And you don’t have to deploy real money.

Step 4: Take your strategy live

Finally, once you are convinced and confident, you can take your strategy live.

All these 4 steps are highly iterative in nature – for example, you have a good backtest (step 2) but the performance is not so good in paper trade (step 3). You will need to go back to the drawing board (step 1).

Be very careful while iterating as this could lead to what is commonly called “over fitting”. So, you can have a fabulous back test by carrying out multiple tweaks to your model and parameters only to see your strategy failing miserably in real life.

Therefore, paper trading is a good “fail first” mechanism. Ideally, you should paper trade a strategy for some duration before taking it live. This is the reason that on sharpely, we don’t allow users to directly live trade their strategies. It is always paper traded first.

Below is a visual representation of the iterative process described above.