Great question. This is why it helps to formulate an idea, design the experiment, and then test rather than going right into experimenting. There are seemingly infinite ways to put a model into action when accounting for risk appetite and trading strategy. I am by no means an expert on this as I am still experimenting myself.

For example, in a long-only strategy, you could stack different models forecasting 10 strongly correlated stocks to define a trading strategy. To keep this example simple, the model would run in the aftermarket hours and take execute market buy orders at the market open and sell 5 minutes before the day’s close. I would set a threshold to determine the minimum % change for a trigger-- let’s use .5% for example. If tomorrow’s price forecast is 0.6% greater than today’s, then the model would spit out a 1 to proceed with purchasing that stock. Now repeat this process for the other 9 correlated stocks. If your entire model's output sum (max being 10 and min being 0) comes out to be 7 then it may trigger your broker to take action.

This is where your position sizing would kick in. You may want to only have a 1% total portfolio risk for the day. So if you invest 10% of your portfolio into each one of those 10 correlated stocks (100% allocation daily) you would have a stop loss trigger to liquidate a stock's position when that return hits -1%. If positions do not hit the stop loss then they will be liquidated at the end of the day.

This is the level of specificity needed before going build any model. You have to fully map out any idea to clearly understand how you are going to test it. Once you have designed your model you need to extensively backtest it while keeping the strategy's integrity (e.g. AMZN as one of your stock 10 stock positions and start the backtest data before AMZN is even listed on the exchange).

I hope this answered your question! If not let me know and I will try to elaborate.

**Disclaimer: In case it wasn’t clear to anyone reading this, do not take the above strategy as a real strategy is purely an example I came up with on the top of my head.

I am a data junkie working to kick my addiction to MS Excel with Python.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store