Hello there:
I have taken the course "Trading with Machine learning: Regression"" …And given the information I learned there… I came out with an Idea…:
In this course… multivariate linear regression is used to predict the prices… and the price of gold is used as an example… using many features an OHLCV data of course…
I was thinking about using the exact method for predicting the low price of a stock… for the next month … and buy a Call option at that stock…with a strike price at the same price predicted by the machine learning lmultivariate linear regression… the expiration date of this call option should be in a month… and also…I don't plan to sell it… I plan to exercise this option…a few days before it expires…
Since it is a call option… once I excercise this… I am going to be long on a stock… and I plan to sell it at a higher price… on the market…
Since I am still learning options… I was wondering if this could be a good strategy ???
What do you thnk about this ?? Can that be good… ? or maybe I am wrong…
Hey Ghery,
This sounds interesting, but the success of your strategy heavily relies on the accuracy of your regression model's predictions.
To understand whether a strategy performs well you must backtest it. On the basis of the results you can decide whether you wanna take the strategy live. Even after that you should definitely practice risk management measures because past performance does not guarantee future returns.
Hope this helps!
Thanks,
Rushda
Hello Rushda:
Thanks a lot or your answer I really appreciate that… You said that I shoudl practice risk management measures… but … The worst thing that can happen when going long an option is that the money used to buy the option be lost… right??
Yes, I have also thought about putting an stop loss below a certain level… (when going long a Call option it can also be sold right?? )
There is also an important aspect to be mentioned here and that is…How many features should be used ?? I am getting very confident with moving averages… since I have seen that they correlate very well with the target variable I want to use (that is the difference between the Open and the Low…)
besides that… what other measures for risk management would you recommend me??
Hey
Yes, you're right options can expire worthless and in such cases, the entire premium paid for the option would be lost.
Setting a stop-loss is definitely a great idea! In addition to this, there are a handful of risk management techniques that you can adopt. You can read more about it in this introductory blog. Remember, risk management is a critical aspect of trading. It's essential to tailor your risk management measures to your specific trading style, risk tolerance, and investment goals.
Regarding the number of features to use in your regression model, it's a good practice to consider a range of relevant features to enhance the predictive power of your model. While moving averages can be informative, you may also explore other features such as volume indicators, technical indicators, or fundamental data specific to the stock or sector you're trading. Experimenting with different combinations of features and conducting thorough backtesting can help you identify the most effective predictors.
Hope this answers all your questions 
Instead of BUY CALL you could SELL a PUT BELOW the strike what your model has predicted.
However as Rushda Ansari summed it up. it comes down to your model's predection accuracy and portfolio risk management.