Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Stock performance prediction

Hi guys, in few days I am going to present my bachelor thesis, which goal was to put together an investment portfolio for a fictional fund containing various stocks. The thesis has been finished with a specific proposal, but I believe part of the presentation should be dedicated to performance prediction of the fund.

Now, I am still new to quantopian, with only limited capability to create algorithms, and my question is : Is it possible to create a simple, basic prediction for stock performance based on historical data?

Specifics are following:

time frame - 1 year
stock portfolio - RDS.A 30%, XOM 30%, EPD 25%, TTA 10% and CVX 5%
no fees, taxes and other factors are considered in this scenario

Thanks a lot

2 responses

Maybe using Bayesian inference?

When creating a tear sheet from a backtest you can use the option live_start_date = ‘dd-mm-yyyy’ to see the Bayesian cone from your live start date, but only on historical data I believe.

I would also add to Joakim's suggestion and include round_trips in the tearsheet, i.e.:

bt.create_full_tear_sheet(round_trips=True, hide_positions=True, live_start_date = "2014-08-01")  

Round trips gives you various performance stats like win rate, payoff ratio, etc.
Hope this helps and good luck on your thesis presentation.