Prerequisite: complete the three Quantopian tutorials described in homework 4, and review the following lectures:
Factor Models How fundamental data can be used in factor models.
Factor Analysis with Alphalens The statistics of determining whether a factor is suitable for a long‐short equity algorithm.
You do not have to report anything related to these tutorials lectures, it is only to prepare you for this case study.
Study and run the case study: Traditional Value Factor. https://www.quantopian.com/lectures/case‐study‐traditional‐value‐factor
Steps:
1. Explore alternative fundamental factors, and extend the original program with your new fundamental factors.
2. Evaluate your factors and the original factors of the lecture using Alphalen’s full tear sheet based on return, information analysis, and turnover. Analyze your results (you do not have to explain every graph and every indicator. It is enough if you discuss your results in about half page).
3. Implement a trading algorithm using your selected factors, and backtest with Quantopian.
Build the following table based on the performance tab of Quantopian’s backtest facility:
Using original factors
Using your factors
Total Returns
Specific Returns
Common Returns
Sharpe
Max Drawdown
Volatility
Changes introduced in the simulations
4. Include a graph as the following for the two simulations:
5. Discuss the results of your algorithm when you run it with the original factors and with your selected factors (about 1 page without including graphs or tables). Explain why your results are better, equal or worse than the simulation using the original factors.
6. Conclusions. What did you learn from this exercise? How can you improve your algorithm for a future test? (you do not have to run an extra test, this is only a discussion). (about half page)
