I do this list of tips so I review before I do some back-testing analysis.
main questions (from John Bogle)
- What is the rationale or hypothesis?
- What is the empirical evidence?
- What are the implementation results?
Tips to remind
- Do a test of your system with 2008.
- If you look for correlation, you probably will find it, it does not mean that it will work in future.
- Test in not the best conditions and compare with other systems of
- Check volatility, tests on periods with low and high volatility is useful to get conclusions.
- Track the results in a word document with screenshots of the work done, the variables you touched and the specific results.
- Prepare set of 5 to 10 stocks for the back-test, track the results of all these stocks.
- Survivor bias: you are not adding the companies that have not survived.
- Look-ahead Bias: you have all the data in front of your eyes and can make trades according to that data, this provokes you get a false understanding about the behavior of your system.
- In sample bias : you use a sample data that is always the same or similar, so then your system adjust to this sample and lose effective.