23 January 2019
- Using one of the data sources outlined in the most recent GitHub post, please choose any stock(s) or ETFs you'd like to analyse which have historical data for at least 2 years.
- Leave 6 months of data (Oct17 - Mar18) for testing your strategy, i.e. do not use data across this horizon in your backtest.
- Feel free to use any technique to analyse the data: ML, Technical analysis, correlation, momentum etc... but you must not use any obvious hindsight bias (e.g. "only short stocks in 2008")
- Your strategy can be long-short or long-only but must have some turnover of positions (i.e. not just buy and hold for the entire period)
Discussion of different strategic approaches and common approaches/pitfalls to back testing will be encouraged during the session, but the main aim is to get everyone working hands-on with some data and creating a strategy from start to finish.