For example,
I trade XLK (The US Technology Select Sector SPDR® Fund).
The current constituents: https://www.sectorspdr.com/sectorspdr/sector/xlk/holdings
However, its historical constituents are different in the past.
how can I re-simulate XLK?
If yes, how?
Hi,
Please read the instructions of the competition
You can only pick stocks which are part of the S&P 500 at the day of selection. Be aware that some stocks can either drop out from the index or enter the index, depending on decisions from the selection committee. Stocks can also drop out from the index if they are de-listed from the exchange, which typically happens following a merger or acquisition event. When a stock drops out of the index and if it’s part of your portfolio, the position has to be closed.
Example:
Netflix can be part of your portfolio starting only from 17 Dec 2010.
Alcoa Inc cannot be part of your portfolio after the 31 Oct 2016.
Therefore you cannot access any ETF for this competition. The goal of this competition is to select 50 stocks among the investment universe and outperform the equally weighted S&P 500.
Hi,
The gist of the competition is to come up with your own logic of how to combine stocks so that your strategy outperforms the S&P 500 index. A systematic portfolio is a portfolio that rebalances its components based on rules (you need to define them) so that it adapts to market conditions in the future (not only in the past - that’s called over-fitting, and that has no value when it comes to investing in the future). In a nutshell, you have to create your own portfolio not replicate an existing portfolio.
In order to test your portfolio, we will compute its returns and risk profile over undisclosed period(s) to check whether it does well on unseen data or if you have just fitted the data you had at your disposal (2006-12-31 to 2017-12-31).
By trying to replicate xlk, you introduce a high bias to your portfolio, excluding all sectors but one. If an event like the burst of the dot-com bubble in 2001 happens again, your portfolio will suffer massive draw downs (NASDAQ fell 78% back then). Usually, practitioners favour a diversification approach to portfolios rather than putting all eggs into one basket.
That being said, you can obviously use sectorial information to try to switch to riskier sectors (high beta) when the market is bull and to more defensive sectors (low beta) when the market becomes more risky. The function getTickersSP500 has an industry code column. Read about GICS sectors to understand how the code works and how to extract valuable information from them.
Hope this helps
Thanks Herve and Phdamay.
I know I can only pick stocks which are part of the S&P 500 at the day of selection. But does it mean that I can backtest some of the S&P 500 historical constituents which went bankruptcy before to avoid surviorship bias?
Such as WorldCom, Lehmen Brothers, Eron.
Looking forward to your reply
Have a look in the getTickersSP500() function that you can find on Qlib that Alphien team has prepared for you. This functions gives you all the list of tickers for a particular period. This functions gives you specifically the date it was included / excluded from the S&P 500 (obviously they are excluded on the date they go bankrupt). Delisted stocks are available to avoid survivorship bias that is indeed very important in the competition ranking.
Generally speaking we suggest exploring all functions made available to you, that’s the best way to start. Thank you for the feedback.
Hope that helps, happy coding. Lionel.