The SEC recently fined $35M to F-Squared Investments, for misleading investors with its AlphaSector ETF rotation strategy. Apparently, even the professionals make serious back test mistakes – in this case, prompting the SEC to pursue a substantial fine!
The following is one of the statements made by the SEC in its proceedings of December 22, 2014, showing how F-Squared misled investors with its back tests:
“The inaccurate compilation of historical data substantially improved the AlphaSector strategy’s advertised hypothetical and back-tested historical performance. If an investor made a hypothetical investment of $100,000 on April 1, 2001 (assuming a reinvestment of dividends and no further contributions or withdrawals), the investment would have been worth approximately $128,000 on August 24, 2008 if invested in the S&P 500 Index. With accurately timed (but still hypothetical and back-tested) signal implementation, the same investment in F-Squared’s hypothetical ETF sector rotation strategy would have been worth $138,000. However, by implementing the hypothetical and back-tested signals one week early, F-Squared advertised the investment as worth $235,000 – an increase of approximately 350% more than if F-Squared had applied the signals accurately.“
Misleading Investors With Biased Back Tests
The SEC had additional reasons to fine F-Squared other than the above, and I suggest you have a look at the proceedings. Still, it’s important for you, as a prudent investor, to realize how important it is to ensure the back tests you consider for your investment decisions are not biased to show particularly good results. It starts with using accurate data of course, but that’s only the beginning. In F-Squared’s case, they outright cheated by looking ahead in the data by one week when they implemented their trading signal. That is clearly unethical and needless to say, I would suggest you don’t invest your money with these people! 😉
Lack of ethics aside, the main take away for you as an individual investor is to question the validity of back tests. Is there reason to believe there was some form of look-ahead bias? Was there curve overfitting? How certain can you be that past outperformance is likely to continue in the future?
Always probe deeper and aim to understand how the strategy was tested before considering any investment. Doing this extra leg work up front will pay off handsomely later on, even if it only means that you decide not to invest in what may later turn out to be a scam.