Let's say I have an investment universe consisting of equities in SP500 at time t, and that the universe changes at each t+1. I.e. at time t you can only invest in an equity that is in the SP500 at that time t.

Now let's say I have a strategy that I backtest using different parameters for the holding period / rebalancing period over the entirety of my data.

Wouldn't the resulting "optimal parameters" be very susceptible to Look-Ahead Bias, since they're defined over and tested over the entire period?

  • $\begingroup$ I don't know if it can be considered a 'look-ahead bias', but there is an issue here because you are using the same data over and over again to test Multiple Hypotheses. You may want to break your data into 2 periods, try the different rebalancings on the first period, select the best and test it on the second period. The performance in the second or OOS period can be considered unbiased. $\endgroup$ – noob2 Jan 29 at 10:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.