a backtest is an incredibly useful research tool…. if it is done right. If it is done wrong, as most are, it gives a false sense of confidence or worse, is used irreverently.
When done right a Backtest may well be the best tool an investor has.
WFOOSMPBT is the gold standard of backtests. Just like GIPS (Global Investment Performance Standards) is the gold standard for performance reporting, there is a right way to create a backtest and, silly acronym or not, the industry could use some conformity in implementing a set of standards.
This standard can be applied to security selection, trade timing or portfolio construction, any or all the above.
So, what is WFOOSMPBT?
WF…. Walk Forward: this means that the backtest is based on creating a condition that creates the investment action at some point in the past and then applies it forward from then, but of Couse still in the past.
OOS: Out of Sample. Sample data is the data used to create the decision. That decision must then be tested in a separate environment outside of the time period used to create the decision so to not contaminate the test with bias.
MP: Multi-Period. The backtest needs to show over an extended period (I like two full market cycles) that it works. Imagine a backtest that just recommends a buy at the beginning of a bull market. Big deal. One rule works in one market condition. But does it work in multiple market conditions? Does it work across varying market conditions? Enforcing a Multi-Period backtest protects the investor from these traps.
BT: Backtest. Any investor that does not backtest quantitative strategies is just a gunslinger with a calculator.