Option FanaticOptions, stock, futures, and system trading, backtesting, money management, and much more!

When Performance is Irrelevant (Part 5)

This blog mini-series started out with some mutual fund prospectus examples (here, here, and here) to understand and critique how investment performance is reported. Today I continue with more excerpts from Jaclyn McClellan’s article discussing mechanics of performance comparison.

Aside from using rolling periods, another way to generate a larger sample of [potential] performance records is through Monte Carlo simulation. This has been mentioned multiple times in this blog (e.g. here, here, and here).

     > Also, many of the robo services haven’t been
     > around long enough to present meaningful longer-
     > term return figures, hence the reporting of
     > backtested results.

I am okay with using backtested results as long as the backtesting is done in a realistic manner. In order to know this, the methodology must be explicitly presented. While this is routinely done in peer-reviewed scientific journals, I have rarely seen such detailed description in financial materials like a prospectus or seminar.

     > Each company stated that they have historical
     > performance figures available; however, there
     > were caveats. In order to see historical
     > performance for Alpha Architect, MarketRiders,
     > and Rebalance IRA, you must have an account.

I don’t need to be a member of Costco to walk through the store and see what kind of merchandise they offer. Why should I need to open an account simply to sample past performance?

     > Alpha Architect will send historical figures
     > to prospective clients on a case-by-case basis,
     > or refer clients to their book, “DIY Financial
     > Advisor: A Simple Solution to Build and Protect
     > Your Wealth” (Wiley Finance, 2015) where the
     > strategies are defined, explained and analyzed.

This sounds like a less-invasive marketing tactic than requiring me to open an account.

In summary, perhaps nothing in finance is respected more than the almighty performance record. Aside from the proper legal documents, to open a hedge fund I need little more than an audited performance record. I don’t need a degree. I don’t need a license. I don’t need a resume.

Unfortunately though, reasonable doubt challenges the validity of most traditional performance records whether or not they are officially audited. Many people think the key is what performed best in the past. I disagree. The key is what performed well in the past and/or has the best chance of performing well into the future. To determine this we need to assess for fluke occurrence. Large sample sizes, Monte Carlo simulation, and walk-forward analysis are means to this end: three tools not traditionally implemented when reporting investment performance.