August 23, 2014

Do Backtesting Trading Strategies Work?

As retail trading platforms continue to add more and more professional features, mom-and-pop investors have been gaining access to backtesting capabilities. thinkorswim (owned by TD Ameritrade) has long had robust stand-alone trading software with countless tools for active traders, including a feature called thinkBack that lets investors use years of historical data to place simulated option trades.

Using recognia Strategy Builder capabilities built into my TradeKing account I was able to develop a trading strategy within a few minutes that would have returned 38.4% annually over the past 5 years! Another customer found a strategy that would have resulted in a 91.5% annualized return!!! Take that, S&P 500, with your crummy 14.1% during the same time period. Interestingly, while the tool lets you choose from a variety of stock metrics to build your strategy, it only offers results based on buying-and-holding for 3 months at a time (or selling short for 3 months), with the ability to include trailing stops if you'd like. The ability to buy-and-hold for a year or longer does appear to exist. Conveniently, holding a position for only 3 months means investors would probably be racking up some pretty hefty trading commissions with all those transactions.

When backtesting stocks, hindsight is 20/20!Fidelity is in the backtesting mix as well. There's web-based strategy testing for all customers, or the Wealth-Lab Pro desktop platform for customers who have traded more than 36 times over a 12-month period (with at least $25,000 in assets). The web-based tool goes back 10 years, while the desktop software goes back 20. Fidelity will even send you alerts when opportunities arise that match your strategies!

Not to be outdone, Schwab also offers a Strategy Tester as part of the company's StreetSmart Edge software. You're probably beginning to see a pattern here. In spite of the age old adage, "past performance does not guarantee future results," there are plenty of tools to help you analyze past performance in search of identical future results.

Ultimately, there will always be traders that believe there's some magic formula that will help them beat the market and brokers have the right to offer capabilities to help them achieve this goal. But ultimately, there are people trading against you everyday who have been doing this for years, or who have a PhD is mathematics or some other hyper-quantitative discipline, with computers than exist solely to recognize patterns and operate within millionths of a second. And investors need to ask themselves why brokers offer these tools. More confidence in your trading ideas = more trades, and more trades = more commissions.

I do believe there's a lot of value in analyzing long-term historical data to identify how a particular ETF or asset class has generally performed in the past, how it may correlate to other assets, and to identify high-level patterns (for example, historical data shows small cap stocks pretty consistently more volatile and prone to bigger swings than large caps). But trading short-term based on historical strategies is too rich for my blood.

Before you consider trading based on backtested results, ask yourself how many strategies you developed that underperformed the market. I was probably about 50/50 on the strategies I tried, and a few of them dramatically underperformed. Hindsight is 20/20!

1 comment:

  1. i am inspired from this post....the individual who make this post it was an extraordinary human..