stocks

Limitations of Using Backtests for Stock Market Research

Stock Market Guides is not a financial advisor. Our content is strictly educational and should not be considered financial advice.

Stock Market Guides relies on backtests for its research.

A backtest involves taking any given trading strategy, writing it in code (an algorithm), and applying it against historical prices of stocks to see how the trading strategy might have performed in the past.

 

Backtests Involve Theoretical Trades, Not Live Trades

Although backtesting offers powerful insights about which trading strategies might have an edge, it's important to understand that backtests are not live results. Since we didn't make these trades live, we can't say with certainty what exactly our entry and exit prices would have been.

There are market conditions that could have affected our ability to enter or exit the trades exactly as the backtests show. A sudden lack of liquidity is an example of such a condition.

 

Backtests Might Reflect Hindsight

Another limitation of backtests is that they are performed with the benefit of hindsight. Sometimes being able to see the past can expose you to the risk of “curve fitting”, where a trade strategy can be pinpointed that might generate optimal backtest results but that doesn’t actually have a true long-term edge.

We work diligently to minimize this risk by employing stringent statistical requirements for our backtests. It’s nevertheless still a risk involved with relying on backtests.

 

Errors Are Possible With Backtests

Another limitation of backtests is that they are created by humans, and therefore subject to human error. There may be errors in our work.

Also, our backtests rely on data provided by third party data services. It's possible there are some errors in their data.

 

Backtests Don't Account For Brokerage Commissions

The backtest results we publish don’t include commissions, which vary by broker, although many brokerages in the USA offer commission-free stock trades. The results also don’t account for slippage.

 

Option Backtests Require Extra Assumptions

Backtesting presents more challenges with options than stocks. Options have substantially less trading volume than stocks, which makes their price history less dependable for backtesting purposes. In addition, options haven't been available for as long as stocks. For example, "weekly" options only first became available in 2005.

Because of these challenges, we created a proprietary model for backtesting option performance. That model is based on the underlying stock's price history, and we apply it to stocks going back to their inception (which, in some cases, is prior to the advent of weekly options).

We developed an internal measure of a stock's price volatility that corresponds with the extrinsic value of options. We use that to determine the strike price of the option that corresponds with the delta value we use for our option picks (94 or higher). Doing this allows us to estimate the strike price and contract price of any given trade in our backtests.

Since assumptions are involved with backtesting options, it means that our option backtest results have even more limitations than our stock backtest results.

 

Take Backtest Results In Stride

Due to all these limitations, we have to take the results of the backtests in stride. There is no way for anyone to be certain that backtests accurately reflect what live results would have been since they weren't actual live trades. And beyond that, past results aren’t a guarantee of future performance.

The bottom line is that there is real risk involved with taking trades, even when backtest results look encouraging. Stock Market Guides is not responsible for that risk; you are responsible for that risk.

If you have any questions about backtests or their limitations, you contact us any time and we'll be ready to help.