Market Regime Indicators for Algorithmic Trading

I recently came across a few articles about using market regime indicators in one’s algorithmic trading strategies.  Market regimes are trending periods in the market, namely, a bullish market or a bearish market.  Depending on one’s trading style, having a quantitative way of ascertaining whether the market is trending up or down is important and can help manage risk.

The first article is from SystemTraderSuccess.com, called Testing Market Regime Indicators.  In this article, Jeff Swanson tests the 200 period simple moving average, the rate of change indicator, the smoothed adaptive momentum indicator, and the relative strength ranking indicator to see which is most effective in determining the market’s overall trend.  I won’t spoil the outcome, so please read the article to see the nuances behind his conclusions.  As I’ve propounded several times here, the character of the market changes, and when it does, old strategies and indicators will fail, while new ones will provide an edge.  Such is the case with the market regime indicators that Swanson tested, as the indicator that he crowns is not unilaterally the best in all types of markets.  He’s more recently followed up with another article using the ROC, RSI, and TSI indicators in another variation of trading based on market regime.

The next article comes from Jay Kaeppel at Optionetics.com, appropriately titled, “It Doesn’t Have to be Rocket Science – Example 267.”  In the article, Kaeppel demonstrates how it’s possible to trade the stock market based on an indirect indicator, which in this particular example is the high yield bond index (he also has an article on using the Nasdaq 100 volatility index to trade high yield bonds, if you want to see another example).  Please read the article for his analysis.

I think pairing these two articles is a great way to show how to begin thinking about algorithmic trading.  One can take various indicators and test them to see how they do, then begin to combine them to see if performance can be improved.  Taking it a step further, one can introduce external data to see how it influences the market one is trading with the caveat that one should be careful not to let things get out of hand by throwing in the kitchen sink to fit the model to the data.

(via thewholestreet.com)

Leave a Reply