Wednesday, January 17, 2007

Narrow Range Days: What Comes Next?

A while back, I mentioned that I use a measure called the relative range to assess the market's daily volatility. The relative range represents the current day's high-low range relative to the average high-low range for the prior 20 trading sessions. In general, when we've had large declines on expanded relative ranges, we've seen bullish market behavior going forward. On Tuesday, however, we saw a particularly narrow range day in the S&P 500 Index (SPY), with a relative range of about 64% (Tuesday's range was 64% of the prior 20-day's average range). Moreover, Tuesday's was the narrowest range in the past five trading sessions.

Going back to 2004 (N = 759 trading sessions), we've had 152 NR5 days: days in which the current range was the narrowest in the last five trading sessions. When the NR5 days also were days in which the relative range was below 70% (N = 102), the next day in SPY averaged a loss of -.11% (46 up, 56 down). Conversely, when the NR5 day was a day with a relative range of 70% or greater (N = 50), the next day in SPY averaged a gain of .17% (29 up, 21 down). Overall, for the entire sample, the one-day average price change in SPY has been .03% (415 up, 344 down). Narrow ranges with respect to the past five and twenty trading sessions have been associated with subnormal near-term returns.

In general, returns after low relative range days have not been impressive. Since 2004, we've had 187 occasions in which the relative range has been below 70%. The next day in SPY has averaged a loss of -.03% (92 up, 95 down). For the remainder of the sample, the next day in SPY has averaged a gain of .06% (323 up, 249 down). When we've had a small relative range and SPY has been down for the day (N = 65)--as was the case Tuesday--the next day in SPY has averaged a loss of -.06% (28 up, 37 down).

These numbers are not so dramatic as to suggest selling on narrow ranges, but they certainly show that there is no bullish edge to a narrow day in the S&P 500 Index. Pretty consistently, narrow days have yielded below average returns in the short run since 2004.

9 comments:

Anonymous said...

Is there any evidence that looking at NR7 yields a better outcome than 5 days?

Anonymous said...

Interesting volatilty figures. They compare favorably with my own research (based on closing price rather than ranges) that low volatility tends to persist.

Brett Steenbarger, Ph.D. said...

Hi David,

Interesting question; thanks. I'd be surprised if there's a huge difference; my guess is that something like NR20 might be interesting. But I'll take a look and post results if there's anything exciting. Thanks for the idea! It would be interesting to test a volatility pattern that could be part of a Trade Ideas screen.

Brett

Brett Steenbarger, Ph.D. said...

Hi John,

You're right: volatility does display serial correlation. Traders often think big breakouts follow periods of low volatility. That isn't the case overall, however. Thanks for the observation--

Brett

Anonymous said...

Doc,

Interesting stuff. Thanks for writing the book. Great stuff. I recently added an outline of my trading plan based on 120 pages of Enhancing TP to my blog. I hope many newbies read the post and buy your book. I may have only understood 10% of the book but it's a damn good 10%. I have also decided to slow down trading and focus on finding out what really interests me about the markets and then work on my strengths. I plan on creating a strategy and practicing it till it becomes natural. I blamed psychology for my lack of profits but now I realize that a lack of knowledge and experience is responsible for my overall problems. My next book will be Mr. Forman's book.

Thanks,
LP

Brett Steenbarger, Ph.D. said...

Thanks for the note, LP. John Forman's book is an excellent resource--

Brett

Anonymous said...

Hi Doc. I love NR7s, I find them most useful in the context of the range, IE if it comes after a sharp advance or after a sharp decline. In that same context of where it is in the range, it dosn't have to be an NR7 or NR5, but can be a doji close, where the close and the open are almost the same.

Brett Steenbarger, Ph.D. said...

Hi Cucca,

Good point; it's worth looking at the context in which the narrow day appears. I'll be tackling that topic in a future post; thanks for the comment--

Brett

Sudarshan Sukhani said...

I track NR7 days on the end of day charts. I do not consider NR7 days that are visibly inside a trading range. For the others, I assume these days will mlead to a reversal or a new breakout in the original trend - a consolidation.
I step into a lower time frame, the next day, on a five minute chart, then take trades only in the direction of the trend as shown on the five minute. Some EOD charts for the NSE50 Index (India) are posted on my blog:
http://www.indiatechnicals.blogspot.com/