Thursday, January 29, 2009

Reversal and Momentum Effects Following Surges in Stocks Making New Highs

One pattern that I've seen play out on multiple timeframes is that price breakouts that lead to a surge in the number of stocks making new highs tend to be followed by consolidation/reversal in the short run, but continuation thereafter. This sets up useful trades in which initial pullbacks following these surges can make nice entry points for subsequent upside moves.

Going back to late 2002, when I first began collecting these data (N = 1572), I looked at what happened in the S&P 500 Index (SPY) following days in which the number of stocks making new 20-day highs across the NYSE, NASDAQ, and ASE exceeded 1500 (N = 213). Interestingly, ten days after the surge in new highs, SPY averaged a gain of only .05% (122 up, 91 down), compared with an average gain of .08% (787 up, 572 down) for the remainder of the sample.

When we go out 30 days, however, we find that the surge in new highs leads to an average gain of .85% (143 up, 70 down), compared with an average rise of only .13% for the rest of the sample (801 up, 598 down). The surge in strength does not necessarily result in significant near-term gains, but it is common to see momentum carry the market higher in the longer run. In a future post, I will illustrate this pattern at intraday time frames. The links below illustrate the importance of understanding momentum dynamics in markets.


jpfreemon said...

Your test was run during a time frame when SPY was in a general uptrend for 5 of the 7 years tested.

Frame the data during periods when SPY was above it's 200 day MA vs periods when SPY was below it's 200 day MA.

DG said...

Reversal as you define it is measured statistically, for the aggregate market. It is helpful as a "heads-up" of when to look.

Activity-wise in the market I like to use scanning and screening to help me find the WHAT, the individual issue. Also I wonder how broad an index you like to compare price against other measures. So far, my best efforts seem to work within the WILSHIRE 5000 [ $WLSH ] [ intraday ], and Tom Holt's Value Line Arithmetic, which would be UNWEIGHTED [ $VLE ] as an End of Day Measure. To avoid the "Co-Linearity Trap" I tend to use Price + another measure for comparisons.


prunto said...

absolutely best trading stuff on the web...keep it coming

ivanhoff said...

Great post, Dr. Brett. I have similar experience with individual stocks at different time frames. High volume and huge price appreciation during the first several bars (I use 10 min), followed by sideways or priceways correction above the daily vwap, followed by a new breakout.