Wednesday, December 13, 2006

Stock Market Psychology: What You See Is Not What You Get

In my recent post, I showed how unaided perception of the market can be misleading. A market that looks as though it is rising over three different time periods ends up having worse near-term returns than an average market. Because of this, the naive trader--one who makes decisions solely based upon the appearance of a chart--is at an objective disadvantage relative to the trader who digs beneath appearance with careful analysis.

How about declining markets? Does a visual appearance of a downtrend alert us to future weakness, or is it a similar perceptual distortion when we extrapolate a negatively sloped line into the future?

Since 2004 (N = 734 trading days), we've had 134 occasions in which the S&P 500 Index (SPY) has been down on a one, four, and nine day basis. We could reasonably call this a short-term "downtrend". Four days later, SPY is up by an average .37% (84 up, 50 down), which is stronger than the average four-day gain of .08% for the remainder of the sample (330 up, 270 down). In other words, a market that is likely to be perceived as weak--down simultaneously across three time frames--has shown better returns than the average market.

Once again, however, let's extend our experiment. We'll pull out those occasions in which the market is down more on a nine-day basis than on a four-day basis and down more on a four-day basis than on a one-day basis. That creates our visual downtrend (N = 47). When the market looks as though it's in a steady downtrend, the next four days in SPY average a gain of .65% (33 up, 14 down), much stronger than the average market. Conversely, when the market is down over the three time periods but not in a visual configuration of a downtrend (N = 87), the next four days in SPY average a gain of only .21% (51 up, 36 down). In other words, the worse the market has looked, the better its near-term returns have been.

So let's summarize the results from the two experiments. Average returns over the next four days are as follows:

SPY up over three time periods and in a visual uptrend: -.11%
SPY up over three time periods and not in a visual uptrend: -.02%
SPY down over three time periods and in visual downtrend: +.65%
SPY down over three time periods and not in visual downtrend: +.21%

The naive visual trader who bought when the chart looked strong and sold when it looked weak would have lost consistent money. Conversely, the psychologically sophisticated trader who understood the limitations of human perception could have profited quite nicely. It is not paranoia to believe that the market systematically reallocates capital from naive hands to sophisticated ones: a process of natural selection every bit as ruthless as that found in the jungle.


kalius said...

You always post reasearch like this on SP 500, do you think similar research should yield good edges on individual stocks, currencies, ETF's, commodities?

Brett Steenbarger, Ph.D. said...

Hi Kalius,

You're one step ahead of me! I will eventually expand the research to the ETF universe, which will search for edges in different sectors, trading styles (value/growth), international markets, and commodities. Thanks--


Anonymous said...

Dear Brett,

first I want to thank you for the amazing quality and productivity of your work, you are providing the trading community with invaluable material. Regarding your analysis of the "apparent trend", it looks to me that the timeframe reference that you consider does not take into account the real story. If you take the monthly charts as the reference and you do the same exercise taking into account its "apparent" trend from 2000 to 2003 and then from 2003 to today, you will end up with the conclusion that the reverse conditions based on the 1,4 and 9 days were beautiful shorting opportunities from 2000 to 2003.


The Market Speculator said...

I am really looking forward to your sector analysis!


Brett Steenbarger, Ph.D. said...

Thanks, Paul; some of the historical patterns in the sectors and international equity markets are very worth following. I hope to get that up and running for '07--


Brett Steenbarger, Ph.D. said...

Hi Stephane,

You're absolutely correct; excellent point. What I'm doing in the analyses is identifying patterns in the most recent market "regimes". I'm not trying to identify patterns that could be traded mechanically across all markets.