Short term timing in the SPY ETF: Today’s price change impacts tomorrow’s return

Have you ever met an investor who feels like their short-term market timing is cursed?  If so, you are not alone.  Investors tend to buy high and sell low, which causes them to underperform the very assets that they choose to invest in.   If you are unfamiliar with this concept, you might like to read the following article.

In our last article on short term stock index timing, we used just one day’s worth of data to document that market strength tends to be followed by below average returns, and market weakness tends to be followed by above average returns.  In other words, by acting in a way that is contrary to the average investor’s impulses, the disciplined trader can develop a trading edge.

By the way:  I have an article coming out in the December 2012 issue of Active Trader Magazine that fleshes out this concept and develops an interesting trading model.  If you like swing trading ideas, I encourage you pick up a copy when it comes out in November.

Lets extend this concept and look at how today’s percent change (yesterday’s close to today’s close) impacts the average return over the next 24 hours.  What has been the average return over the next 24 hours (close to close) if the market is closing higher vs. if it is closing lower?

The average return after an up day was -.02%, and the average return after a down day was .10%.  Note also that the average close-to-close return for all days during the test period was .04%.  This fits the pattern that we have previously documented –  On average, short term strength is followed by weakness, and short term weakness is followed by strength.

Lets break up the data a bit further so we can look at how the magnitude of the change up vs. down has impacted the average forward return.  In the following graphic we look at percent change from yesterday’s close in segments (0 – 1%, 1 – 2%, 3%+, etc) and then look at average return over the next 24 hours:

The overall pattern is quite distinct:  Larger percent gains for the current day on average lead to larger expected losses over the next 24 hours, and larger percent losses lead to larger average gains.  Notice that these results mirror what we found in our prior article, when we applied this concept to  just one day’s worth of data.  Lets take the above concept and view it as a trading strategy.   How has buying (or staying long) when the market is down greater than 1% and holding for 24 hours compared to buy and hold?

I believe this graphic demonstrates how powerful these concepts can be.  Note, in the above study the “system” was only in the market about 15% of all trading days.

If you found this study interesting, I have posted some supplemental material here.  Specifically, we review one additional short-term pattern, and observe what happens when we combine it with the factor mentioned in this article.

How can this information be used?

  • For short term traders, the concepts outlined here and in our last short-term timing article can be refined into effective trading strategies.
  • Many intermediate term traders have been taught that buying on strength is the best approach to timing trades.  While this can work, evidence suggests that buying on short term market weakness can also be very effective.
  • Finally, this information can help investors avoid being “one of the statistics” who under-perform their own investment choices by buying high and selling low.

Enthusiasm has been building for the soon to launch NAS Trading member service – If you have yet to do so, sign up for our free email list here.  You will receive free strategy reports and have the opportunity to stay informed on the Developments at NAS Trading.

Good Trading.

Leave A Reply (2 comments so far)

  • Bobgun45

    What does the graph look like when you factor in transaction costs?

  • nastrading

     Hi Bob, trading costs are going to vary from trader to trader.  The 1%
    down idea has had an average forward 24 hour return of about .29% vs.
    the average return for all days of about .04%.   For a trader using a
    brokerage that caters to active traders, the transaction cost will be .5
    cents per share or less.   So on 1000 shares the round trip commission
    would be about $10.  Throw in another $20 for the bid/ask and that is
    $30.  So the expected return with the S&P500 at around 1400 is about
    $400, which would be reduced by $30 for transaction costs.  So at least
    over the simulation period, the return per trade that gets compounded
    is reduced by less than 10%.   Keep in mind when we look at average
    returns we are looking at averages.  There is considerable volatility
    around this average.   Regards, Nat