iBankCoin
Joined Nov 11, 2007
1,458 Blog Posts

S&P 500 Beneath the 50 Day Avg. for More than 50 Days, Now Above. What Now?

Today, after trading beneath the 50 day moving average for more than 50 days (52 days, to be exact), the S&P 500 surged to close above the average. What can we expect going forward in the short and intermediate time frames?

You have likely already heard the hubbub across the financial landscape about the S&P and Dow Jones averages closing above their 50 day moving averages (MA50). Frankly, these averages are helpful only if they are not used to make binary trading decisions. Just because the index closes above it, or below it, doesn’t mean one should immediately go out and buy or sell. However, a close above the MA50 does serve as a warning that a change of trend may be underway. In fact, if we do use it as a binary signal, the decision may be to buy the first pullback after a close above the MA50 (the results below will illustrate why this may be a profitable decision).

But I digress.

What we want to look at is how we should incorporate the amount of time spent beneath the average into our trading decisions. Surely, being back above the MA50 is a positive development, but just how excited should we be? Should we be chomping at the bit to initiate new positions, or should we be more patient? If we consider how long the index has traded beneath its MA50 before again closing above it, we can answer that question.

The Rules:

Buy SPY at the close if

  • It has traded for more than  X days beneath the MA50 and then closes above the average
  • Sell Y days later

No commissions or slippage included. All SPY history used.

The Results:

Analysis of Results:

Let’s start with the green line.  There were 123 setups and 34 trades held the full 100 days. A market that has traded for just a few days beneath the MA50 and then closes above it may have completed a healthy pullback. Or, the green line may represent a market that has vacillated around the MA50. Regardless, this setup is the least volatile going forward but also achieves the smallest returns after 100 days. Why the low volatility and low returns? My guess is that a market that is vacillating around the MA50 is may tend to have some momentum remaining from the previous uptrend (hence, the low volatility) but may also be subject to increasing downward pressure (hence the lower returns).

The blue line, with 19 setups and 16 trades held the full 100 days, is more rare. We can see that a market that has traded for a month or so beneath the MA50 will be more volatile going forward, after it regains the MA50.

The red line is the closest model to our current market. This condition is rare. With over 17 years of SPY data available, there were only 6 setups, and all 6 were held for the full 100 days. This setup is the most volatile going forward but also achieves the highest returns after 100 days. Note the multiple pullbacks averaging near -2%.

What Does It All Mean?

The bottom line is this: In all of the models, SPY does not achieve an average return greater than 1% until 55 trading days after the setup. The >25 and >50 models do not achieve an average return greater than 1% until 4 months (90 trading days) have passed. The risk of another downside move appears to be greater than that of an upside move.

Patience will be a profitable virtue.

While there is growing evidence that the bear market may be waning, volatility will remain very high. Going forward, this is still a trader’s market. There will be plenty of pullbacks to buy and rips to sell. Putting some cash to work on the dips may be appropriate for longer-term investors, but be prepared for a volatility whipsaw.

If you enjoy the content at iBankCoin, please follow us on Twitter

31 comments

  1. synackf1n

    Nice study.

    • 0
    • 0
    • 0 Deem this to be "Fake News"
  2. Bozo on a bus

    Looks like we can expect continuing volatility ahead.
    And on a related topic, if I am not mistaken, isn’t the ROC5 is now above the ROC252 for two days in a row? This has really been jumping back and forth.

    • 0
    • 0
    • 0 Deem this to be "Fake News"
    • Woodshedder

      Bozo, yes, the ROC5 is back above the ROC252 and has been since Friday. As of today, it would be short again. I’ll post this formally tomorrow.

      • 0
      • 0
      • 0 Deem this to be "Fake News"
      • Bozo on a bus

        OK, great. Any thoughts on what the back-and-forth action mean?

        • 0
        • 0
        • 0 Deem this to be "Fake News"
        • Woodshedder

          Bozo, back and forth on what? ROC5 and ROC252?

          • 0
          • 0
          • 0 Deem this to be "Fake News"
          • alf44

            “Woodshed” … seriously …

            …have YOU actually EVER traded ?

            I mean … REALLY TRADED ???

            Spare me your lame-ass comeback !!!

            The “answer” is obvious … you ridiculous farce !!!

            There are some good “bloggers” here !!!

            You ain’t one of them !!!

            We’re both Southern Boys … dig this …

            You’re as useless as tits on a boar hog !!!

            🙂

            .

            • 0
            • 0
            • 0 Deem this to be "Fake News"
          • Bozo on a bus

            Boy, go to bed early for a change, and miss a lot…
            Sorry, I should have been more clear.

            My thinking was in the last bear market, the switch was unambiguous – the ROC252 jumped pretty fast past the ROC5. Now, we have the signals: bull, bear, bull, all in just a few weeks. Sometime I would like to go back and see which scenario is more typical – a single fast switch, or multiple switches before the trend is clear. It’s not obvious from your chart.

            • 0
            • 0
            • 0 Deem this to be "Fake News"
            • Woodshedder

              Bozo, I’m not sure it really matters. I did go back a bit myself, and what I found were most switches stuck, or a single fast switch, as you describe. But there were some that had a few switches before settling on one side.

              • 0
              • 0
              • 0 Deem this to be "Fake News"
  3. Superpositron

    Nice stuff Wood. As always.

    • 0
    • 0
    • 0 Deem this to be "Fake News"
  4. huh?

    Is the flame fest over kids? Jessus, Im sure the majority of people here are older then me, but sure as hell dont act it.

    This is what Im watching. I could be wayyy off and the number of head fakes blow my mind.

    http://imageshack.us/f/856/carryonsir.png/

    • 0
    • 0
    • 0 Deem this to be "Fake News"
  5. RC

    lol @ thread.

    • 0
    • 0
    • 0 Deem this to be "Fake News"
  6. tendollartommy

    bs like that rant above really should be banned. it reduces the blog’s interest for everyone else. don’t like Shedder’s analysis? say so, or come up with a counter idea. but dopey name-calling ain’t even funny, it’s just plain useless….

    • 0
    • 0
    • 0 Deem this to be "Fake News"
    • Woodshedder

      Ten Dollar- I appreciate your support. I typically refrain from commenting in such an un-gentlemanly manner. Luckily, idiots like Alf are very rare on my blog. I should not have even responded to him. I apologize for my crass responses. Next time I will give serious consideration to just banning the commenter.

      • 0
      • 0
      • 0 Deem this to be "Fake News"
      • huh?

        *hat tip* good sir. We can’t be running a gentleman’s club with these non-gentleman types.

        • 0
        • 0
        • 0 Deem this to be "Fake News"
      • huh?

        *hat tip* good sir. We can’t be running a gentleman’s club with these non-gentleman types.

        • 0
        • 0
        • 0 Deem this to be "Fake News"
  7. jkw

    Are the results actually more volatile with more time beneath the moving average, or is it just a sample size issue? The standard deviation of the mean of a sample decreases as the sample size increases (1/sqrt(n)), so 134 samples will tend to have a smooth average, while 17 samples will have the average jumping all over the place, even if the individual setups all have the same volatility. If you want to compare volatility, you have to measure the volatility of each setup and then average those measurements together.

    • 0
    • 0
    • 0 Deem this to be "Fake News"