Category Archives: ROC Studies
I’ve said over and over how hard it is to find a setup on the indices that produces a bearish result over an intermediate term. Almost all setups confirm that the market likes to go up more than go down. Imagine my surprise to find that the past couple of days have made a truly bearish setup.
- Buy $SPY at the close when it gains more than 4% in 2 days.
- Sell $SPY at the close X days later.
- No commissions or slippage included.
- All $SPY history used.
Some Additional Stats:
Next Day Winning Percentage: 47.62%
5 Day Winning Percentage: 60.00%
Number of Setups: 63
Number of Trades Held 50 Days: 22
It looks like another spike higher is to be expected over the next week. But if history is our guide, any move higher should be sold. $SPY has fallen, on average, greater than 2% over the next 3 weeks.
Previous Post: High Tight Flag for Thursday
Recently, the Dow Jones has been outperforming the S&P 500. The study will demonstrate that the greater the outperformance of the Dow Jones vs. the S&P 500, the worse the intermediate term results for both indices.
The performance of the Dow Jones and S&P 500 has diverged. The Dow is making new highs while the S&P is struggling to surpass the highs made in October, November, and December 2o11.
I somewhat randomly chose the 30 day rate of change (ROC30) of the Dow Jones (DIA) and S&P 500 (SPY) to measure the difference in performance of the two indices. When looking at their charts, both indices appeared to have bottomed about 60 days ago. I just split the number in half. Unscientific? Maybe. I will take a look at the 60 day rate of change as well, but not in this post.
What does the outperformance by the Dow Jones mean over the intermediate term, if anything?
Buy SPY or DIA at the close if
- The difference between the DIA ROC30 and the SPY ROC30 is greater than X.
- Sell at the close Y days later.
- No commissions or slippage included.
- All SPY and DIA history used.
I know the graph is busy. Bear with me. We’ll start from top to bottom of the legend on the right side of the graph.
The white and pink lines show a small divergence between DIA and SPY where the difference between the two is more than 0.5%. Note that the average performance of both ETFs track each other fairly well and make several forays into positive territory.
Next are the blue and red lines. With the difference greater than 1%, we began to see the DIA (red line) outperforming SPY (blue line). This is not really ground-breaking. I think it just shows that the Dow’s momentum persists over the time period. Both ETFs finish in positive territory after 50 days.
As we examine the green and purple lines where the DIA’s ROC30 is 1.5% or more greater than SPY’s, performance begins to head into negative territory around the 17th day. While DIA still outperforms SPY, neither ETF manages to get back to positive territory.
Finally, the brown and orange lines show that as the DIA outperformance grows to 2% or more, performance over the next 50 days worsens.
What Does It All Mean?
I’m not sure. I’m not sure why performance for the differences of greater than 1.5% begins to drop around the 17th day. Perhaps when capital moves to the Dow it is a flight to safety, and after a month or so, whatever was perceived as risk has abated.
There were plenty of samples. Even the most restrictive requirement of > 2% still had 30 samples at 50 days out. When the divergence begins around the 17th day, there were 68 samples averaged to show the > 1.5% result.
The issue I can’t ignore is that it is pretty difficult to produce an average return for these ETFs that doesn’t finish in positive territory after 50 days. Markets have a tendency to go up, and most returns over 50 day periods will reflect this. I have looked at dozens of setups (if not a hundreds or more), and even when I was sure they were bearish over an intermediate term, most of the time the results showed they were in fact neutral to bullish. Therefore it is significant to me that the >1.5 and >2 setups both produce a negative return after 50 days.
This is a very simple study. Perhaps I’ve lucked into something using ROC30. In a future study I will lengthen the ROC and see what happens. Also, I would like to require successive days where the difference is greater than X to see how that changes the results.
I welcome any thoughts and observations.
The 7 day rate-of-change (ROC7) for SPY is 9.84%. Looking back over 40 years, a ROC7 of greater than 9% has occurred only 44 times. Standard thinking would have us looking for a pullback here. Testing this setup shows that the move higher is likely to continue.
I’ve got a short position open that I rode down to the recent new low, and have ridden back up to our current close. I’m looking for a pullback here because I want to close the short for break-even and put some cash to work. Plus, it just makes sense that after the market goes up almost 10% with no pullbacks, that it will pullback. Unfortunately, such a strong and forceful move like we’ve seen over the past 7 days seems to indicate a change in the market, and the change looks like it might be an uptrend. If we do see an uptrend develop, look for volatility to quickly dissipate.
Buy SPY at the close if:
- The 7 day rate of change is greater than 8% (I lowered the threshold to 8% to obtain a larger sample)
- Sell X days later
- Also look at when SPY is beneath its 200 day moving average
- Include 60 years of $SPX data
No commissions or slippage included. All SPY history used.
Analysis of Results:
SPY ROC7>8 had a sample size of 28 setups with 13 trades held for the full 50 days.
SPY <MA200 had a sample size of 21 setups with 9 trades held for the full 50 days.
$SPX < MA200 had a sample size of 31 setups with 16 trades held for the full 50 days. The first trade was in 1962.
These results do not need an in-depth explanation. They are bullish. In the past there has been a small pullback, and then it has been off to the races.
Frankly, these results do not square very well with this study about SPY re-taking the 50 day moving average. I’m inclined to believe that such a powerful move up negates the influence of the 50 day moving average.