Utilizing R-Squared and Linear Regression Slope To Gauge the Lasting Power of the Current Trend
By
Rennie on Wednesday, March 25th, 2009 at 2:25 am
Stock indexes posted solid losses Tuesday, yet volume was light and institutional selling was muted. TICKscore settled at -1, Cumulative TICK +10,000. Note that the 20-day moving average of the Cumulative TICK remains in a solid uptrend, and that’s not likely to change anytime soon. The extreme negative readings from early March have yet to fall off, meaning the moving average will continue moving higher over the next week on even flat-to-negative readings.
Breadth closed in 2:1 negative territory, but Tuesday’s low volume warns against expecting much follow-through. NYSE volume hit a one-week low, NASDAQ volume hit a one-month low, and SPY volume also hit a one-month low, reflecting light institutional participation. Since the inception of the SPY, there have only been twelve other occurrences in which SPY volume hit a two-week low on the same day that NYSE breadth closed at least 2:1 negative. Note that in every case but one the SPY was trading at a similar level or higher four sessions later, suggesting we may not see much downside follow-through over the short-term…
SPY Volume at Two-Week Low, NYSE Breadth 2:1 Negative
03/24/09… S&P500 ??? four sessions later
12/22/08… S&P500 -0.3% four sessions later
12/15/08… S&P500 +2.2% four sessions later
09/22/08… S&P500 +0.5% four sessions later
08/18/08… S&P500 +1.1% four sessions later
07/28/08… S&P500 +2.1% four sessions later
09/17/07… S&P500 +3.3% four sessions later
08/27/07… S&P500 +0.5% four sessions later
10/27/05… S&P500 +3.0% four sessions later
02/12/03… S&P500 +3.2% four sessions later
11/11/02… S&P500 +3.8% four sessions later
10/07/02… S&P500 +6.4% four sessions later
08/10/98… S&P500 -1.9% four sessions later
07/22/96… S&P500 +0.3% four sessions later
03/10/94… S&P500 +1.2% four sessions later
NASDAQ breadth also closed 2:1 negative, which is interesting considering that the Nasdaq100 posted an ‘inside day’, meaning Tuesday’s high was less than the previous day’s high and Tuesday’s low was greater than the previous day’s low. Normally you’ll see a break of the previous day’s low on a 2:1 negative breadth session. Since 1990, there have only been fifteen cases in which the NDX posted an inside day on a 2:1 negative breadth session. Note the tendency for the Nasdaq to trade higher over the next few sessions…
Nasdaq Breadth 2:1 Negative & Inside Day
03/24/09… NDX ??? three sessions later
12/15/08… NDX +1.9% three sessions later
11/14/08… NDX -7.8% three sessions later
10/27/08… NDX +14.0% three sessions later
07/14/08… NDX +3.1% three sessions later
11/19/07… NDX +0.4% three sessions later
07/27/07… NDX -0.6% three sessions later
06/22/07… NDX +0.6% three sessions later
03/02/07… NDX +0.6% three sessions later
01/05/07… NDX +1.7% three sessions later
08/11/06… NDX +5.5% three sessions later
10/18/05… NDX +1.8% three sessions later
02/24/03… NDX +0.1% three sessions later
01/30/03… NDX -1.4% three sessions later
10/09/02… NDX +11.2% three sessions later
09/20/01… NDX +2.0% three sessions later
Neither of these setups are rock-solid buy signals given their small sample size. They do imply generally higher prices, however, which fits with the potential ‘runaway’ scenario outlined in recent columns. And there’s further evidence in that regard. First, a description of R-squared and linear regression as described in the charting section of the site…
Linear Regression is a statistical tool that utilizes what?s known as the ?least squares? method to plot a straight line through price. Similar to a moving average, the linear regression line is often used to better understand the underlying trend. Unlike a moving average, however, Linear Regression does not exhibit as much delay. Since the indicator is ?fitting? a line to the data points rather than averaging them, the Linear Regression line is more responsive to price changes. This chart highlights two important technical indicators based on the linear regression technique – R-Squared and Linear Regression Slope. Slope tells you which direction the linear regression line is pointing (above zero is bullish, below zero is bearish), while R-Squared attempts to quantify the strength of the trend. Specifically, r-squared values show the percentage of movement that can be explained by linear regression. For example, if the r-squared indicator is currently at 70%, it means that 70% of the movement of the S&P500 can be explained by linear regression. The other 30% would be considered random noise.
On Tuesday, the R-Squared figure for the S&P500 hit 90%, reflecting an unusually strong uptrend. It’s been nearly two years we’ve seen a 90% reading with a positive slope. Historically, this has been a good indication that the S&P is likely to remain on firm footing over the next ten trading days. The last thirty occurrences are noted in the table below, along with the S&P’s performance over the next two weeks…
SPX R-Squared at 90%, Linear Regression Slope Positive
03/24/09… S&P500 ??? two weeks later
04/25/07… S&P500 +1.2% two weeks later
11/23/05… S&P500 -0.7% two weeks later
11/09/04… S&P500 +1.1% two weeks later
12/29/03… S&P500 +1.1% two weeks later
06/06/03… S&P500 +0.8% two weeks later
03/08/02… S&P500 -1.3% two weeks later
04/23/01… S&P500 +3.2% two weeks later
01/25/01… S&P500 -1.8% two weeks later
12/31/99… S&P500 -0.3% two weeks later
11/18/99… S&P500 +0.6% two weeks later
07/02/98… S&P500 +3.5% two weeks later
03/20/98… S&P500 +2.1% two weeks later
02/27/98… S&P500 +1.8% two weeks later
02/10/98… S&P500 +2.3% two weeks later
06/19/97… S&P500 +2.1% two weeks later
01/20/97… S&P500 +1.3% two weeks later
11/13/96… S&P500 +3.3% two weeks later
09/19/96… S&P500 +1.4% two weeks later
05/21/96… S&P500 +0.8% two weeks later
04/29/96… S&P500 +1.1% two weeks later
01/29/96… S&P500 +6.0% two weeks later
12/08/95… S&P500 -0.9% two weeks later
09/12/95… S&P500 +0.9% two weeks later
07/14/95… S&P500 +0.5% two weeks later
05/04/95… S&P500 -0.2% two weeks later
08/27/93… S&P500 +0.3% two weeks later
03/08/93… S&P500 -1.3% two weeks later
10/29/92… S&P500 +0.5% two weeks later
01/03/92… S&P500 -0.1% two weeks later
06/03/91… S&P500 -2.0% two weeks later
The S&P was higher two weeks later in the majority of cases, but more significant is that downside potential was limited to 2% or less in every case. This implies an area of intermediate-term support around SPX 790 that should be good through April 7th.
Utilizing R-Squared and Linear Regression Slope To Gauge the Lasting Power of the Current Trend
By Rennie on Wednesday, March 25th, 2009 at 2:25 amStock indexes posted solid losses Tuesday, yet volume was light and institutional selling was muted. TICKscore settled at -1, Cumulative TICK +10,000. Note that the 20-day moving average of the Cumulative TICK remains in a solid uptrend, and that’s not likely to change anytime soon. The extreme negative readings from early March have yet to fall off, meaning the moving average will continue moving higher over the next week on even flat-to-negative readings.
Breadth closed in 2:1 negative territory, but Tuesday’s low volume warns against expecting much follow-through. NYSE volume hit a one-week low, NASDAQ volume hit a one-month low, and SPY volume also hit a one-month low, reflecting light institutional participation. Since the inception of the SPY, there have only been twelve other occurrences in which SPY volume hit a two-week low on the same day that NYSE breadth closed at least 2:1 negative. Note that in every case but one the SPY was trading at a similar level or higher four sessions later, suggesting we may not see much downside follow-through over the short-term…
SPY Volume at Two-Week Low, NYSE Breadth 2:1 Negative
03/24/09… S&P500 ??? four sessions later
12/22/08… S&P500 -0.3% four sessions later
12/15/08… S&P500 +2.2% four sessions later
09/22/08… S&P500 +0.5% four sessions later
08/18/08… S&P500 +1.1% four sessions later
07/28/08… S&P500 +2.1% four sessions later
09/17/07… S&P500 +3.3% four sessions later
08/27/07… S&P500 +0.5% four sessions later
10/27/05… S&P500 +3.0% four sessions later
02/12/03… S&P500 +3.2% four sessions later
11/11/02… S&P500 +3.8% four sessions later
10/07/02… S&P500 +6.4% four sessions later
08/10/98… S&P500 -1.9% four sessions later
07/22/96… S&P500 +0.3% four sessions later
03/10/94… S&P500 +1.2% four sessions later
NASDAQ breadth also closed 2:1 negative, which is interesting considering that the Nasdaq100 posted an ‘inside day’, meaning Tuesday’s high was less than the previous day’s high and Tuesday’s low was greater than the previous day’s low. Normally you’ll see a break of the previous day’s low on a 2:1 negative breadth session. Since 1990, there have only been fifteen cases in which the NDX posted an inside day on a 2:1 negative breadth session. Note the tendency for the Nasdaq to trade higher over the next few sessions…
Nasdaq Breadth 2:1 Negative & Inside Day
03/24/09… NDX ??? three sessions later
12/15/08… NDX +1.9% three sessions later
11/14/08… NDX -7.8% three sessions later
10/27/08… NDX +14.0% three sessions later
07/14/08… NDX +3.1% three sessions later
11/19/07… NDX +0.4% three sessions later
07/27/07… NDX -0.6% three sessions later
06/22/07… NDX +0.6% three sessions later
03/02/07… NDX +0.6% three sessions later
01/05/07… NDX +1.7% three sessions later
08/11/06… NDX +5.5% three sessions later
10/18/05… NDX +1.8% three sessions later
02/24/03… NDX +0.1% three sessions later
01/30/03… NDX -1.4% three sessions later
10/09/02… NDX +11.2% three sessions later
09/20/01… NDX +2.0% three sessions later
Neither of these setups are rock-solid buy signals given their small sample size. They do imply generally higher prices, however, which fits with the potential ‘runaway’ scenario outlined in recent columns. And there’s further evidence in that regard. First, a description of R-squared and linear regression as described in the charting section of the site…
On Tuesday, the R-Squared figure for the S&P500 hit 90%, reflecting an unusually strong uptrend. It’s been nearly two years we’ve seen a 90% reading with a positive slope. Historically, this has been a good indication that the S&P is likely to remain on firm footing over the next ten trading days. The last thirty occurrences are noted in the table below, along with the S&P’s performance over the next two weeks…
SPX R-Squared at 90%, Linear Regression Slope Positive
03/24/09… S&P500 ??? two weeks later
04/25/07… S&P500 +1.2% two weeks later
11/23/05… S&P500 -0.7% two weeks later
11/09/04… S&P500 +1.1% two weeks later
12/29/03… S&P500 +1.1% two weeks later
06/06/03… S&P500 +0.8% two weeks later
03/08/02… S&P500 -1.3% two weeks later
04/23/01… S&P500 +3.2% two weeks later
01/25/01… S&P500 -1.8% two weeks later
12/31/99… S&P500 -0.3% two weeks later
11/18/99… S&P500 +0.6% two weeks later
07/02/98… S&P500 +3.5% two weeks later
03/20/98… S&P500 +2.1% two weeks later
02/27/98… S&P500 +1.8% two weeks later
02/10/98… S&P500 +2.3% two weeks later
06/19/97… S&P500 +2.1% two weeks later
01/20/97… S&P500 +1.3% two weeks later
11/13/96… S&P500 +3.3% two weeks later
09/19/96… S&P500 +1.4% two weeks later
05/21/96… S&P500 +0.8% two weeks later
04/29/96… S&P500 +1.1% two weeks later
01/29/96… S&P500 +6.0% two weeks later
12/08/95… S&P500 -0.9% two weeks later
09/12/95… S&P500 +0.9% two weeks later
07/14/95… S&P500 +0.5% two weeks later
05/04/95… S&P500 -0.2% two weeks later
08/27/93… S&P500 +0.3% two weeks later
03/08/93… S&P500 -1.3% two weeks later
10/29/92… S&P500 +0.5% two weeks later
01/03/92… S&P500 -0.1% two weeks later
06/03/91… S&P500 -2.0% two weeks later
The S&P was higher two weeks later in the majority of cases, but more significant is that downside potential was limited to 2% or less in every case. This implies an area of intermediate-term support around SPX 790 that should be good through April 7th.