SPX 14-day RSI Triggers an ‘Overbought’ Buy Signal
By
Rennie on Sunday, March 21st, 2010 at 6:00 pm
Stocks closed lower Friday on a sharp uptick in volume due to triple witching, although Friday’s 1.98 billion shares for NYSE floor volume was actually light compared with previous triple witchings. In recent years NYSE volume has usually surpassed 2 billion shares on triple witching – June ’08 and March ’06 are the last instances when volume was under 2 billion – so in that respect it wasn’t a particularly active expiration.
Nonetheless, NYSE floor volume was up over 100% from Thursday’s session while consolidated volume only increased by 36%. This discrepancy triggered the 3-5 day pattern recently discussed in this February 17th column. If you’ll recall, when NYSE volume makes an unusually large leap relative to consolidated volume, that day’s move tends to continue over the next 3-5 sessions. With S&P futures settling below the open, this points to a close below Friday’s settlement in the Wednesday-Friday time frame of this week. However, with a limited amount of consolidated volume data to work with (just over 3 years), I’m wary of giving this signal much weight.
The run-up heading into Friday’s expiration still suggests we’ll see a rebound on Monday (see last Thursday’s column). If the volume setup above proves correct, we may not see much follow through initially. I’d also note that the 20-day moving average of the Cumulative TICK continues to trend lower after topping out on March 11th, which can be a sign of limited upside potential.
Some new charts will be debuting in the end-of-day charting section beginning next week. Among these is the percentage of stocks above their 20-day moving average. We’ve recently finished compiling the last six years of this data – here’s a look at the current chart as well as the long-term version. One point I find particularly interesting from the longer-term chart is that there wasn’t a single instance in the 2005-2007 period in which the percentage of stocks above the 20-day average exceeded 80%. Yet since the March 2009 bottom, we’ve seen this statistic over 80% not just one or two days, but for extended periods of time. For instance, we’re just coming off a string of fourteen consecutive sessions over 80%. That just missed eclipsing the record of fifteen consecutive sessions over 80% in July/August ’09 and did exceed the eleven consecutive sessions in March/April ’09. In both cases the market entered into a consolidation phase once this statistic dropped below the 80% level, but downside was limited and the market was solidly higher one, two and three months later. Two cases is hardly a convincing sample size, but the fact that we never even saw this statistic move above 80% prior to 2009 indicates the market remains in an unusually broad-based rally phase.
In this column from last Thursday I noted the SPX 14-day RSI closed over 75 and highlighted the market’s typically solid performance in the month following such an unusually high reading. The edge from the high RSI reading alone wasn’t quite strong enough to be considered significant, which is the reason it wasn’t added to the board. Today I’d like to build on that study with a setup that does show a significant historical edge. Note from the current chart that the SPX 14-day RSI closed under 70 following Friday’s selloff. I went back and reviewed all instances in which the RSI closed over 75 for the first time in at least 14 sessions (to avoid any duplicate signals) and then closed under 70. Readings over 75 are generally bullish looking out one month (this holds true over a long test period), so an initial pullback under 70 should provide a favorable entry on the long side for an intermediate-term trade. Testing backs up that assumption. Below I’ve listed every instance in which the 14-day RSI closed under 70 after closing over 75 at some point within the last ten sessions…
S&P500 14-day RSI Closes <70 after closing >75 Within Last Ten Days
03/17/10 RSI >75… 03/19/10 RSI <70… ???
08/03/09 RSI >75… 08/06/09 RSI <70… Higher S&P five sessions later
10/23/06 RSI >75… 10/27/06 RSI <70… Higher S&P six sessions later
11/12/04 RSI >75… 11/16/04 RSI <70… Higher S&P five sessions later
01/05/04 RSI >75… 01/09/04 RSI <70… Higher S&P five sessions later
11/23/98 RSI >75… 11/30/98 RSI <70… Higher S&P five sessions later
03/19/98 RSI >75… 03/27/98 RSI <70… Higher S&P five sessions later
06/12/97 RSI >75… 06/23/97 RSI <70… Higher S&P five sessions later
10/04/96 RSI >75… 10/09/96 RSI <70… Higher S&P five sessions later
02/07/96 RSI >75… 02/15/96 RSI <70… Higher S&P five sessions later
12/05/95 RSI >75… 12/07/95 RSI <70… Higher S&P five sessions later
09/11/95 RSI >75… 09/22/95 RSI <70… Higher S&P five sessions later
06/22/95 RSI >75… 06/26/95 RSI <70… Higher S&P five sessions later
03/24/95 RSI >75… 03/31/95 RSI <70… Higher S&P five sessions later
02/03/95 RSI >75… 02/17/95 RSI <70… Higher S&P five sessions later
08/30/94 RSI >75… 09/01/94 RSI <70… Higher S&P nine sessions later
02/04/93 RSI >75… 02/08/93 RSI <70… No higher S&P 5-10 sessions later
12/08/92 RSI >75… 12/10/92 RSI <70… Higher S&P five sessions later
02/05/91 RSI >75… 02/20/91 RSI <70… Higher S&P five sessions later
05/21/90 RSI >75… 05/25/90 RSI <70… Higher S&P five sessions later
07/31/89 RSI >75… 08/04/89 RSI <70… Higher S&P five sessions later
05/19/89 RSI >75… 05/23/89 RSI <70… Higher S&P five sessions later
01/27/89 RSI >75… 02/09/89 RSI <70… Higher S&P six sessions later
08/10/87 RSI >75… 08/18/87 RSI <70… Higher S&P five sessions later
06/22/87 RSI >75… 06/24/87 RSI <70… Higher S&P eight sessions later
03/23/87 RSI >75… 03/27/87 RSI <70… Higher S&P five sessions later
01/15/87 RSI >75… 01/23/87 RSI <70… Higher S&P five sessions later
02/18/86 RSI >75… 02/19/86 RSI <70… Higher S&P five sessions later
05/20/85 RSI >75… 05/22/85 RSI <70… Higher S&P six sessions later
08/07/84 RSI >75… 08/15/84 RSI <70… Higher S&P five sessions later
10/11/82 RSI >75… 10/15/82 RSI <70… Higher S&P five sessions later
In 29 out of 30 occurrences stretching back to 1982, the S&P bounced back from its first sub-70 RSI reading to close at a higher level 5-10 trading days later. That 97% accuracy level is significantly greater than the 76% random chance for a higher S&P close 5-10 sessions later in the same period of time, and implies we should see the SPX above 1159 in the March 26th-April 1st time frame.
If all setups are fulfilled, we’re looking at a rally Monday, a potential selloff into mid-week followed by a resumption of the rally into the end of the week and/or the first few days before Good Friday.
SPX 14-day RSI Triggers an ‘Overbought’ Buy Signal
By Rennie on Sunday, March 21st, 2010 at 6:00 pmStocks closed lower Friday on a sharp uptick in volume due to triple witching, although Friday’s 1.98 billion shares for NYSE floor volume was actually light compared with previous triple witchings. In recent years NYSE volume has usually surpassed 2 billion shares on triple witching – June ’08 and March ’06 are the last instances when volume was under 2 billion – so in that respect it wasn’t a particularly active expiration.
Nonetheless, NYSE floor volume was up over 100% from Thursday’s session while consolidated volume only increased by 36%. This discrepancy triggered the 3-5 day pattern recently discussed in this February 17th column. If you’ll recall, when NYSE volume makes an unusually large leap relative to consolidated volume, that day’s move tends to continue over the next 3-5 sessions. With S&P futures settling below the open, this points to a close below Friday’s settlement in the Wednesday-Friday time frame of this week. However, with a limited amount of consolidated volume data to work with (just over 3 years), I’m wary of giving this signal much weight.
The run-up heading into Friday’s expiration still suggests we’ll see a rebound on Monday (see last Thursday’s column). If the volume setup above proves correct, we may not see much follow through initially. I’d also note that the 20-day moving average of the Cumulative TICK continues to trend lower after topping out on March 11th, which can be a sign of limited upside potential.
Some new charts will be debuting in the end-of-day charting section beginning next week. Among these is the percentage of stocks above their 20-day moving average. We’ve recently finished compiling the last six years of this data – here’s a look at the current chart as well as the long-term version. One point I find particularly interesting from the longer-term chart is that there wasn’t a single instance in the 2005-2007 period in which the percentage of stocks above the 20-day average exceeded 80%. Yet since the March 2009 bottom, we’ve seen this statistic over 80% not just one or two days, but for extended periods of time. For instance, we’re just coming off a string of fourteen consecutive sessions over 80%. That just missed eclipsing the record of fifteen consecutive sessions over 80% in July/August ’09 and did exceed the eleven consecutive sessions in March/April ’09. In both cases the market entered into a consolidation phase once this statistic dropped below the 80% level, but downside was limited and the market was solidly higher one, two and three months later. Two cases is hardly a convincing sample size, but the fact that we never even saw this statistic move above 80% prior to 2009 indicates the market remains in an unusually broad-based rally phase.
In this column from last Thursday I noted the SPX 14-day RSI closed over 75 and highlighted the market’s typically solid performance in the month following such an unusually high reading. The edge from the high RSI reading alone wasn’t quite strong enough to be considered significant, which is the reason it wasn’t added to the board. Today I’d like to build on that study with a setup that does show a significant historical edge. Note from the current chart that the SPX 14-day RSI closed under 70 following Friday’s selloff. I went back and reviewed all instances in which the RSI closed over 75 for the first time in at least 14 sessions (to avoid any duplicate signals) and then closed under 70. Readings over 75 are generally bullish looking out one month (this holds true over a long test period), so an initial pullback under 70 should provide a favorable entry on the long side for an intermediate-term trade. Testing backs up that assumption. Below I’ve listed every instance in which the 14-day RSI closed under 70 after closing over 75 at some point within the last ten sessions…
S&P500 14-day RSI Closes <70 after closing >75 Within Last Ten Days
03/17/10 RSI >75… 03/19/10 RSI <70… ???
08/03/09 RSI >75… 08/06/09 RSI <70… Higher S&P five sessions later
10/23/06 RSI >75… 10/27/06 RSI <70… Higher S&P six sessions later
11/12/04 RSI >75… 11/16/04 RSI <70… Higher S&P five sessions later
01/05/04 RSI >75… 01/09/04 RSI <70… Higher S&P five sessions later
11/23/98 RSI >75… 11/30/98 RSI <70… Higher S&P five sessions later
03/19/98 RSI >75… 03/27/98 RSI <70… Higher S&P five sessions later
06/12/97 RSI >75… 06/23/97 RSI <70… Higher S&P five sessions later
10/04/96 RSI >75… 10/09/96 RSI <70… Higher S&P five sessions later
02/07/96 RSI >75… 02/15/96 RSI <70… Higher S&P five sessions later
12/05/95 RSI >75… 12/07/95 RSI <70… Higher S&P five sessions later
09/11/95 RSI >75… 09/22/95 RSI <70… Higher S&P five sessions later
06/22/95 RSI >75… 06/26/95 RSI <70… Higher S&P five sessions later
03/24/95 RSI >75… 03/31/95 RSI <70… Higher S&P five sessions later
02/03/95 RSI >75… 02/17/95 RSI <70… Higher S&P five sessions later
08/30/94 RSI >75… 09/01/94 RSI <70… Higher S&P nine sessions later
02/04/93 RSI >75… 02/08/93 RSI <70… No higher S&P 5-10 sessions later
12/08/92 RSI >75… 12/10/92 RSI <70… Higher S&P five sessions later
02/05/91 RSI >75… 02/20/91 RSI <70… Higher S&P five sessions later
05/21/90 RSI >75… 05/25/90 RSI <70… Higher S&P five sessions later
07/31/89 RSI >75… 08/04/89 RSI <70… Higher S&P five sessions later
05/19/89 RSI >75… 05/23/89 RSI <70… Higher S&P five sessions later
01/27/89 RSI >75… 02/09/89 RSI <70… Higher S&P six sessions later
08/10/87 RSI >75… 08/18/87 RSI <70… Higher S&P five sessions later
06/22/87 RSI >75… 06/24/87 RSI <70… Higher S&P eight sessions later
03/23/87 RSI >75… 03/27/87 RSI <70… Higher S&P five sessions later
01/15/87 RSI >75… 01/23/87 RSI <70… Higher S&P five sessions later
02/18/86 RSI >75… 02/19/86 RSI <70… Higher S&P five sessions later
05/20/85 RSI >75… 05/22/85 RSI <70… Higher S&P six sessions later
08/07/84 RSI >75… 08/15/84 RSI <70… Higher S&P five sessions later
10/11/82 RSI >75… 10/15/82 RSI <70… Higher S&P five sessions later
In 29 out of 30 occurrences stretching back to 1982, the S&P bounced back from its first sub-70 RSI reading to close at a higher level 5-10 trading days later. That 97% accuracy level is significantly greater than the 76% random chance for a higher S&P close 5-10 sessions later in the same period of time, and implies we should see the SPX above 1159 in the March 26th-April 1st time frame.
If all setups are fulfilled, we’re looking at a rally Monday, a potential selloff into mid-week followed by a resumption of the rally into the end of the week and/or the first few days before Good Friday.