Anticipate Market Crashes with this Two-Day Price Pattern

A much weaker-than-expected employment report sent stocks skidding lower Friday. Institutional selling was heavy, reflected by the -75 closing TICKscore and -106,800 closing Cumulative TICK. S&P futures posted an unfilled downside gap, only the fourteenth unfilled downside gap on a Jobs Report Friday since 1990. Previous occurrences led to a flat-to-down day for the S&P one day later in all but one case…

Unfilled Downside Gap on Jobs Report Friday
06/04/10… S&P futures ??? next session
07/02/09… S&P futures +0.2% next session
01/04/08… S&P futures -0.1% next session
09/07/07… S&P futures -0.3% next session
08/05/05… S&P futures -0.2% next session
08/06/04… S&P futures +0.1% next session
07/06/01… S&P futures +0.5% next session
03/09/01… S&P futures -4.2% next session
11/07/97… S&P futures -0.6% next session
07/05/96… S&P futures -0.7% next session
04/08/96… S&P futures -0.3% next session
03/08/96… S&P futures +1.1% next session (*)
08/05/94… S&P futures +0.1% next session
05/06/94… S&P futures -1.0% next session

Open interest in E-mini S&P futures jumped by over 100,000 contracts on Friday’s selloff, another generally negative sign. That represents more than a 3% increase in open interest, and when that large an increase coincides with an unfilled downside gap, it usually precedes further weakness over the next 4-5 sessions. Below is every instance since 2002 in which ES open interest increased 3% or more on a day when the S&P’s intraday high is less than the previous day’s low…

ES Open Interest +3% on Unfilled Downside Gap
06/04/10… S&P futures ???
05/20/10… S&P futures -0.8% four sessions later
02/17/09… S&P futures -5.2% four sessions later
10/06/08… S&P futures -15.4% four sessions later
06/26/08… S&P futures -1.7% four sessions later
02/05/08… S&P futures -0.4% four sessions later
01/15/08… S&P futures -5.7% four sessions later
01/04/08… S&P futures -0.1% four sessions later
09/07/07… S&P futures +1.7% five sessions later
02/27/07… S&P futures -1.7% four sessions later
05/17/06… S&P futures -1.3% four sessions later
09/21/05… S&P futures +0.6% five sessions later
08/06/04… S&P futures +0.2% five sessions later
07/06/04… S&P futures -0.1% four sessions later
05/10/04… S&P futures +0.2% five sessions later
07/10/03… S&P futures -0.8% five sessions later
02/04/03… S&P futures -1.6% four sessions later
09/12/02… S&P futures -2.9% four sessions later
07/19/02… S&P futures -0.9% four sessions later
03/20/02… S&P futures -0.9% four sessions later
02/19/02… S&P futures +2.2% five sessions later

Out of 20 occurrences since 2002, 15 led to a lower S&P 4-5 trading days later. Only 2 out of 20 led to any sort of meaningful gain (>0.6%), suggesting limited upside potential this coming week. Note that because open interest always increases sharply in the week before expiration, any occurrences of this pattern in the five days preceding the quarterly expiration were not considered.

With the major averages all falling in excess of 3% on Friday, I’d like to bring your attention to a rare but surprisingly accurate ‘crash warning’ pattern. Utilizing Dow Industrials data back to 1900, I noticed that all major market crashes have been preceded by a very similar two-day price pattern. It begins with a drop of at least 3% on day one. Day two sees the Dow once again drop hard, recording at least a 2% drop but less than day one on a percentage basis. That’s the trigger for a crash or at least a sharp down day (>3%) within the next week.  The most bearish thing the market can do after a big percentage drop is to stage another solid drop that’s not as big as the first day’s. That indicates sellers are maintaining the upper hand but that pressure hasn’t reached climactic levels. The result is usually lower prices over the short-term and quite often a large one-day selloff of more than 3% within the next five sessions. There have been 26 occurrences of this pattern since 1900 utilizing Dow Industrials data, all of which are listed in the table below…

Dow Industrials Close Down 2%+ Following 3%+ Down Day,
Down Less than Previous Session, Loss Does Not Exceed 5%
11/17/08… Dow -3.3% two sessions later
11/06/08… Dow -0.0% three sessions later
10/27/08… Dow +12.3% three sessions later
10/08/08… Dow -7.3% one session later
07/22/02… Dow -1.1% one session later
10/15/87… Dow -4.6% one session later
11/20/73… Dow -2.3% three sessions later
09/10/46… Dow +3.6% three sessions later
12/09/41… Dow -0.2% one session later
10/08/32… Dow -4.4% one session later
09/13/32… Dow -5.7% one session later
05/25/32… Dow -3.3% two sessions later
04/29/32… Dow -1.0% two sessions later
04/06/32… Dow -0.4% one session later
01/04/32… Dow -0.5% one session later
12/09/31… Dow -2.0% one session later
11/27/31… Dow -1.7% one session later
10/14/31… Dow +5.1% three sessions later
09/30/31… Dow -1.0% one session later
09/19/31… Dow -0.8% one session later
09/14/31… Dow -0.6% one session later
06/21/30… Dow -1.5% two sessions later
11/12/29… Dow -5.3% one session later
10/24/29… Dow -0.2% two sessions later
12/07/28… Dow -5.1% one session later
05/08/15… Dow -1.1% one session later

Note that in 23 out of 26 occurrences the Dow posted a subsequently lower close (below the setup day’s close) within the next three sessions. That’s an impressive record considering the market is already extremely oversold following heavy back-to-back losses.

Even more interesting is the tendency for this price pattern to lead to a large 3%+ down day within the next five sessions. Consider that since 1900, there have been 328 days in which the Dow Industrials fell more than 3% in a single session. That’s out of 29,755 sessions, meaning the random chance for a 3%+ down day has been about 1 out of 100. Contrast that with the track record below, which is a reproduction of the table above. Instead of noting the Dow’s performance relative to the setup day, I’ve noted the largest single down day within the next week…

Dow Industrials Close Down 2%+ Following 3%+ Down Day,
Down Less than Previous Session, Loss Does Not Exceed 5%
11/17/08… Dow -5.1% two sessions later
11/06/08… Dow -4.7% four sessions later
10/27/08… Dow -0.8% two sessions later
10/08/08… Dow -7.3% one session later
07/22/02… Dow -1.1% one session later
10/15/87… Dow -22.6% two sessions later
11/20/73… Dow -3.4% three sessions later
09/10/46… Dow -0.3% two sessions later
12/09/41… Dow -0.3% three sessions later
10/08/32… Dow -4.4% one session later
09/13/32… Dow -5.7% one session later
05/25/32… Dow -6.2% four sessions later
04/29/32… Dow -2.2% three sessions later
04/06/32… Dow -5.0% two sessions later
01/04/32… Dow -2.2% five sessions later
12/09/31… Dow -3.4% two sessions later
11/27/31… Dow -3.6% four sessions later
10/14/31… Dow -0.2% three sessions later
09/30/31… Dow -6.8% four sessions later
09/19/31… Dow -7.1% four sessions later
09/14/31… Dow -5.5% four sessions later
06/21/30… Dow -3.5% two sessions later
11/12/29… Dow -5.3% one session later
10/24/29… Dow -12.8% three sessions later
12/07/28… Dow -5.1% one session later
05/08/15… Dow -4.2% five sessions later

Note the consistent theme of unusually large down days within a week of this price pattern. Both the ’87 crash and the ’29 crash occurred within three days of this price pattern, as did the mini-crash in October ’08 and numerous instances in the 30’s. In total, 19 out of 26 occurrences (73%) saw the S&P drop 3%+ in a single session within the next week. The random chance for a 3%+ down day within the next week is only 4%, so this is a major edge over random. When a 3%+ down day for the Dow is followed by a slightly less intense but still solid down day, it’s a crash warning. For this setup to be triggered Monday, the Dow needs to close down at least 2% but less than 3.2%, so it’s a pretty tight window. Even if not triggered Monday, something tells me this won’t be the last time we’ll be on crash watch.

Categories : Archived EOD Commentary, End-of-Day Commentary

 Print This Post