One of the things that struck me after the drop last week was the level of my SPX base indicator at Wednesday’s close. My base line is drawn from a combination of several other indicators combined (via formulas I developed through much trial and error) and normalized to fit into a 1-1000-point framework. The SPX baseline at the close Wednesday October 10th was 28.74. As it turns out, that was the lowest I had ever recorded on the SPX. In fact, this was only the fourth close below 50 I had ever recorded. Among the things that experience has taught me after watching my stuff is that extremes are NOT good fades in the short term. When the baseline indicator closes above 900 (or has several closes in a row above 850ish) the market tends to stay at or near the highs and even moves higher most of the time. Selling attempts get bought and it takes a while for the extreme bullishness to wear off. The exact same thing happens in reverse when the indicator closes below 100, although it isn’t uncommon for the extreme bearishness to be reversed quicker. And the reason for that is simple. The MAJORITY of players in the market are playing the market long…..period. And that will never change.
Anyway, I had done a study of closes above 900 but had never done one on closes below 100. Now I have. My data goes back to January of 1999 by design. I made the decision to do it that way because I wanted the data to more accurately reflect the period where personal computers came into play and more players were able to access the market directly. We are talking basically 2.5 months shy of twenty years of data. In short, we are talking 4,979 trading days. In that period, there have only been 91 days where the SPX base indicator closed below 100 including the two days last week. So, this happens on average once every 55 trading days although there are times when we go over a year without even one day with a close like that.
What I did was look at closes five days after the extreme low close and 10 days after the extreme low close. At first, I treated clusters of back-to-back extreme lows as one event counting from the first day. In all those years, there have only been ten times where there were two-day, back-to-back extreme low closes. Last Wednesday and Thursday was number ten. (There have been two times with three and one time with four.) I changed my mind and decided to treat every single instance the same. That is, the clusters were treated no differently than the rest…..every low close was looked at in the five and ten day close windows.
Here are the real numbers for five day and ten day closes after the extreme low close reference day. Note: these are 5-day closes after the extreme close reference day and 10-day closes after that same day. It is NOT 10 days after the 5-day close day.
- Average 5-Day point gain is 16.05 pts.
- Average 5-Day percentage gain is 1.03%.
- Largest 5-Day point gain was after February 8, 2018 – 150.20 points.
- Largest 5-Day percentage gain was after July 23, 2002 – 13.17%.
- Average 10-Day point gain is 17.02 points. (Gains between the 5-Day and the 10-Day are quite muted, on average.)
- Average 10-Day percentage gain is .96%.
- Largest 10-Day point gain was also after February 8, 2018 – 166.30 points.
- Largest 10-Day percentage gain was after September 20, 2001 – 8.64%.
- The five day close appears to be somewhat critical., though not a slam dunk. There have been 35 instances where the close at 5 days has been below the extreme reference day close. Of those 35 instances, there have been 23 subsequent 10 day closes BELOW that same reference day close.
I could have looked at the numbers a hundred different ways, but I have spent enough time doing this as it is. One more thing that I did look at was the largest 10-Day gains. There have been 12 instances where the 10-Day gain (closing basis) has been above 5%. So I thought…..I will look at the 30-Day close for each of those occasions (30 days AFTER the low close reference day……not 30 days after the 10-Day close). However, I compared the gains at 30 days to the gains at 10 days for reference to determine if there was continuity.
Of those 12 instances, the 30-Day close has been ABOVE the 10-Day close 9 times. The average point gain has been 8.35 points and the average percentage gain has been 1.62 %. However, the numbers are skewed a bit because the most spectacular FAILURE here was also post February 8, 2018…….a LOSS of 159.04 points and a percentage loss of 5.79% compared to the 10-Day close. If the February 2018 event is left out, the average gain for the other eleven 30-Day closes (as compared to the 10-Day close) is 23.56 points and the average percentage gain is 2.29%.
One immediate takeaway here is what has happened over the last five days has been normal. The second takeaway here is (unless the Bubblebutts go completely crazy today) the somewhat muted 5-Day reaction compared to the extreme low close reference day (October 10th, 2785.68 close) bodes much better for the market going forward than it would have with a quick, wild, and sharp three-figure bounce. Of course, if they go wild into ten days, then I would be afraid of a failure on par with February's failure. Selah.