When Markets Bottom

03/04/09 8:45 AM EST

Good Wednesday morning...We begin the day with some buying interest in the pre-market futures as we try to recover from a late dip into the close yesterday. 

One of the market platitudes I saw bandied about yesterday was "well, we can't bottom today because it's a Tuesday, and the market doesn't bottom on Tuesdays".

Besides the obvious correction that markets can do whatever the h*ll they want, whenever they want to do it, the statement above is simply false.  It's probably a good time to look at the market's tendency to bottom during certain parts of the calendar.

The charts below use the S&P 500 index from 1928 through the present.  What we're seeing are those times that the index forms at least a three-month low (a day where there are no lower closes three months before or after that date).  It is broken down by month, by day of the week, and by day of the month.

From the charts, we can see that the absolute ideal time to form an intermediate-term low would be the 9th trading day of October, especially if it's a Monday or Tuesday (by the way, in 2009 the 9th trading day of October just happens to be a Tuesday).

So what about now?  Well, March is fairly well-represented, accounting for 10% of all intermediate-term lows.  That's a greater proportion than 8 other months.  And forming a low yesterday would be the most likely of all, as Tuesdays account for 28% of all lows. 

Monday and Tuesday combined account for more than 50% of all lows, while Thursday is particularly under-represented.  If you squint, you may notice that the percentages don't quite add up to 100%, which is because there used to be trading on Saturdays, and they accounted for 4% of the lows (not shown).

As for the day of the month, that's a lot more spread out.  There are some spikes in the percentages from the 6th - 9th trading days (just after retail investors get their monthly statements?) and another one after the 14th of the month.

The worst of all the days were the 2nd and 12th of the month.  That's not so good, since yesterday was the 2nd trading day and thus least likely to see a bottom form.

These are obviously just averages and tendencies, and they are taken in a vacuum without any consideration of the fundamental, technical or sentimental condition of the market.  But it's worth knowing that the market can, indeed, bottom on a Tuesday (or a Monday, or a Wednesday...).  Just simply based on the data from the tables, it would seem more likely that we'd hit a low next Monday or Tuesday, which would be the 6th and 7th trading days of the month.

If we step out of the vacuum and look at the market's structure, it's obvious that the streak of selling pressure leading up to this point has been incredible.  At least yesterday showed the smallest loss of the past week, suggesting a drop in selling urgency.

Let's go back and look for any other time since 1928 that the S&P dropped for five straight days to a new yearly low, losing at least 5% during the stretch, with the last day showing the smallest loss of the bunch.

 DATE  -  OUTCOME

12/15/30 - 1 day from an intermediate-term low

05/19/31 - 11 days from an intermediate-term low

12/12/31 - 3 days from multi-week rally

02/10/32 - Marked the low before multi-week rally

10/22/57 - Bear-market low

06/26/62 - Bear-market low

08/15/74 - Failure

10/01/74 - 3 days before bear-market low

08/09/82 - 3 days before bear-market low

10/08/08 - 2 days before multi-week rally

Some of those days really stand out, as they marked either the end of a major decline, or within days of it.  On four of the ten occurrences, we were within a whisper of some of the greatest buying junctures in the history of the market.

Even during the early 1930's - analysts' favorite worst-case comparison to our current market - the S&P was within 3 days of at least a multi-week rally every time but once.  I'm not a huge fan of analogs, but yesterday we did look at one time period that is exceptionally close to our current market in a number of ways, and that comparison also suggested an important low should occur soon.

Speaking of streaks, there is something else notable about our current market.  Both the S&P 500 and Nasdaq Composite have been down for 11 of the past 12 days.  That has happened only one time since 1971...which was August 10, 1982.  The market formed its bear-market bottom two days later.

Bottom line - Intermediate-term

For the past couple of months, I've been repeating the same first paragraphs in this part of the summary - we had a bunch of positive studies during November/December, they failed during mid-January, so in order to get a positive outlook again we either needed another pessimistic extreme in Dumb Money Confidence, or a much-improved technical picture.

That latter condition ain't gonna happen anytime soon.  So we needed an extreme in the Dumb Money, and we reached that last week.  We haven't quite reached the 13% extreme we did at prior lows (and almost certainly won't due to stubbornly complacent put/call readings), but when we've seen the Dumb Money this low, and then saw further short-term deterioration in price like we have, it has been near inflection points (see here and here).

Given that, and other positives we've discussed recently, I've been on the lookout for signs of extreme intraday selling pressure and/or a price pattern similar to what we outlined on Tuesday.  We kinda-sorta got some of those extremes on Monday, but not quite to the panic levels reached at other lows.  As for the price pattern, we could see the 3rd part of that today, but we'd have to hold well after this morning's gap up and not retreat more than about 0.75% or it would lessen the probability we were witnessing a low.

If we fail again today, I'm still looking at that 675 area as a good spot to add a tranche of intermediate-term longs, with risk down to 640-660 or so (coinciding with a 61.8% retracement of the bottoms formed in 1974 and 1982). 

Bottom line - Short-term

Yesterday was a very choppy day that didn't do much to clear up our picture.  We didn't hold the opening gap, which violated the potential (but unlikely) pattern of forming a low, and we didn't decline enough to generate the kinds of intraday "panic" extremes we've seen at past inflection points.

At various points during the day it looked like we were either going to take off on a major upside reversal, or crash into the mid-600's, and we got neither.  Right after the close of regular trading the futures dropped to 681, but have recovered more than 20 points in a steady climb since then.

That leaves us with a large gap up (so far), which again ushers in the potential bottoming pattern.  For that to take effect, we would have to hold this morning's gap relatively well, preferably not losing more than about 0.75% from the open.  That would roughly coincide with yesterday's lows, so perhaps a slight undercut of those levels would be OK as long as we quickly recovered.  Should that occur and we move back above 710 that holds, I would look to become more aggressive on the long side into the afternoon and close.

Again, if we fail, I would be looking for a move towards 675ish, preferably coinciding with intraday panic readings in breadth and volume, but would be concentrating more on intermediate-term positions than short-term ones at that point.

All the best,