Every Day, More Extremes

12/02/08 9:00 AM EST

Good Tuesday morning...we begin the day with heavy buying interest in the pre-market futures.  Asset classes almost across the board are reversing a part of yesterday's wide-range moves.

In a sign of just how widespread the damage was, there was only one stock in the entire S&P 500 that rose yesterday (Rohm and Haas), and another that was unchanged (AutoNation).  The rest closed lower, marking one of the worst-breadth days in the index's history.

Looking at the other times we saw less than 10 advancing stocks in the S&P, on September 29th the index rebounded nearly 60 points the following day then fell apart immediately thereafter.  On October 15th, it jumped about 40 points the next day, fell back the following day, rallied another 45 points, then fell apart.  On November 19th, it dropped 54 points the next day, then rallied for the next five before yesterday's fiasco.

Not too much of a pattern there other than at some point the losses from the bad day were mostly reversed in the short-term before leading to yet another pullback.  They all showed positive three-day returns, but after that, not so much.

Not surprisingly, those were also the days when we saw the heaviest volume on the sell side.  Yesterday, only 3% of the volume on the NYSE was concentrated in stocks that were up on the day.  Since the '87 crash, the S&P has shown a strong bias to rebound immediately from such a low one-day Up Volume Ratio.

Out of the 10 occurrences of comparably low one-day volume ratios since then, it bounced back the next day 9 times.  The one failure was the last one on November 19th, which we know was quickly made up over the next couple of sessions.

Prior to 1987, the tendency was to see more downside follow-through before any kind of a rebound.  Out of the 26 instances from 1950 up to the '87 crash, the S&P bounced back the following day only 11 times.  That highlights something we've talked about often - since 1987, and even more so over the past decade, stocks tend to rebound more quickly, and more severely, from one-day extremes than it did in the past.  Only in the very short-term though.

When the S&P violated its October lows in mid-November, I had to change my thinking.  Instead of being relatively comfortable with the idea that we were likely in the midst of an intermediate-term bear-market rally based on the overwhelming number of extremes generated in October, I'd have to switch to more of a trading approach that capitalized on short-term extremes, on both the long and short sides of the market.

On November 21st, the Dumb Money Confidence dropped to an extreme of only 13%, and we saw a wicked rally off of that, so it's apparent that traders are still willing to try buying into panic conditions.  Each subsequent rally has been shorter in duration than the last, but it's still good to know that when the next true oversold extreme gets registered, we should again look to buy into the morass.

On the flip side, we did see buying interest early last week in spite of short-term overbought conditions for the first time since July.  The next overbought reading last Friday brought us up to the 900 area that seemed like it should prove difficult to overcome, and yesterday was the result.  It's a good longer-term sign that we were able to see additional gains after the first overbought signal, but I'm still interested in looking to the short side on overbought readings.

I was thinking we might have a decent shot at a long setup if last week's overbought readings wore off with the S&P holding above 850ish, but with this kind of volatility that price level wasn't realistic.  It did provide something of a backstop intraday, but the last couple of hours of trading obviously violated that.  800 - 820 should prove to be another support zone, and we're bouncing off there today.

This volatility is exceptionally difficult to trade on a multi-day time frame, and I'm avoiding that until we hit another clear extreme one way or the other.  The highest-probability trades are either exceptionally long-term or exceptionally short-term (intraday).

Yesterday we had a clear trend day down, and it's possible (though less likely) we could see the opposite today.  As usual with these large gaps, I'll be watching the first hour of trading - if we make higher intraday highs after the first hour, then we should be able to make further gains as the day wears on, with a lower probability of a downside reversal.  If technical levels matter, then 840 and again 850 should prove to be resistance, especially if they're approached within the first half-hour.