Tuesday, October 12, 2004

While reactions to earnings reports from YHOO and INTC will likely have an impact on otherwise "normal" market behavior, today's action in the NDX is interesting. There have been 30 times this year where the NDX opened for trading at least 0.5% below where it closed the previous day.  Only 10 times did it fully close the gap that same day by trading above the previous day's close.  While that seems like a positive development, three days later the index was actually LOWER 7 out of the 10 times, with an average return of nearly -1%.  90 days later, it was negative every time and the average return was nearly -5%.  So while today's performance does look promising, and the earnings reports after the bell will likely shape tomorrow's activity, I thought it notable that previous "gap-filling" instances were not necessarily a good sign going forward. 3:34:59 PM

	Friday, October 01, 2004

In the last comment, I pointed out the apathy regarding semiconductor shares.  With the nice rally since then, not much has changed in that respect (Rydex assets are still very low), so I would not be too quick to step in front of the up move there. Today's action, though, is moving many of our contrary indicators to true extremes as can be seen from the intraday charts.  Our shortest-term model on the Nasdaq is now down to 13%, one of the lowest readings seen so far this year.  Past instances approaching this level occurred on 4/2, 5/25 and 9/13, which mostly lead to consolidation or weakness in the week ahead.  We have positive seasonality for October that runs the first few days of the month, with weakness in the middle.  I suspect what we will see from this point is a struggling move higher over the coming days, with some weakness to follow probably by late next week. The much talked-about implied volatility measures (VIX, VXO and VXN) are pushing to new lows, which has mostly lead to weakness going forward.  This is another sign that it is probably not a good idea to chase this strength right now, but wait for some weakness which should come by next week. 11:45:20 AM

	Wednesday, September 22, 2004

As today's price action in the broader market certainly looked like it could have longer-term implications, I took a look at past instances when the S&P put in a similar performance.  Specifically, I looked at all times since 1962 when the S&P set a new 21-day high yesterday, but today closed lower than the prior 10 days' lows.  The results were interesting... Out of the 42-year period I studied, such a pattern has occurred only six times, the last one being on 1/4/00.  Right off the bat, it should be noted that six occurrences does not give us any kind of statistically valid sample.  Nevertheless, out of the 6 instances, the S&P recovered enough to exceed the prior high only once within the next 10 days, but it managed to do so 3 times within the next 21 days.  The average close the next day was neutral, with 50% of them closing positively and the average return equaling 0.1%.  After 10 days, the S&P was higher 2 times and the average return was 0.7%.  However, after 90 days the index was higher every time, with an average return of over 7%.  It did not make any difference if the volume was higher or not on a day like today. Overall, quite frankly the results are more positive than I thought they would be.  There was some weakness in the short-term, but it wasn't very pronounced.  It is notable to me that the market went on to new highs within 90 days every time, suggesting whatever weakness we do see going forward could be temporary.  Also notable I think is the fact that only once did the S&P exceed the prior highs within 10 days, telling us that most often it struggled or made minimal upside progress in the two weeks following such a pattern. One other thing I looked at was the amount of bullish sentiment displayed at the time, using the Investor's Intelligence survey.  Today's occurrence happened while we have the 2nd-highest amount of bullishness, with the occurrence on 1/9/86 being the only one that was higher.  After 10 days, the S&P on that occurrence was lower by nearly 1%, but recovered not long after that. In case you are interested, here are the other dates that popped up:  9/26/63, 1/9/86, 5/8/89, 11/15/91, 7/20/99, 1/4/00. 9:46:30 PM

	Tuesday, September 07, 2004

I think it's notable that there is a negative correlation between the performance in the S&P the day after Labor Day and its performance the rest of the week.  Since 1950, there is evidence that the larger the gain the day after the holiday, the weaker the return for the balance of the week.  When we restrict our look to only those times the S&P made a major move the day after the holiday (say a gain or loss of 1% or more), then the correlation becomes quite large at -0.41.  Even though that correlation is high, due to the small sample size it is not necessarily statistically significant (meaning it could happen by chance alone). Still, I think it's an interesting enough stat to mention in light of today's gains so far.  From the looks of history, a good day after Labor Day was a poor barometer of gains for the rest of the week. 1:27:25 PM

	Wednesday, August 25, 2004

Just a quick note to mention that the intraday put/call ratio I monitor (which includes all options, all exchanges) is at one of its lowest levels all year.  The only other times which match this level are 1/16, 3/2, 3/20 and 7/20, all of which came close to peaks in the Nasdaq.  When combined with the other data in the comments from last night, it is looking increasingly likely that we are getting close to a short-term pullback. These short-term bursts off potential intermediate-term lows are tough to fight, but we're seeing the types of extremes now that normally lead to at least a pause even in strong uptrends.  It'll be difficult to bust above resistance and stay there when we see these kinds of readings. 11:36:02 AM

In the widely-watched Investor's Intelligence survey, bullish respondents fell 4% to 39.6%.  This can now be considered historically low, and it is the fewest amount of bulls since the March 2003 lows.  Bearish responses picked up a bit, but are still well below what was seen in March.  So overall, the bull ratio (bulls / (bulls + bears)) is now about where it was in mid-April 2003.  Even though the S&P is 200 points higher now than it was then, we are seeing about the same sentiment situation in this survey. This is not a short-term timing device, as I've showed before, but by simply looking at its chart we can see that when the survey reaches extremes in sentiment, it is a consistent forecaster of market turning points.  If we could see one more week of increasingly bearish responses, it should serve to confirm that we are putting in a longer-term low. 8:50:41 AM

	Friday, August 20, 2004

Something for short-term traders to keep in mind is the market's tendency to show declines the day after an option expiration.  Today being expiration, it would apply to Monday's trading. The bias for the day after has been negative for at least the past 10 years, but it has become very prounounced lately.  Since the beginning of 2003, there have been 19 expirations.  The Monday following has been positive only 6 times (32% of the time), and the average return is a rather large minus 0.7%.  The average negative day has shown an average return more than twice as large as the average positive day, so there appears to be a definite negative skew there. One other note regarding this is that when expiration week showed a rally (as it has this week), then the day following expiry was higher only 25% of the time, and the average return was nearly -1%. There are obviously a number of things that could happen over the weekend to shift the traditional pattern, but I do think it's notable enough to mention that there has historically been quite a negative tone to Monday's trading.  Heading into the weekend, that may be something to factor into your risk profile. 11:31:13 AM

	Wednesday, August 18, 2004

The widely-watched Investor's Intelligence sentiment survey came out with its latest results showing 43.6% bulls (down from 48.4%) and 28.7% bears (up from 24.2%).  This is the biggest drop in bullishness in a while, and the survey is now about where it was at the lows in May in terms of overall opinion.  It is not showning what I would call an extreme in bearishness by any measure, but it's a big (and good) change from what we saw in June. The lowrisk.com survey also came out, and its results were essentially unchanged from last week.  The 4-week moving average of this data, as posted to the site, is still sitting in extreme territory, suggesting that population at least is about as bearish as at any other time over the past 7 years. The change in I.I. is another good sign.  Many are waiting for the survey to reach the extremes it did at the lows in 2002 and 2003, but we have to remember that these are newsletter writers.  They traditionally do not do as well attracting subscribers by giving bearish opinions (even if they turn out to be correct, oddly enough).  So as we saw in 2000 and 2001, it will take some time for these guys to become truly bearish - likely another protracted bear market.  I would like to see a less bullish attitude from them, but considering the market environment we have seen over the past year and a half, the fact that it's now around the May lows gives a small degree of comfort. 8:56:55 AM