Dec 30, 2002

sentimenTrader.com

Thursday, January 2, 2003 8:30 PM EST

I have referenced the NYSE TICK several times today, so for those of you who are unfamiliar, the TICK is simply the number of NYSE-listed companies that traded higher on the last trade minus those which traded lower. According to my quote vendor, the highest the TICK reached today was +1441. I show that 3,414 NYSE-listed symbols had shares traded today, so at one point during the day 1,441 more of those symbols were trading on an uptick than a downtick. This is extremely unusual, as I show only 6 days in the past three years with a high TICK of greater than +1300. Another interesting fact? Two of the top-10 TICK readings over the past three years occurred on 1/3/01 and 1/3/02. The TICK is commonly used as a sign of strength or weakness, but it can also be interpreted as a sign of panic, both up and down. Typically when we see TICK readings that are truly extreme, it is a sign of "get me in at any price" or "dump my shares no matter what". This is emotional trading, and very often precedes a reversal.

I mentioned the high level of the NYSE intraday cumulative TICK in the intraday comment and model change today. The indicator, which sums up the past 13 half-hourly TICK readings, became even more extreme at the close with a final reading of +7503. This has only been exceeded twice since the bear market began, with a +7545 reading on 10/4/01 and a +7986 reading on 7/29/02. The odd thing about those two readings is that they came within 10 days of a major panic low, and our current reading basically came out of nowhere. The two prior readings were also followed by 3-5 days of retracement which erased the prior days' gains, before another major leg higher.

The only other time since the beginning of the bear market in March 2000 we have seen the extreme conditions of a +7000 cumulative TICK and at least 70% in the price oscillator was on May 17th, 2001. After this instance, there was a small three-day rally before a more severe mini-correction which ultimately lead to the waterfall decline into the September low. If we relax our conditions a bit and only look for instances with a cumulative TICK above +6500 and a price oscillator reading of 65% or greater, then we get five occurrences: 5/17/01, 9/28/01, 5/14/02, 7/5/02 and 7/29/02. Five days later, the S&P was lower every time except 9/28/01, in the rally which followed the low after the market re-opened. The average five-day return in the above instances was -2.3%. If we take out the 9/28/01 occurrence, then the average return drops to -3.4% If we go out to 10 days, then all occurrences were again negative except 9/28/01 and 7/29/02 (which showed a very slight positive return). The average return after 10 days was -3.1% and if we again take out the September 2001 reading, the average return drops to -4.8%. We only have five occurrences here, so it's not like we can derive any statistically significant conclusions, but what seems apparent is that these occurrences of extremely high cumulative TICKS and overbought price oscillators tend to occur during explosive one-day moves. In fact, each of the five occurrences were recorded during one-day moves in the S&P of at least 20 points. It was also typical to see a modest one- to two-day upside continuation before the downtrend resumed in force.

Since the STEM.MR should also be giving a low reading during these explosive moves, I checked the average reading that model gave on the five days mentioned above. The average was just under 9 (the lowest was 3 and the highest was 13), so our current reading of 13 compares favorably but is a bit high to be "average".

With over a 2% decline in the S&P since the last reporting period, one would have expected the Investor's Intelligence survey to show a drop in the stubborn bullishness is has shown recently. Not so. Once again, the bullish ratio remained constant, and in fact even increased a tad. This steadfast optimism, whether it be for the positive seasonality or what, certainly has to be considered bearish in the intermediate-term. I went over some statistics of this phenomenon last week, and the results were as you would expect (see the December 26th Daily Commentary in the archives for details).

For only the second time in seven months, the NDX Down Pressure registered a reading of 0% today. This indicator measures the volume and magnitude of all point moves in the 100 component stocks of the Nasdaq 100. Amazingly, there was a total of 97.71 points gained and a grand total of 0.12 points lost. $0.12 lost out of 100 stocks is rather extreme, I would say. The only other day with such a lopsided reading was July 5th, which is not a good precedent if you're looking for continuation to the upside. In the S&P 500, the Down Pressure today was 1%, with 478.66 points gained and only 2.10 points lost. This is the lowest number of total points lost, and the lowest Down Pressure reading, in the past five months.

If we can compare our current situation to those over the past two years, and this is arguable, then today's extreme measures portend a small one- to two-day continuation at best before reality sets in and we retrace today's gains. That has happened without fail in comparable circumstances, so I believe it's a high-odds situation to play out now. What happens after that is the question. Is today's explosion a sign of built-up demand that was just waiting for an excuse to buy (in the form of a better-than-expected economic number)? Or was it something structural having to do with the time of year (perhaps mutual funds putting whatever cash they have left to work)? Since the first two trading days of the year are historically positive, and two of the top TICK readings in the past three years occurred on 1/3/01 and 1/3/02, I am lead to believe that it is a structural phenomenon and not a sign that we have miraculously turned the corner. If we simply look at what happened in the past when we recorded the readings we're seeing now, the outlook is not positive. Therefore, I believe betting on lower prices over the coming days now presents the best risk/reward scenario. If we get a gap up tomorrow (particularly a large gap up), I would look aggressively to sell short against it. If we don't get a gap up, I would view any rally that occurred to be shortable for a short-term (1-5 days) move at least.

Disclosure: long QQQ calls, long QQQ puts

This disclosure is not intended as trading advice in any form. It is meant as a note to subscribers that I have a position directly affected by my market outlook. Although I take great pains to remain objective in my commentaries, I believe it is only fair that readers should know that I have taken positions in accordance with my market outlook.

Monday, December 30, 2002 9:51 PM EST

You will recall that I recently discussed the Seasonality Index created by Jay Kaeppel, which is a number I post to the site each day. For any given day, this index assigns one point to four different positive seasonality influences:

The two days immediately prior to an exchange holiday (New Year's Day, Presidents' Day, East, Memorial Day, July 4th, Labor Day, Thanksgiving and Christmas).
The last trading day of the month and the first four trading days of the next month.
November 1st through the 3rd trading day in May.
The most favorable 14 months of the 48-month Presidential election cycle. This begins on October 1st two years prior to each Presidential election, and ends on December 31st of the following year.

These four biases were chosen due to their consistency in out-performing other days, and have proven themselves over time. Tomorrow, December 31st, will receive the highest possible score of "4", meaning that each positive bias above will be present. There have been 26 other days since 1950 that have also received a score of 4. 19 of the 26 days ended positively, for a success rate of 73%. Those days have averaged a daily return of 0.27%, with a maximum of 2.08% and a minimum of -0.49%. The average gain was 0.47% and average loss was -0.26%. If you had invested $10,000 at the beginning of the period, by buying the S&P at the close on the day before the day with a score of 4 and selling at the next close, that $10,000 would have grown to $10,731 by 2000. This is a return of 7.3%, with your money only being invested essentially for a total of 26 days. The bottom line is that it is not wise to count on being short intraday tomorrow, and in fact I would instead look aggressively for long candidates. This is with the understanding that this particular bias will last for tomorrow only. Thursday and Friday will have Index readings of 3, which is still quite positive, though not to the degree of tomorrow.

Today's NYSE volume was about 25% below the 200-day average, and if history is a guide, tomorrow will be slow as well. Over the past 20 years, the last trading day in December has averaged volume 85% of average, with 18 of those years showing below-average volume.

The few short-term positives we had coming into today have dissipated, as the STEM.MR model, NYSE cumulative TICKS and price oscillators have all alleviated their oversold conditions. About the only positive from a sentiment-type perspective I can find for the near-term is the Seasonality Index as discussed above. Friday will also have a particularly positive seasonal bias, so that's something to watch as well. Other than that, I don't see any particular edge from our models or indicators. With a lack of clear signals from those and the light holiday volume, this is a period to be trading lightly if you trade at all.

- Jason Goepfert

Disclosure: long QQQ calls, long QQQ puts