sentimenTrader.com

All rights reserved.

Thursday, October 17, 2002  8:34 PM EST

In yesterday's commentary I mentioned the statistics regarding recent gaps up and down in the Nasdaq 100, and how it affected the close and next morning.  Taking it a bit further, I looked at all gaps in the Nasdaq over the past 5 years, and how long it took them to ultimately (if ever) get filled.  For those of you not familiar, by "filled" I mean price goes back and closes the gap at some point.  In these five years, there have been 100 gaps.  For the purposes of this study, I considered a gap to be an open more than 2% away from the previous close.  It has been easier for the NDX to gap in the past year or so because the average is 1/5 what it used to be.  At the peak in 2000, a 2% gap would have been 80 points, so the majority of our sample gaps comes in the past two years.  Here is the breakdown:

We can see that there were about an equal number of gaps up and down, and remarkably similar filling statistics.  A rather substantial number of gaps get filled sometime during the same day (40% of the time for gaps up and 36% of the gaps down).  A great majority of gaps are filled within 30 days.  Another interesting tidbit is that when there are two gaps up within 5 days of each other, the second gap gets filled the same day just over 60% of the time, instead of the 40% on a "normal" gap.  However, on gaps down, the second gap within 5 days gets filled the same day only 32% of the time.  This phenomenon is likely due to the downward bias over the past couple of years, which contained a majority of our gap sample size.  When we have two gaps up within a short period of time, that has typically coincided with overbought conditions.  And as we well know, overbought in the context of a downtrend does not last long. 

So how can we use this information?  Well, let's take gaps down, for instance.  If we gap down and price begins to fill the gap immediately, it may be a good bet to sell short the attempted gap-filling rally, since there is a 64% probability of the gap not being filled that day.  However, within a relatively short period of time, there is a decent bet that the gap WILL get filled, so you may want to look for upside reversals in the short-term (in conjunction with other sentiment measures and your own methodology, of course).  I think it's safe to say from the results here that there is a natural inclination for gaps to ultimately get filled, and usually within 30 days.

I said in the intraday note that there was no clear edge in either direction from a sentiment perspective, and indeed equities continued to chop around for the rest of the afternoon.  I also mentioned that since we gapped up on above-average volume, there was a good chance we would close higher but gap down tomorrow.  With after-close earnings reports, that gap down does not look likely, and in fact we will most likely gap up again.  This will be the third gap up in a relatively short period of time, but the sample size of the study mentioned above is not large enough to come to any conclusions about this type of situation.

Regardless of the gaps, we are becoming overbought once again in the short-term.  Unlike the previous two times we were short-term overbought over the past two weeks, we will likely get a gap up opening (if the current futures hold).  When we combine the gap statistics above with the fact that we are somewhat overbought, there is a very good chance that the gap will get at least partially filled tomorrow.  I would suggest that the short-term traders among you be fairly aggressive shorting a large gap up tomorrow, if it fits your methodology and risk tolerance.

The equity put/call ratio has had quite a wild week.  After doubling yesterday from the previous day (which has never happened in the past 7 years), the p/c ratio was halved again today.  The cause was a huge surge in call volume, greater even than Tuesday's number.  At the same time, put volume shrunk to the lowest amount in the past few days.  I'm not a big fan of trying to guess the "why" behind our indicators, but considering the elevated nature of the 21-day put/call ratio, I think it's safe to say that the call volume was not influenced to a great degree by closing transactions but rather new speculation on additional upside.  That's not a positive, especially considering the OEX put/cal ratio continues to be rather high (meaning the more savvy index options traders are not betting on the upside to a great degree).

Yesterday the 5-day moving average of the TRIN reached the lowest level in a decade (!), and today the 10-day average could be considered overbought, both on the NYSE and Nasdaq.  Although an overbought 5-day average has not had particularly bearish overtones, even during this bear market, overbought 10-day readings HAVE been bearish.  As I've said many times over the past few months, overbought TRIN readings in the context of a daily downtrend have been very effective in the past couple of years, and it is beginning to urge caution at this point.

Short-term, we are overbought and much more upside will make us VERY overbought, so the risk/reward is skewed to the short side.  I would view strength at this point with suspicion.  Longer-term, as I have stressed since last Wednesday, the intermediate-term picture is positive, and I would view significant pullbacks to be buying opportunities.

Wednesday, October 16, 2002  9:01 PM EST

Before I comment on today's action, I wanted to present the results of a research request from a subscriber.  He asked me to look at gaps up and down, how they influence the close, and how the close influences the next morning.  On top of that, he also asked what impact volume had on the stats.  I looked at the Nasdaq 100 over the past six months, since that market reflects gap opens more accurately than the other broad-based indexes and because that time frame encompasses pretty much everything we've experienced over the past two years.  The charts below lay out the average return and probability of the market being up for each of four scenarios:

1.  Gap down on below average volume

2.  Gap down on above average volume

3.  Gap up on below average volume

4.  Gap up on above average volume

The average return is the close-to-close return.  So, if we gap up on Tuesday, the "AVG CLOSE" would be how Tuesday closed compared to Monday.  "NEXT OPEN" would be how Wednesday opened.  I considered a gap to be an open 1% away from the pervious close, and average volume to be the past one month's average. 

I realize this is all probably confusing, so let's just look at the charts and go over an example:

Let's go over the top left chart first - a gap down open on below average volume.  When that happens, the day ends positive 40% of the time, with an average return of (1.1%), while the next morning is essentially a draw.  The top right chart shows that when we have a gap down on ABOVE average volume, the day closes positive only 29% of the time (meaning it has a probability of closing lower of 71%) with an average return close to (1.5%).  However, it also has a good chance of gapping up the next morning, to the tune of a 0.65% average.

When we gap UP on below average volume, the day ends positively only 60% of the time, with an average return of 0.80%.  Considering that we initially opened up over 1%, this means that the market actually sold off a bit during the day (on an average day).  We then have a tendency to gap down the next morning, with an average gap of (0.40%). 

Now for the most remarkable statistic.  When we gap up on ABOVE average volume, we end higher an astounding 90% of the time, with an average return of nearly 4%.  Part of this high average return is due to a very strong +10% day, but even without that data point the average gain is 2.86%.  When this happens, we gap down the next morning 80% of the time, with an average gap of nearly minus 1%.

This was not just an exercise to keep me busy after market hours.  There is some real, practical trading knowledge to be gleaned from these statistics.  For short-term traders, when we have a strong gap up, it may be best to take profits near the close, since there is a very strong chance that we will gap significantly lower the next morning.  Today we gapped lower on above average volume and closed poorly, so there's a very real chance we could have a gap up open tomorrow.  These statistics do not take any other trading information into account, so I am by no means suggesting that traders place or close trades based on this information alone.  But if you are considering the risk/reward of holding your positions overnight, you may want to take the above charts into consideration.

Today's persistent negative tone worked off the overbought condition in virtually all of our short-term indicators.  The cumulative TICK indicator on the NYSE is working toward oversold, and if we get there tomorrow it will present an interesting situation.  If we are in the midst of a new 30-minute uptrend, then we will be oversold within an uptrend, which as you know is when these indicators are at their most effective.

I mentioned yesterday that the "hot money" from the Rydex timers were pouring into the bullish funds and out of the bearish funds at a pace that almost invariably leads to short-term weakness.  That trend continued when yesterday's numbers were released, as the RSI spread reached an unsustainable +80.  Today's weakness will undoubtedly relieve the outflow from the bearish funds, so we should see the RSI spread drop over the coming days.  But I suggest that going forward, the short-term traders among you pay close attention when the RSI spread reaches an extreme (this chart is posted to the site daily).

I stated yesterday that we were seeing some aggressive call buying by equity options traders, which is never a good sign.  Call volume dropped off dramatically today, as CBOE equity call volume was over 775,000 contracts yesterday but only 431,000 today.  Put volume picked up a bit, but that difference in call volume was enough to move the put/call ratio from 0.54 yesterday to 1.15 today.  I have not found one-day readings to be particularly meaningful, but if it keeps up (and we see the OEX traders begin to commit to some call positions), then I would consider it constructive.

The Investor's Intelligence survey fell to 28.4% bulls and 43.2% bears.  You've probably already heard about this, or likely will over the coming days, since this survey is so widely watched and the figures are so dramatic.  This is the lowest bullish ratio (bulls / (bulls + bears)) since a string of low readings in 1994, so it will probably be a popular topic among casual sentiment observers.  I pointed out last week the positive intermediate-term effects of wide bull - bear spreads in this survey, so this weeks reading only reinforces the likelihood that we will have more sustained upside over the coming weeks.

Today's decline was on relatively light volume, so that's one for the bulls.  If we can get some more downside over the next day or two, it should set up some nice buying opportunities for both short-term and intermediate-term traders.  Follow whatever methodology it is that you follow, but if we get some positive signals out of our shorter-term indicators, you should be aggressive in exploiting whatever long signals your methodology generates.

Tuesday, October 15, 2002  8:50 PM EST

During this bear market, there have been 21 instances of the S&P being up 4 days in a row.  Only four of them went on to become 5 days in a row.  These occurrences have tended to precede market underperformance as detailed below:

We can see that out to three days following four up days, the S&P has tended to trail a random day in both average return and probability of being positive, and it has also lead to some of the largest short-term declines seen during this downtrend.

For the Nasdaq 100, there have been 27 such instances of it being up four days in a row, 8 of which went on to be 5 days in a row and 2 which were 6 in a row.

For the Nasdaq as well, being up four days in a row has tended to precede some short-term weakness.

The pressure on the downside that we saw until late last week obviously has been reversed, and to a great degree.  I keep an indicator that compares points gained on the S&P (and NDX) versus points lost, as well as the volume pouring into each set of issues.  From that, I compute a "Down Pressure Index".  For example, today there was a cumulative 642 points gained in the securities that make up the S&P 500, and only 23 cumulative points lost.  When coupled with volume, it creates a Down Pressure Index of 6%.  The graph below is a 3-day moving average of this index since the beginning of July:

We can see that after reaching a peak of 85% (extremely oversold) last week, we've now cycled down to 16%, suggesting that there has been a lot of points gained with a significant amount of volume in the past three days.  If we look at a four-day average, today's reading would be the lowest in this lookback period.

From this same information, I create a "Modified TRIN" on the S&P 500 and Nasdaq 100.  The "modified" nomenclature is used because instead of looking at simply up issues versus down issues, as the regular TRIN does, I look at the number of points gained or lost by those advancers and decliners.  I believe this is a more accurate way to view the actual breadth of the day.  In any event, today's Modified TRIN on the S&P 500 was 0.24, which is a three-month low.  On the NDX, it was 0.19, which has been exceeded a few times in the past quarter, with all but one leading to lower prices short-term.

The point of these indicators is to determine if we have seen "too much, too fast".  That's a slippery concept to grasp, and one certainly open to interpretation.  But by looking at actual points gained or lost, I think we can get a better feel for it.  If recent history is any guide, then what we have seen over the past four days is certainly both too much and too fast, and we are due for a breather.

I mentioned yesterday that I would be very interested in the Rydex asset flows last night.  They didn't disappoint, as the flood out of the short funds continued.  This has served to push the three-day RSI spread between the weighted bullish and bearish funds to +70.  When this level has been reached over the past two years, it has lead to short-term weakness every time but once (which preceded a small rally which ultimately failed).  Of course, there is no intrinsic property of this indicator which would suggest the market is overbought (unlike the indicators discussed above), but the traders who use these funds - particularly the leveraged funds - tend to be "hot money" which jump on every trend as soon as it seems to pick up steam.  Usually, by the time they have piled in to a great degree (reflected by our RSI spread), the trend is about to either consolidate or reverse.  Since they are now so convinced that there is more upside in here, it makes it less likely to be so.

We also saw some pretty optimistic trading by the equity options traders, as the equity put/call ratio approached the lower 1.5 standard deviation band (you can see the chart each day on the site).  The OEX ratio continues its elevated ways, and the spread between the two is remaining neutral to slightly bearish.  In the S&P 500 options, our put/call bid/ask bias ratio is relatively bearish at 0.51 (reflecting some rather heavy selling of calls and buying of puts), but the volume was extremely heavy and this is undoubtedly being influenced by option expiration, so I'd rather not try to read too much into this indicator at this point.

At the close, the STEM model dropped into strong negative territory, as it is now significantly below its lower standard deviation band.  This is a longer-term model than the STEM.MR model, and suggests that weakness is to be expected over the next few days.  Coupled with a low STEM.MR model, the chances are high that weakness is imminent.

As I write this, it is apparent that the weakness has already begun, due to disappointing earnings reports.  I don't care if you're a subscriber or free trialer - if there's one thing I wish you to gleam from this service, it's that when sentiment reaches one extreme or the other, when an important news event is forthcoming, it takes a better-than-usual catalyst to continue the trend.  For example, last Thursday it would have taken news that was truly horrible in order for the market to continue down, and after today it would have taken some unexpectedly positive news to keep the uptrend going.  When the release is disappointing or even ho-hum, you can expect a quick reversal.  We should see that play out again tomorrow.

Because of the reasons outlined above, I was prepared to suggest that the risk/reward for short-term traders would lie with shorting any additional strength tomorrow.  With the apparent gap down that we will likely get, the situation becomes murkier, but I believe that if we do gap down and begin to fill that gap, then look for subsequent failures.  With our overbought situation, failed rallies should prove to give high-odds short setups.  For intermediate-term traders, nothing has really changed.  The market is acting as it should if we have put in an intermediate-term low, so trailing stops should still be used, but you may not want them especially tight.  It depends on how the market reacts to whatever downside pressure we get, but I don't see any reason to believe that there is not more upside left to the rally that began last week.

Monday, October 14, 2002  7:28 PM EST

Tonight, you will notice that the STEM.MR model has been revised, completing the transition of the four models to better reflect their original intentions.  Like the others, and a growing number of indicators, I am now using standard deviation bands around a longer-term moving average to determine when the model becomes significantly bullish or bearish.

Similar to the cumulative TICK indicators, the STEM.MR model entered today quite overbought, and considering that we were still in a downtrend, odds were high that we would see some weakness.  We got the weakness, but unfortunately it was in the form of a gap down and the markets spent the rest of the day chopping higher.  The day's action was enough to relieve the tension of the overbought condition for most of our indicators, and even though we closed up on the day, we're less overbought than we were yesterday.

One possible negative for the near-term comes from the S&P 500 options.  Our put/call bid/ask bias ratio closed today at a very low 0.22, one of the lowest readings we've recorded.  However, the ratio is being skewed by what may be one institution creating a large position that is almost certainly a hedge.  It looks like one entity sold over 4000 October 825 calls and bought 4000 October 825 puts.  They also sold about 4000 December 825 calls and bought about 4000 December 825 puts.  There's no real way for me to tell if it was indeed one institution or what the strategy was, but I'm pretty certain it was not an all-out bearish bet against the market.  There was nothing of note in other option activity.

The Rydex asset levels are not released until late in the evening, so that information could have something of an impact on the near-term outlook if it makes the RSI indicator I talked about this weekend even more extreme.  These traders have recently been throwing money around in a rather volatile fashion, so tonight's report should be telling.

With the bond market closed, volume was very light and it's hard to make much of such a session.  The one thing it did was relieve a good deal of our overbought condition, so at this point I would have to say that the chances of another push up are significantly greater than they were entering the day.  With a slew of earnings releases about to hit the tape, the only bias I have at this point is a very slight negative one.  If we get news that isn't quite as positive as it could be, the ingredients are in place that could cause something of a selloff.  But with the impressive strength the market has recently shown and the positive intermediate-term outlook, I don't believe a complete breakdown is imminent, and in fact we could easily see another leg up.  In sum, I can just as easily see us go up as down, which is no edge at all, so I'm on the fence for now.

- Jason Goepfert