May 13, 2002

sentimenTrader.com

Friday, May 17, 2002 3:53 PM CST

Please check back Sunday for the regular weekly commentary. All daily charts (except the CBOE ratios) have been updated as of today's close. Have a great weekend.

Thursday, May 16, 2002 3:55 PM CST

Unless you're selling options and taking in premium, it doesn't get much more boring than what we saw today. Fortunately, there is reason to believe that volatility will pick up within a day or two. Today we had an inside day on both the S&P and NDX, and it was also the narrowest range of the past 7 days in the S&P and the narrowest range since March 27th for the NDX. The last few times we have seen such a narrow range in the NDX have led to significant movements soon after. In the S&P, today's range was only 54% of the 10-day average range. Over the past four years, there have only been 77 days where the range was 55% (or less) of the 10-day average. The average increase in range the next day was just over 100% (meaning the day after such a narrow range day typically had a range twice the narrow day). A whopping 92% of the time the range increased the next day, with an average increase of over 10 points. That bodes well for increased volatility.

But which direction will that volatility take us? After the close, DELL announced some good new, which sparked the futures to make a bit of a run. Whether that can continue into tomorrow is the big question. We have somewhat conflicting readings in our shortest-term indicators, as the put/call ratio stayed relatively high all day (which is bullish) but the VIX dropped to 9% below its 10-day moving average (which is bearish). It's somewhat unusual to see these diverge so much, so at times like this I find the TOPM model useful as something of a tiebreaker. That index today is on the high end, suggesting calls are being bid heavily (in relation to puts). This would lead me to place more weight on the dropping VIX as a sign of complacency as opposed to high put/call ratios as a sign of fear.

Our STEM.MR model didn't move enough to give any indications, as the opposing indicators cancelled each other out for the most part. The STEM model, however, finally dipped below .30 on the slow line. This is in SELL territory and I will be watching very closely for a reversal. A signal is not generated until the signal line makes a reversal so that we minimize the chances for selling into a continually rising market.

This leads me to believe that there may be another final push to this rally, but it will ultimately fail and take us back down to challenge the recent lows.

Wednesday, May 15, 2002 3:45 PM CST

A pretty interesting day in the market, but boring sentiment-wise. This morning, we got lucky by covering the SELL signal at almost the low tick of the day before the rally. That rally was impressive, especially in the technology indices, but fizzled in the early afternoon for whatever reason (I've seen about six rumors now that supposedly accounted for it). That failure near resistance didn't do much to our sentiment indicators, as they stayed fairly steady, although low, throughout the day. Maybe that's an ominous sign - when a failure right at resistance doesn't do anything to generate fear.

The STEM.MR model never really went anywhere after we covered the SELL signal when it went into neutral territory. I mentioned last Wednesday that when short-term sentiment is as negative as it was then, it doesn't take much of a catalyst to spur buying (or short covering). We had mediocre economic news, but it was spun in a very positive light. Now entering today, with extremely overbought short-term sentiment, another batch of mediocre economic numbers was spun the other way. It pays to be aware of this short-term sentiment when important numbers are about to be released.

The STEM model dropped quite a bit, like I mentioned yesterday it might. We have another 10 half-hourly readings that are high and when those drop off it could have a dramatic effect on the signal line in this model (depending of course on what readings replace them tomorrow).

The VIX and VXN displayed similar action to yesterday and our breadth oscillators went nowhere. However, in the next few days we will be dropping some relatively large negative readings from the advance/decline ratio (10-day moving average), and if we get any more of a rally, that oscillator could become quite overbought by Friday.

An early peak at the sentiment surveys shows a slight increase in bearish posture from the II crowd but an increase in bullishness in the others. Still a mixed message there, but overall rather constructive.

There's not much to add today that wasn't said yesterday, as today really did nothing to alter the picture. Tomorrow will provide some more clues.

Tuesday, May 14, 2002 9:00 PM CST

The continuation of the rally today triggered the sell signal in the STEM.MR model as discussed yesterday. As the day wore on, short-term sentiment became extreme enough to get the model to a reading of .60, which is essentially zero. I went back over the past two years and checked the past occurrences when the model was .60 or less. Out of approximately 5800 1/2 periods, there were 35 where the model was .60 or less. The following table shows the percent returns for the S&P 500 after these occurrences. The returns are assuming that one goes short when the model reaches this level. So the % profitable row would mean that, for example, 69% of the time the market was down 1 1/2 days later and it was down 77% of the time 2.5 days later. The max return would be the most profitable short position while the min would be the least profitable short. I don't lay it out this way to make it seem as though this would be a viable trading system, although from the looks of it it may (marginally at least). I do this so that we can objectively look at the odds of what may happen by looking at the recent past.

                  3hrs 1day 1.5day 2day 2.5day 3day 5day 7day 10day 15day
random        0.0 -0.1    -0.1    -0.1    -0.2     -0.2   -0.3    -0.4    -0.6     -0.9
STEM.MR   0.0 -0.2    -0.5    -0.8    -1.0     -1.1   -1.5    -1.1    -0.8     -1.2

% profitable 49    51      69      66      77        71     60      49      51       60
max             -1.0 -2.3   -2.4    -4.4    -5.2     -7.0   -7.5    -8.9   -10.9    -9.2
min             0.9   1.5     1.4     1.6     2.9       3.6    3.2     4.0     6.3      4.5
avg pos       -0.4 -0.9    -1.0    -1.7    -1.6     -1.9   -3.1    -3.9   -3.0     -3.2
avg neg        0.4   0.6     0.7     0.6     1.2       1.1    1.0     1.6     1.7      2.0

So what conclusions can we draw from this? Well, the first conclusion is that this model level depicts overbought markets fairly well. There weren't too many large rallies after this level was reached. You can see from the avg pos vs. avg neg rows that the average market decline was greater than the average market gain (avg pos would be a positive return for a short position - meaning a market decline) across all time periods. The largest gain was also greater than the largest loss, meaning that the market dropped by a greater degree than it rallied, also across all time periods. So probability says that the market should be lower several days from now than it is today, but if the rally does continue, it probably wouldn't be by much.

Optimism returned in full force once we broke the recent highs. Options traders certainly determined that the rally will continue, as the open interest - put/call spread widened to .32 from under .10. That's not necessarily high, but it sure is no longer oversold on a short-term basis. Obviously, the VIX and VXN also dropped pretty significantly. The reason the daily commentary is late today is that I was researching the past occurrences when the VIX dropped below the 10% band then reversed and closed back inside, like it did today. Unfortunately, the results were inconclusive. I could say that it usually portended an increase in the VIX and a market decline, but that conclusion, while true, is very marginal. I was also interested in looking at past occurrences of the VXN dropping as much as it has the past few days, but since this indicator only goes back one year, there was not enough data for me to be comfortable with any conclusions.

The STEM model may be close to giving a SELL signal if we get another day of upside, as the signal line will be dropping a day of very high readings. This is quite unusual, since at the end of last week it would have been quite easy to get a BUY indication if we had gotten a day or two of downside. This slower-moving model typically doesn't move that much in a few days, but the extreme readings over the past two days have made quite an impression.

The breadth ratios are quickly cycling back up to overbought from oversold just a few days ago as well.

Although the gains of the past two days have been impressive, there are two problems. One is the base from which they sprang was not as solid sentiment-wise as it has been at past intermediate-term bottoms. Also, the rally has not shown the explosive breadth and volume that typically signals the end of a severe decline. That does not mean this rally cannot hold or continue, but the odds are against it being THE bottom everyone seems to be looking for.

Monday, May 13, 2002 3:58 PM CST

We had another one of those tinder-box days today like I mentioned in the most recent weekend commentary. Our short-term model, STEM.MR, was on the cusp of 90 going into today's session, and when you combine an oversold short-term market with an oversold longer-term market, the results can be explosive.

Even though we had a weak BUY signal on that model going into today, a rally isn't particularly what I wanted to see. Like most others, I wanted a day or two of pressured selling - the kind where margin clerks are selling out accounts instead of waiting for money to be wired in or customers to do it themselves. Don't take that as I want anyone to suffer - I certainly don't - but that's the kind of action that washes out weaker holders and allows those looking longer-term to step in at better prices. That's the kind of action that gives us more sustainable rallies.

Now, we're back to a situation similar to last Wednesday, although not as strong. NYSE and Nasdaq volume today was the lowest in two weeks, and the breadth ratio (advances / (advances + declines)) was less than it was on Wednesday (although not by much). If this is THE bottom everyone is pining to see, we would typically see volume and breadth explosions on a scale much greater than we're seeing here.

We're getting somewhat mixed signals from the options crowd, as the put/call ratios and volatility indices (VIX and VXN) are showing increased confidence of a rally, but the TOPM index (see the April 28, 2002 weekly commentary in the archives for a discussion of this index) is showing that puts were being bid surprisingly heavily today. Normally, all of these ratios point in the same direction, but we're getting a divergence here. I wouldn't read too much into any of them at this point.

The STEM.MR model is right back down to SELL territory, and if we get a continuation of the rally past the first 1/2 hour tomorrow morning, we will probably get at least a weak SELL indication from that model. The other models continue to be frustrated by rallies when they are at the cusp of giving BUY signals. We were very close to BUY's in STEM, AIM and the about-to-be-released Composite model, but today sets us back a couple of steps there. We really need to see a couple of days of heavy selling and pessimism to get them over the hump and into BUY territory.

I talked a bit about volatility in the May 5, 2002 weekly commentary and that we had seen five of the past six days exceed a 1.5 standard deviation volatility band. As of today, 9 of the past 12 days have seen that level exceeded and three of those days have gone beyond the 2.0 standard deviation band. In a "normal" market, it would be unusual to see the 2.0 band exceeded once, and certainly not twice. Three times is extremely unusual. If increased volatility does precede bottoms, this is another good sign.

Like I said this weekend, sentiment is at a place where a decent rally could certainly occur. However, at investable market bottoms (and shortable tops) we typically see a confluence of signals from our models, and we just don't have that at this point. I would prefer to see this rally fail and challenge (or slightly exceed) the recent lows to get our signals, but we'll just have to follow the market and let it tell us what to do. Price always comes before indicators.