There are two gaps on the S&P index. The first is at 1089 (green) and the second is at 909 (purple). Historically at this point in the market cycle the market will drop in the area of the blue circle. That is, it generally retraces the amount of the previous major wave which ends up at about the mid-point of the prior market crash (2008). Still, that's a pretty big circle so we're going to have to break out some price projection techniques to try and narrow things down. The biggest question for me is whether the purple gap is a possible target or a bogey.
Historical Context
Let's review the historical context. First 1977:
1977 ended up being more of a correction than a crash, and so the actual corrective pattern is not an exact match. Regardless, the green and purple arrows match up almost identically. We can see that the actual bottom for 1977 ended up right at the first "higher low" from the previous rally. This would correspond to about 1050 on our own market and would neatly fill the infamous SPY 106 gap:
Let's now take a look at 1949 which is the other historical prototype for this market:
Here indeed we had a stupendous crash which nevertheless ended up in about the same location relative to the gap (technically not a gap but a "thin area"). In 1949 however there was no obvious purple gap as the market was rather zig zaggy with plenty of coverage.
Crash Patterns
The 1977 bear market is not a good analog for our current crash. At this point I'd like to take a closer look at 1949 as well as a few other crashes that bear a similar imprint to our current crash. Let's zoom in on 1949 and compare it to today:
To me, the trademark feature of the 1949 crash is the volatility at the bottom. As you can see from this picture, the volatility at the bottom is somewhat symmetrical with the "stairway" pattern that led to the crash. Our own market had a single step and as such I would expect at least one false rally (aka short term trading opportunity). The wild cards in our own market are the two blue "X" marks. The first X is from a higher point and thus we might expect it to produce another lower point. Indeed I think that's a possibility but it's important to note that these two X's were not part of the stairway pattern. As such I believe they may play out into a larger correction. But first let's take a look at two other crashes that bear a similarity to our current crash. First 1962:
Here I'm comparing the 1962 crash on a daily chart with the S&P futures on a 30 minute chart. The S&P futures dipped much further in the after hours of August 8th, all the way to 107.7. Often the S&P futures provides a much clearer wave pattern than the index or ETF. The similarities between these charts are striking. They include a single step into the crash, a curved crash and a retracement back to the 50% mark of the crash. Even the wave pattern leading into the retracement looks similar.
There are some subtle differences though. If we look withing the crash curve we can see that 1962 had a well defined 3rd and 5th wave. In our current crash, the 5th wave appears the longest. This is a warning sign that the previous wave count is not complete.
Secondly, the wave patterns in the retracement are opposite of each other in these crashes. 1962 started motive while today's retracement started as a zig zag. Next 1962 had a 3 wave while today's crash had a curved motive wave. Finally, on the way down, 1962 had a clear step pattern while today's crash was a motive slide up to today.
As such, I believe the 1962 pattern is analogous in that they both "retest" the bottom, however 1962's retest was part of a larger motive wave upward and thus did not actually make bottom while today's retest (I believe) is finishing out the prior crash wave and thus should make a new bottom. Still, the analogy demonstrates that we need to be on the lookout for the possibility that we will not get a new bottom, and more probably I think, that if we do get a new bottom it will not be much lower than 1077! Again the infamous SPY 106 gap looks like a good end point.
Finally let's look at 2002 versus today:
Amazing right? Once again we see nearly the exact same pattern. It is worth noting however that this 2002 "crash" was actually the end phase of the dot.com crash. So 1962 was on the way up; 2002 was on the way down; and today is essentially sideways. The lesson here is that the same supply/demand patterns show up regardless of the major trend.
The big difference between 1962 and 2002 however is that the "retest" actually made a new bottom as I am expecting for our own market. Note the blue lines next to the curve and how 2002, like today, finished the curve with the longest drop. This is the indication that a new bottom will occur after the retracement. I believe the pink circles mark the analogous points although our current retracement does point towards a deeper "retest" than occurred in 2002. We'll explore this further in the article.
Worth noting is what happened in 2002 after the bottom was finally reached. This I believe is a likely scenario for our own market. What looks like a secondary retracement is actually the motive wave from the 1962 market, along with the retest that doesn't quite touch bottom. This creates the symmetry with the aforementioned "X" marks that occurred in our own market back in May.
Oh look, the market is plummeting as I write this...
Fibonacci
After first analyzing historical context and wave morphology I usually turn to Fibonacci levels to attempt to identify likely points of support:
Not surprisingly, we had a major Fibonacci confluence zone at 1098 which is where the market bottomed out (although the futures did go further) and just a hair lower than the first major gap. More difficult is predicting exactly where the next wave will stop as we have levels at 1060, 1040 and 1016 which correspond with clear landmarks on the chart. We also have the doomsday fib, last line of support, at 936 which is a hair above the 900 gap.
Possibly these three levels are an indication that we will see the volatility that would be synonymous with the 1949 crash. Or possibly they mark the stop and retest from the 2002 market. It is difficult to say, and dangerous to make an assumption at this point. What these levels do however is set a good exit point for shorts (1060) and tell us that it is time to focus in when the market gets to these points.
Let's use some other price projection techniques.
Gap Extrapolation
I am probably not the only person to use this technique but I've never read about anyone using it so I'll claim invention. Gap extrapolation is based on the observation that gaps eventually get filled. So when we see a gap created we know that, eventually, the market will come back to refill that gap. Since we also know that market retracements tend towards Fibonacci ratios we can therefore project out, or extrapolate, a bottom (or top) based on how far we expect the retracement to occur.
So in this picture I've drawn hypothetical Fibonacci retracements from the 3 possible price projection points we came up with out of standard Fibonacci theory. The gap is difficult to see on this SPY chart because of an after hours block order so I've added a black bar so that you can see the gap (it is a stupendously wide gap). Note also the black support/resistance line.
The bar that most likely fills the gap on a 2002 style bounce would be the middle black bar. 119.42 is a 50% retracement from 1040. The green line, 1016, has the 50% retracement landing in the middle of the gap which is also meaningful. Often very large gaps only get partially filled.
A 38.2% retracement is only possible from the 1060 position. If this occurs then the implication is that the bear market will be ongoing and that the next stop may indeed be the 900 mark. If the market only makes a 50% retracement then some sideways action would be expected, and given the historical context I'm putting a higher probability on this and so 1040-1016 seems more likely now.
Wave Ratios
Waves are always difficult to identify but we have some landmarks we can use based on the S&P futures chart. The best landmark is retracements. A 4th wave retracement typically ends at the previous 4th wave's vicinity. As such, I believe these lines indicate the wave count. Since our wave 3 (long white line) is much longer than wave 1, there is no rule that indicates how long our current wave 5 should be. However, guidelines are that the wave 5 will either be the length of wave 1 (which it cannot be, because it was such a steep retracement) or it will be .618 times the length of wave 3. That would put the bottom at 1063. So once again, we're in the ballpark.
The most troubling thing about the current wave however is how steep it has already been. We only see one clear motive wave which took us from 1206 to 1118. We still need room for two more motive waves! One possibility is that wave 3 is the same length as wave 1 (88 points). From 1147 down this would have wave 3 bottoming at 1060. Since our wave 2 is flat we would expect wave 4 to retrace back to 1093. Then wave 5 would likely be .618 X 88 which would create a bottom at 1038.
This is really the least explosive type of wave. The most explosive wave would have wave 3 ending at 1004 and wave 5 ending at 987. I do think the prior scenario is a good one though as a lot of people will be going long at 1060 and that 1040 will look like a minor retest. We'll know more of course once we see this 3rd wave play out.
Disturbances
There are three things that disturb me currently.
1) The S&P gap at 900
This gap adds uncertainty. Given the macroeconomic conditions it is not out of the question to see the S&P at 900 (Dow at 9000). I don't think anyone would be surprised or shocked.
2) The most recent gap
This is a chart of DIA which tracks the Dow rather than the S&P. I use it because it doesn't have the spurious block/algorithmic trades that fog up the S&P chart:
You'll have to click on this to get a bigger picture to see all the gaps but even from this distance you can see the size of the most recent gap (6). Only the breakaway gap from the 2008 crash (1) rivals it in size. Basically, on Friday the Dow Jones opened 280 points lower than it had closed. That's an incredible loss of confidence without any trading activity. Such gaps are usually reserved for extreme market crashes like 2008. The length of wave 1 (as discussed in the previous section) was suitably long given the gap but I'm concerned that it is an indication of extreme bearishness that might be an omen for a lower end point.
It's worth looking at how gaps provide a roadmap for the market. The bear gaps are indicated by black numbers and the bullish gaps by green letters. One important lesson is to always trade in the direction of the gap. This is especially the case for breakaway gaps that occur at the end of a turning point. One could have gone long at point A or point B and experienced incredible profits, even weathering out the flash crash. Likewise, in our current market, points 5 and 6 are clear indications to go short. Runaway gaps (gaps that occur in the middle of a trend such as points 1, 3, 4) are more difficult to read. Often we'll see a gap at the end of a trend and won't know in retrospect whether it is runaway or not.
For my money, I believe that gap 6 is most closely related to gap 2. It's just that the size of the gap gives me pause.
3) Elliot Theory
Elliot Theory tells us that at this stage we should be seeing a 3 wave correction, not a 5 wave. The crash pattern however is almost certainly painting out a 5 wave. I think wave reading at the major trend level is difficult if not dangerously misleading. I prefer the historical context but let's analyze the possibilities. The first possibility is that this crash does indeed end the 3 wave by either of these possible counts:
The blue count might be the "obvious" count seen from a distance. We know from close up observation however that this is probably not correct. The white count has the little step counted as wave 1. This is all out of proportion and so seems unlikely as well.
Given that this current wave is a 5, we should be expecting a zig zag 5/3/5 pattern. That might look something like this:
If we're going to get to the 900 range, then the zig zag is how we're likely to get there. Note that the middle wave (3) should be easy to identify as a non-motive wave and give us plenty of warning if this is going to happen.
A zig-zag however is indicative of a market that is in a hurry to correct in order that it can proceed upward. This doesn't really match the current economic conditions which look to play out over a longer time period. I've trained myself to recognize however that usually when a wave pattern is unclear that it means that the pattern being played out is much larger than expected. Given the historical context, we should expect this bear/sideways market to take much longer than a few weeks to play out! So perhaps we're looking at something more like this:
While it is difficult to come up with a satisfying wave count for this scenario, I believe nevertheless that it is the most likely. It matches the historical context (1978 being a false rally). It also provides the most time for the market to play out the economic conditions. A possible QE3 coming out of the Jackson Hole meetings might provide the context for such a false rally.
This pattern also leaves room for another rally and retracement as wave 1 of the next big boom. Such a pattern occurred in the 1979-1982 market.
Timing
Note the lines in this chart:
The S&P is on a clock of sorts. Throughout the previous corrective period the dips occurred in the middle of each month. During the flash crash correction the peaks occurred in the middle of the month. Likewise, we now observe that this previous peak occurred in the middle of the month.
Generally speaking, the opposite turning points occur at the beginning of the month or thereabouts. Unfortunately these are not quite as regular as the peaks and valleys but given the current pattern and timelines I believe we can expect a bottom around the end of this month. Coincidentally, this corresponds with the timing of the Jackson Hole meetings. It's hard to imagine, given the current market state, the S&P clock, the retracement projections, etc that the results of Jackson Hole wouldn't provide a buying opportunity. Whether the effect lasts or not is harder to say. I believe however that we will be able to watch for ongoing wave patterns based on this same clock.
Most importantly, note the two black lines denoted by points "A" and "B". These were both points that I call "truncated markets". In both instances the market looked like it was going to take off but instead immediately slipped into a steep correction. I believe the cyclical timing is the key to catching these. Those were on the bull side. Now I believe it would be the opposite. Watch out for a steep dip (wave 5) towards the end of the month that is in reality the beginning of a bounce.
Other Related Assets
First let's look at the dollar:
The market requires a falling dollar in order to rise. Currently, the market will fall if the dollar rises. It is also likely to fall if the dollar stands still or wiggles around. Thus, we've had a market crash during what is essentially a sideways movement of the dollar. As everyone knows, the market has been on life support and without a falling dollar it is doomed.
It does look though like the dollar is falling. the white arrows point to clearly motive waves. The trend for the dollar has turned down. Now, on the minor trend level it does look set to climb a few notches before this overall trend continues. The green lines are enough to get the market down to the 1060 level I believe.
Now let's look at the long bond:
It's been a hell of a ride for 30 year treasuries. The yield spread recently dipped below 2%. The ride is not over yet. We are clearly making out a sideways pattern. This is very high probability for a 4th wave which means treasuries have another notch up. Wave counts can always extend but I believe we're seeing the 5th of a 5th here which should be followed by a correction back at least as far as the arrow shows if not further. Overall, treasuries are in the territory that they were in at the height of the financial crisis. It's all out of proportion with what is actually going on in the equity markets so I wouldn't be at all surprised to see a significant retracement.
Gold
As noted in my previous blog piece, gold has recently gone parabolic. This never ends well. However, like treasuries it appears as if the wave pattern is not played out yet. I'm estimating 1930 as the top. Gold should at least fall back to 1750 but most likely to 1500. Unfortunately, very few are ever willing to see at the top of a parabola. Greed takes over.
Market Survey
This blog post has taken much longer than I had anticipated but I can't finish without a quick market survey.
I'll start with Corning because a good friend of mine trades it and it got a mention in Barron's this week. I'm concerned though. Here's a weekly chart:
This is big picture but note the pattern. A rocky climb over an extended period and then a steep drop off. The last climb has the tell tale ragged action of a B wave. The implication is that this current motive wave downward is not finished. I think the long term support line is indicative of a bounce point around 11 for this stock. Note the thin cover in the 12.50 area which will probably serve as an intermediate stop. My expectation is that it bounces around this point back up to 16-17 and then comes down to finish the 5th wave (as indicated on the chart). If GLW has any implication for the broad market it is that we will indeed see lower lows after a retracement.
We've looked at Microsoft before but let's look again:
So many people are long Microsoft as it is a core component of most indexes and most mutual funds. Microsoft is a darling because of the cash on the books and market position. For those in the technology know however, we know that Microsoft is a sinking ship. The technicals seem to confirm this. Once again on the weekly chart note how MSFT appears to be making a B line for the thin spot which is the equivalent of S&P 900. The wave count is a decent read. Most damning are the stochastics which indicate that MSFT has quite a bit further to fall before it gets a bounce. Long term it probably eventually winds it's way back up into the 30s, after all so many people own it that it will take a lot of misses, and a lot of spent cash before it comes down. Decades perhaps. Still, I wouldn't be long in the near term.
Apple also looks alarming on a weekly chart:
Note how deep the stochastics are running. Note the divergence between composite and RSI (green and red lines pointing different directions). Finally the Elliott Wave count looks like it wants AAPL in the red circle before we get a kick back upwards for a 5th wave.
Home Depot on the other hand looks strong:
Stochastics bottomed already and now are making bullish divergence, as is the composite and RSI. Meanwhile, HD hasn't come back to retest the lows like the rest of the market. I suspect the worst is past for HD, although it will still probably duck back below 30.
Proctor and Gamble also looks relatively strong:
Can't say I like the broadening top, and the X marks the point where one would definitely want to bale on this property, but in the meantime stochastics look good, it is holding up better than the broad market and we have good divergence on the indicators.
Con Ed certainly has weathered the storm:
Looks like it has further to climb although it might just go flat if the general market rebounds. Brewing bearish divergence on the RSI although still early to tell.
Caterpillar I think tells the tale for what happens to the market:
Here on the daily chart we see a Fibonacci confluence zone that predicts a bounce. We also see brewing bullish divergence on the composite. However, the gap marked "eventual" coincides with a .618 retracement line and there's not a lot of support above that line. So I think CAT bounces at 75 but then eventually winds it's way down to 66.
Judging from the market survey I think going long at 1016-1060 is likely to be a temporary position as there is just too much evidence that the market will fall back further.
What we do know from a long term perspective however is that the gaps at 1190 and 1330 provide eventual upside targets. A long position taken in the 1016-1060 range is likely going to have to weather some downside paper loss but should eventually yield a nice profit. We should be on the lookout for a breakaway gap to the upside as a sign to go long with a large position, otherwise a modest position I believe is most warranted for playing the likely retracement off of a 1060 bounce.
0 comments:
Post a Comment