Written by and copyright © 2005-2018 by Thomas N. Bulkowski. All rights reserved. Disclaimer: You alone are responsible for your investment decisions.
See Privacy/Disclaimer for more information. Some pattern names are the registered trademarks of their respective owners.
- Thursday, 2/26/09. Aphids Attack and Yahoo Dirty Data.
- Monday, 2/23/2009. Tutorial Tuesday on Monday!
- Thursday, 2/19/2009. How Now Dow?
- Tuesday, 2/17/2009. Tutorial Tuesday: What You Don't Know About ETFs.
- Thursday, 2/12/2009. Another Exchange Rule Gotcha!
- Wednesday, 2/11/2009. Scrubbed on Account of Weather.
- Tuesday, 2/10/2009. Tutorial Tuesday: Moving Averages and Chart Patterns, Take 2.
- Thursday, 2/5/2009. GLD Wedge. What Does it Mean?
- Wednesday, 2/4/2009. Dow Utilities Goes Horizontal
- Tuesday, 2/3/2009. Tutorial Tuesday: Learn How To Make Money Trading Patterns.
Thursday, 2/26/09. Aphids Attack and Yahoo Dirty Data.
I spotted a bud on my evening primrose and decided to take a closer look. The bud turned into a ladybug. Uh-oh. If there are ladybugs around then there must be aphids.
Yes indeed, as the picture shows, those bugs are aphids. I doused them with insecticidal soap but fear that only gave them a refreshing bath. In fact, this picture was taken after
I sprayed them with soap. If you look closely, you can see one sticking its tongue out at me and smiling (or giving me the finger. It’s hard to tell).
"You need an integrated pest management system"
Ben F. told me a few years ago. My version of that is a blowtorch. I’ll have to put some dollars into the economy and buy one.
# # #
The chart pattern indicator (CPI at the top of this page), has turned bearish again after signaling the start of a bull run yesterday. The indicator should be viewed as a weekly
indicator, so ignore these brief changes.
Yahoo!finance is playing games with their data again. Yesterday, they listed Home Depot’s (HD) low price as 2.70 when it should be 18.24. Plus, their Up/Downgrades page shows data from
February 9. They do this from time to time -- showing old pages instead of updating the information with new content. For the last several days, old pages have appeared more frequently than
new stuff. But at least it’s free.
The indexes might be forming a bottom here, but given the state of the economy, they probably will continue lower. I adjusted Tom’s Targets (see page top, right) again on Feb 23,
a day after changing them.
The Wilder RSI test portfolio finally issued a signal today. I made a change for the new year to only show performance numbers for that year on all of the
test portfolios. Since then, no updates have occurred (See the upper right of this page, under "Test Portfolios and CPI"). I thought I broke something.
I am happy to report that my cost of living was $10,743 last year. That includes everything except for property and incomes taxes. A full 40% of that total went to insurance!
Linda B once accused me of buying two-ply toilet paper and separating the plies. I’m a gardener. I use leaves.
-- Thomas Bulkowski
Monday, 2/23/2009. Tutorial Tuesday on Monday!
The mutual funds where I have my IRA managed to lose 20% in one month, so I got fed up and moved them into cash. I sold off my non-core holdings, too. That was on Friday. This week,
since I am mostly out of the market, I expect it to rebound.
I am writing this on Sunday and looking at the Dow industrials on the monthly chart, the average has not moved below the October 2002 low of 7181, so we are still in a
support zone. Looking at the other indexes, the price projections using a measured move down chart pattern show some of them already reaching the price target for the
measure rule. Thus, I see a rebound happening this week.
My new price targets appear at the top of this page under Tom’s Targets. If you click on the link, a window will appear to explain the number, so I won’t go
though each one here. However, let me discuss how I used the measured move down to get a price target.
The above chart shows the S&P 500 index on the daily scale. The measure move down is the pattern ABCD. The idea behind the pattern is that
the second leg, CD will match the extent, slope, and duration of the first leg, AB. Thus, since you know the extent
and duration of the first leg, you can estimate the price and time targets of the second leg.
Let’s go through the numbers. In the calculation, I use the lower peak at C because it starts the downward swing. You can use the high on January
28, if you prefer. The drop from A to B is 943 - 804 or 139 points. Since price peaks
at C at 875, the full measure target would be 875 - 139 or 736. However, page 505 of my book,
Encyclopedia of Chart Patterns, Second Edition,
pictured on the left, says that the second leg averages 19% shorter (bull market) or 20% (bear market) than the first leg, so I adjust the length by 20% to get a closer target of 763.
You can use half the length and price
reaches or exceeds the target 93% of the time in a bear market.
Since Friday’s low reached 754, I expect a bounce upward to the corrective phase,
BC. The book explains that price stops below the corrective phase 20% of the time in a bear market, stops within the phase 52% of the time, and
rises above it 28% of the time (20% below the top of the measured move and 8% continue higher). Thus, there is an 80% chance that price will make it to the bottom of the corrective
What about the time projection? For Tom’s Targets above, I just guessed, based on how price has been moving in the past. For a more accurate projection, we can use the
measured move down. The AB span from January 6 to 21, is 15 days. Projected from the peak at C gives a
target date of February 24, if the second leg matches the first leg. The statistics show that the second leg averages a day longer than the first leg (46 days versus 45). Since
this measured move down is shorter than average, I will leave the time projection alone.
Finally, notice that the slope of the two legs (AB and CD shown here by two
green lines) are nearly equal, offset by the corrective phase. This is typical behavior for a measured move down. Often the slope of the two
lines will vary somewhat.
-- Thomas Bulkowski
Thursday, 2/19/2009. How Now Dow?.
I show the Dow Jones industrials (^DJI) on the monthly scale to highlight underlying support. I show that support as a horizontal red line stretching back to
1998. This is a formidable support zone that stopped the decline during the 2000 to 2003 bear market and as recently as November 2008.
Will support hold this time? Good question and only time can answer it. However, I show three possibilities.
First, I show route one in red. This is a continuation of the downward price trend. If you listen to the talking heads and the news reports,
this is what will happen. The economy is falling apart, the world is going bankrupt, and our lives will be over in a few months. An in neck candlestick
today supports this since those breakout downward 53% of the time. Of course, that is almost random.
Route two appears in green. It shows a bounce off the support zone back to the bottom of the congestion region , stopping at the round number of
8,000. Since nearly all pullbacks (87% of them) bounce off overhead resistance and continue lower, that is what I show. Right now, this would be my guess
as to what will happen in the coming week or so. However, Tom’s Targets (at the top of this page) do not reflect this scenario.
Finally, we have route 3, shown in blue. I know, the blue looks black. You may be losing your eye sight.
This is for people wearing rose colored
glasses. It shows price pushing through overhead resistance and staging a strong recovery. No one but me believed this scenario until I lowered my price targets at the top of this page
a week or so ago. Hope springs eternal.
I will adjust Tom’s Targets as the direction becomes clear. For the record, the Dow transports have voted for route 1 by pushing below a support region at 2,900. They may be
pointing the way lower. Or not.
-- Thomas Bulkowski
Tutorial Tuesday: What You Don't Know About ETFs.
When the 2x and 3x exchange traded funds (ETFs) came out, I thought I could just load up on them and double or triple the performance of the major indexes in the coming year by
buying the correct ETF. Sadly, that is not the case.
You probably know that there are ETFs that follow the indexes both from the long side and the short side. For example, the DOG fund is supposed to mimic the Dow Industrials from
the short side. Here is how yahoo!finance describes it: "The investment seeks daily investment results, before fees and expenses, which correspond to the inverse of the daily
performance of the Dow Jones Industrial Average index."
A fund like DXD is designed to produce twice the inverse performance of the Dow, but there’s a problem. I call it tracking error.
Here is what the ProShares website says about it.
Investors considering ProShares and other leveraged or short funds with daily objectives should not expect them to achieve their objectives over periods longer than one day
(emphasis added), however.
There are several reasons for this, but the most significant one is index volatility and its effect on fund compounding. In general, periods of high index volatility will cause
the effect of compounding to be more pronounced, while lower index volatility will produce a more muted effect.
I decided to found out how well the funds do over the long term. Since the funds are new to the markets, their performance lifetime is short. The following shows some of the
more popular ETFs.
Notes: DJ 15 Utilities, ^DJU
DJ US Utilities sector index fund, IDU
UltraShort 2x Utilities ProShares, SDP
Look at the above table. ^DJU is the Dow utility average. IDU is a sector fund and SDP is the new kid on the block, a fund that shorts the utility stocks at twice the volatility.
The first row measures the close to close performance from January 2, 2001 to January 2/2002. Over that period, the Dow dropped 25.33% but the index fund dropped 22.87%. That is not
Compare that to 2006 when the Dow gained 25% and SDP gained just 18%. In 2007, the ETF over-performed, beating the Dow by 17% to 11%.
Now look at the last row. The Dow dropped 27.94%, the index fund dropped 28.66% and the 2x short fund gained 28.85%. While a gain is what the fund was designed to do (since it performs
to the inverse of the Dow), it is far from the target 27.94% x 2 or 55.88% return that one might expect.
Dow Jones Industrials
|Year||^DJI||DOG (-1x)||DXD (-2x)||DDM (+2x)|
Notes: DJ 30 Industrials, ^DJI
DJIA short 1x ProShares, DOG
DJIA short 2x ProShares, DXD
DJIA short 2x ProShares, DXD
Above is the comparison for the Dow industrials. The DXD, a 2x short fund, doesn’t come close to twice the inverse Dow’s performance over a year.
The DDM fund at 2x, comes much closer to the -61.48% target.
|Year||^NDX||PSQ (-1x)||QID (-2x)||QLD (+2x)|
Notes: Nasdaq 100, ^NDX
Nasdaq 100 short 1x ProShares, PSQ
Nasdaq 100 short 2x ProShares, QID
Nasdaq 100 long 2x Ultra QQQ ProShares, QLD
The first thing that is tricky about these funds is that they do not track the Nasdaq Composite. Rather, they track the Nasdaq 100.
During 2007-2008, the QID (-25.2%) and QLD (24.51%) funds did not double the performance of the benchmark (33%)
even though they are 2x funds.
In 2008 to 2009, the tracking was closer to the 2x target, 57.58% and -69.52% versus 76.70%.
|Year||^GSPC||SH (-1x)||SDS (-2x)||SSO (+2x)|
Notes: S and P 500, ^GSPC
S and P short 1x ProShares, SH
S and P short 2x ProShares, SDS
S and P long 2x Ultra SP 500 ProShares, SSO
The above table shows the ETFs against the S&P 500 index. Notice in 2007-2008, the 2x (long) SSO fund lost money even as the benchmark gained 2%.
What these tables show you is that holding an ETF for the long term can produce better or worse results than you expected. As Monty Python would say, nobody expects
-- Thomas Bulkowski
Thursday, 2/12/2009. Another Exchange Rule Gotcha!
The storms last night blew through quickly and tore the roofs off a number of homes miles from mine. I have been fortunate in not suffering major damage
in the 27 years I have lived here. But, tornado season is approaching and with global warming, storms tend to be stronger. With the way my luck has been running for
the last two years, I will be toast one of these days.
This morning I placed a market order to buy a stock 3 minutes before the open and then returned to my charts. They were showing the premarket price climbing, so I knew that the
stock would drop in a few minutes of the open, so I canceled the buy order two minutes before the open. Guess what? The cancel was denied even though the market was closed. Thus, the order went through
and the stock dropped by $1.20, minutes after the opening, costing me $840 on the 700 shares. I bought at the high for the day.
So, I called up my broker for an explanation. They said it is an exchange rule that they can deny a cancel on a market order within minutes of the opening bell. To get
a better definition on the rule, I called again (and spoke to a different representative) and they said the rule applied to the last 2 minutes before the open and it was
a rule that Nasdaq has, so any orders going through Nasdaq would be subject to this limit. And he also said that I was within the 2 minute window by just six seconds.
I like learning from my mistakes, but I just wish they were not so costly.
-- Thomas Bulkowski
Wednesday, 2/11/2009. Scrubbed on Account of Weather.
I live in tornado alley, which is a section that runs through the center of the United States. That is where the most tornadoes occur each year. Today, they are expecting
severe weather to begin in about an hour and lasting into the night. Thus I won’t be able to post the usual entry for tomorrow (Wednesday).
If my house is still here tomorrow, I’ll blog then for Thursday’s entry.
-- Thomas Bulkowski
Tuesday, 2/10/2009. Tutorial Tuesday: Moving Averages and Chart Patterns, Take 2.
I released the study of combining moving averages and chart patterns. This is the same as the one I discussed a week ago but botched the risk numbers.
I learned a few things that I want to share here.
For upward breakouts, I consider a chart pattern failure as one that shows price failing to climb at least 5% after the breakout before plunging by at least 20% or
by closing below the lowest low in the chart pattern. Most chart patterns can easily reach the 5% benchmark, so the failure rates are typically less than 5%. However, in
the study I used a 15% failure rate. A move of that magnitude often means a trader can profit from the trade. I found that the average failure rate was:
- For upward breakouts in a bull market: 32.2%
- For downward breakouts in a bull market: 41.0%
- For upward breakouts in a bear market: 39.7%
- For downward breakouts in a bear market: 25.6%
The numbers say that you should trade with the trend for the highest chance of success (upward breakouts in a bull market and downward breakouts in a bear market).
This is something we all know, but the numbers prove it.
You can reduce the failure rate if you filter your selections by using a moving average. Here is what to use.
|Market||Breakout||Moving Average Position|
|Any||Up||The day before the breakout closes below the 9 day SMA.|
|Bull||Down||The day before the breakout closes above the 9 day SMA (best) or below the 50 day SMA.|
|Bear||Down||The day before the breakout closes below the 50 day SMA.|
For example, if this is a bull market and you expect an upward breakout from a chart pattern the next day, then the closing price should be below the 9 day simple moving average.
If this is a bear market and you expect a downward breakout, then the close should be below the 50 day SMA. If your situation does not agree with the combination shown in the table,
then look elsewhere for a more promising trading setup.
-- Thomas Bulkowski
Thursday, 2/5/2009. GLD Wedge. What Does it Mean?
The chart of GLD, an exchange traded fund that, according to yahoo!finance, "seeks to strive to reflect the performance of the price of gold bullion, less the Trust’s expenses. The Trust holds gold, and is expected to issue baskets in exchange for deposits of gold, and to distribute gold in connection with redemption of baskets. The gold held by the trust will only be sold on an as-needed basis to pay trust expenses, in the event the Trust terminates and liquidates its assets, or as otherwise required by law or regulation."
I show the chart on the weekly scale to highlight a descending broadening wedge, shown here outlined by two red trendlines.
According to my book,
Encyclopedia of Chart Patterns, 2nd edition,
the chart pattern breaks out upward 79% of the time. In a bear market, however, the rate changes to 66% of the time. That means an upward breakout two out of every three trades.
Feel like taking a chance and buying it now? I would wait for a breakout (a close above the top trendline) since the trendline can pose formidable resistance. A
partial decline has not appeared but it could if the price of gold drops. That might be a buy setup. In a bear market, price reaches the top of the pattern
just 58% of the time. That would mean a rise of about 12% or so, and that is if everything worked perfectly. And in this market, that is a bet I am unwilling to take.
-- Thomas Bulkowski
Wednesday, 2/4/2009. Dow Utilities Goes Horizontal
First, let me say that I botched the failure rates of the study on moving averages (yesterday’s post). Using a moving average in concert with a chart pattern breakout
can reduce risk but not anywhere near the 85% or higher rates that I was seeing. The reduction is on the order of 10% or so. I will complete the corrections and post the study
when it is completed.
I show the Dow utility average (^DJU) on the daily scale in the chart to the right. What I find interesting about the chart is how price has moved in a horizontal direction since October. This exemplifies
the trading range behavior of the other indices. By that I mean the indexes are range bound as if they are unwilling to trend in any direction for a long time.
The chart shows a symmetrical triangle, highlighted in red. The breakout from a symmetrical triangle in a bear market
is downward but only 52% of the time. In other words, the breakout direction is random.
The utility average is the only index that is in positive territory for the year. Since most of my portfolio is in utility stocks, I am pleased as punch. In fact, I sold more of
CHG yesterday and received a fill at the exact high for the day: 52.23. If the market continues lower, the utilities tend to act as bonds and are a safe haven for investors (so I’ve read).
With one utility offering a yield of about 7.5% (AEE, which I also own), I can understand why.
Anyway, I am not going to predict the breakout direction from this symmetrical triangle, but I hope it is upward. That would likely mean that the other indexes are also rising and
that would be good news for everyone’s portfolio.
-- Thomas Bulkowski
Tuesday, 2/3/2009. Tutorial Tuesday: Learn How To Make Money Trading Patterns.
Addendum: I believe this study is in error because of a mistake in the failure rates. In other words, disregard the following post until I can check the math!
Last week, I discussed how chart patterns fail during the 2000s bull market twice as often as they did in the 1990s. Today, I am going
to tell you how to make money trading chart patterns.
As part of my year end trading review, I am researching new avenues that I had ignored or overlooked in prior years. One of those is to use indicators along with chart pattern buy
signals. I discovered that if you combine the breakout from a chart pattern with price being below the 9 day simple moving average, you will increase your profits and lower your
risk of a failed trade (for upward breakouts).
The chart shows an example of a symmetrical triangle chart pattern, outlined here in blue. The 9 day simple moving average
appears in red. The day before the breakout, the closing price is below the moving average. The inset in the blue
box shows this more clearly. Point B is the close on the breakout day. It is above the top of the chart pattern, signaling a buy. The close the
day before the breakout is at A, and it is below the moving average (which appears as the red line).
That combination, an upward breakout from a chart pattern along with the prior day’s close below the 9 day simple moving average, signals a buy.
As you may know, a bounce often occurs after price makes a swift decline. The dead-cat bounce event pattern is based on that behavior, for example. This
trading setup takes advantage of that situation without relying on such a strong decline.
I tested both simple and exponential moving averages and found that the simple ones work better than the exponential, much to my surprise.
I tested the 9, 20, 50, and 200 day moving averages and did preliminary
work with the 3, 75, 100, and 150 day SMAs also. The 9 day SMA worked best on over 21,000 chart patterns from 1989 to 2009. I found that profits increased marginally but the failure
rate dropped by 86% for upward breakouts and by 85% for downward breakouts in a bull market. In a bear market, the failure rates dropped by 90% (upward breakouts)
and 78% for downward breakouts. Before you get too excited, a failure is when price moves less than 5% away from the chart pattern before reversing and making a large move in the new
direction (at least 20% or closing beyond the side opposite the original breakout).
I ran my actual trades against the 9 day SMA and found that my win/loss ratio increased by 16% and profits soared by 340% using this method. Not all of those trades used chart
patterns and I compared the moving average to the closing price the day before I bought instead of a chart pattern breakout.
I will be releasing the full details of this trading setup in the near future.
-- Thomas Bulkowski