Written by and copyright © 2005-2022 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.

Archives

• April Archive
• February Archive
• January 2010 Archive
• All Archives

Tuesday 3/30/10. Tutorial Tuesday: What is Best Day to Buy or Sell?

The headline tells it all. When is the best day of the week to buy or sell? The answer depends on the market you trade, and if you want to go further, the stock within that market.

To answer the question for the general market indexes, I counted each day in which price closed higher as a percentage of the total of those closing higher and lower (unchanged days are not counted but they are few). The above figure shows the results. Bars below the horizontal black line mean price closed lower more often than higher.

For example, the green bar is for the Dow utilities. Using all data that I have on the index, the utility index closed higher most often on Monday -- 54.3% of the time. It closed lower most often on Thursday, 48.6% of the time.

The following table shows the results in tabular form.

Index	Best Sell Day	Best Buy Day	Data Start
Dow Utilities	Monday or Friday	Thursday	1/2/1990
Dow Transports	Monday	Thursday	1/2/1990
Dow Industrials	Wednesday or Friday	Monday	10/1/1928
Nasdaq	Wednesday, Thursday, or Friday	Monday	2/5/1971
S&P 500	Wednesday or Friday	Monday	1/3/1950

The Best Sell Day column is the one with the most up closes. The Best Buy Day column is the one with the fewest up closes. The Data Start column shows the date of data available for analysis.

So which is the best day to buy? If it's a tech stock, then I would buy near the close on Monday because the index closed lower on that day and you'll be buying at a bargain. The best day to sell would be near the close on Wednesday because the composite closed higher most often on that day but Thursday and Friday also look good.

For a stock in the S&P 500 and the general market, buy near the close on Monday and sell either on Wednesday or Friday.

If you want information on other time periods, then I have posted my research here.

-- Thomas Bulkowski

Monday 3/29/10. Market Monday: The Week Ahead

A Brief Look Back

The following are economic reports that moved the markets last week. The numbers refer to the close-to-close move in the Dow industrials.

For the week...

Economic Reports

The following information is derived from yahoo!finance and sometimes Bloomberg.com with times local to the east coast.

Options

Quarterly expiration date on Wednesday, March 31.

CPI: Chart Pattern Indicator

As of 03/26/2010, the CPI had:

The chart pattern indicator is bearish with 1 of 3 half triangles showing (). Additional triangles are a measure of strength with solid triangles meaning a more reliable signal than half triangles.

Other

Earnings season is over.

The following industries, of 52 that I follow, were the best (1) and worst (52) performing.

My Prediction

The chart shows the Dow transports on the daily scale. Price is already heading lower and I expect it to continue until it reaches the target area. Bordering that area is the 38% retrace of the move up from the February low on the bottom and the January peak on the top.

The Dow industrials, Nasdaq, and S&P 500 have not displayed this type of decline or really hinted of a decline. Nevertheless, I changed Tom's Targets at the top right of this page to bearish. This is based on a feeling more than anything else. Clearly the transports have stopped rising and the Dow utility is also headed lower but the other three are only hinting of a turn.

Looking at individual stocks, some in weak industries are headed lower (oil services come to mind). I plan to sell one oil play at the market open to capture a 20% gain before I lose it all.

The chart pattern indicator, discussed earlier in this report, has turned bearish and could stay that way but it really depends on what the general market does. It suggests weakness ahead but with only 1 half triangle showing, it's a weak signal so far. The 4 employments reports coming out on Friday could be the catalyst but since the market is closed that day, nothing much will happen. Certainly good news has not lifted the markets like one would expect. Will bearish news take it down by a few hundred? Perhaps we will see.

-- Thomas Bulkowski

Thursday 3/25/10. SEC forms and What They Tell You.

When I was doing Dutch auction tender offers, I checked form TO-I. Why? Because that's the SEC form companies use to announce their surprises. Here's a list of forms you may find useful when researching a company. You can go to the SEC website to search for forms and schedules.

Sources: http://sec.gov/info/edgar/forms/edgform.pdf and http://www.secfile.net/SEC_calendar.htm#General_filing_deadlines

-- Thomas Bulkowski

Tuesday 3/23/10. Tutorial Tuesday: Breakout Day Volume Review

I finished my research on volume using chart patterns, and I want to share my findings with you. Here are excerpts from the research.

In the above figure, the black line represents upward breakouts. As breakout day volume from a chart pattern begins well below average, the gain drops but turns upward at the 20% volume mark (horizontal axis). It rises in an arc, approaching the vertical red line. That line separates below average from above average volume. As breakout day volume increases, so does the percentage gain (left scale).

Downward breakouts use the right scale and those numbers represent losses. When breakout volume hits 50% of the average (pink line), it marks the low point. From there, the average loss increases from about 16% to as high as 26%. The results are similar to upward breakouts: As breakout volume increases, the post-breakout decline in a chart pattern also increases.

Here are other findings...

• As breakout day volume increases, chart pattern failures drop.
• Throwbacks and pullbacks occur more frequently after above average volume breakouts.
• For upward breakouts, throwbacks occur 74% of the time if the breakout is on above average volume. For downward breakouts, the pullback rate is 68%.

-- Thomas Bulkowski

Monday 3/22/10. Market Monday: The Week Ahead

A Brief Look Back

The following are economic reports that moved the markets last week. The numbers refer to the close-to-close move in the Dow industrials.

For the week...

Economic Reports

The following information is derived from yahoo!finance and sometimes Bloomberg.com with times local to the east coast.

No options expire this week.

CPI: Chart Pattern Indicator

On 03/19/2010, the chart pattern indicator (CPI) had:

The chart pattern indicator is bearish with 1 of 3 half triangles showing (). Additional triangles are a measure of strength with solid triangles meaning a more reliable signal than half triangles.

Other

Earnings season is over except for those late to the party.

The following industries, of 52 that I follow, were the best (1) and worst (52) performing.

My Prediction

Ilozagrav (and I thought Bulkowski was difficult to spell...) in an email commented about Elliott wave in the S&P 500 index. Since I haven't talked about the Nasdaq as often as other indexes, I show the composite on the weekly scale. Motive wave 1 begins an upward thrust followed by a correction at wave 2. Wave 3 is the longest of the bunch followed by corrective wave 4 and the most recent motive wave 5.

Sounds like a handful, doesn't it? I show the idealized Elliott motive wave with red numbers. Three waves higher (1, 3, 5) and two moving lower, (2 and 4) followed by an ABC correction.

If you're not into Elliott wave the you may not know that if the primary trend is downward, the wave motion flips. You get the 5 waves heading down and the ABC correction trends upward.

For fun, look at the Dow Transports on the weekly scale for 1 year duration for best viewing. How many waves do you count? Answer 7. How can that be?

Since Elliott couldn't get his wave counts to work, he added exceptions called extensions. Any wave can extend and extensions can have extensions, too. Thus, non-believers will tell you that few Elliott wavers can agree on the count and even the God's can't get it right, sometimes.

I agree with Victor Sperandeo who said, paraphrasing, that surfing the Elliott wave is too subjective to be useful...except for the ABC correction. He said that had value.

What does all of this mean for the coming price action?

Let's say the wavers are right. Price should retrace in an ABC fashion as the idealized chart shows. When will that retrace begin? I have no idea. I think that we could get an extension such that wave 5 leads to waves 6 and 7 and maybe more stair-step moves up. So, I remain bullish but cautious.

The chart pattern indicator has flipped to bearish on Friday, and judging from the hit my portfolio took, the general market got whacked much more than the Dow suggested. A large move up on Monday or something during this week could erase the bearish reading. That's happened in the recent past. So, we'll have to wait and see over the coming days. If it hasn't turned bullish by Thursday, then it probably won't. I may have to change my targets then.

-- Thomas Bulkowski

Thursday 3/18/10. New Volume Discovery!

The Dow industrials has closed higher for 7 days now. There's a candlestick called 8 new price lines that's based on 8 higher highs (not closes). It acts as a bullish continuation pattern 53% of the time, which is what I call "near random." It's supposed to act as a bearish reversal but if that were true, there would not be candles called 10, 12, and 13 new price lines. Anyway, it does raise a red flag suggesting that when the turn comes, it could be violent with the Dow losing a few hundred points in one session. As the picture shows, I'm bullish, but concerned.

# # #

I wanted to mention a discovery I made in my research on volume. I released the new page discussing the research, and I'm not done yet. Since I've been working on this since Saturday, I figured I might as well blow the rest of the week and complete the job.

The discovery concerns failure rates. For the test, I defined a failure of a chart pattern as a climb of less than 10% after the breakout. Assume that tomorrow is the breakout day. Total the volume over the last 5 trading days and compare that to the total of the 5 trading days before that. If the most recent 5 days has a lower volume than the prior 5, then do not buy the stock if it should breakout tomorrow.

Why? Because patterns that have a lower volume trend in the two weeks preceding a breakout suffer 31% more failures that do those stocks with volume trending higher.

To find out more, it's best that you read the entire volume page, but you can start with failures.

-- Thomas Bulkowski

Tuesday 3/16/10. Tutorial Tuesday: New Volume Research

I released the new volume study here.

I conducted new research to see if I could prove that high or low volume is beneficial to stock picking. To do that, I used my database and found 15,444 chart patterns going back to July 1991 in 1,093 stocks. I measured from the close the day before the breakout to the ultimate high or low. That's the highest high (upward breakouts) or lowest low (downward breakouts) before price changes trend (a move of at least 20% in the new direction). Thus, this method does not depend on the skill of an exit rule or sell scenario. The results represent perfect trades.

I looked at volume over 3 periods: 3 months, 3 weeks, and 3 days before the breakout. I will discuss findings for upward breakouts from 3 months to 3 weeks in this posting.

In the discussion that follows, imagine you take volume readings over 3 periods and they are 7, 6 and 8, respectively, with 8 being the most recent. The volume trend between 7 and 6 is down and between 6 and 8 it is up. Thus, when I refer to down and up (or whatever combination), you'll know what I'm talking about. The month or week begins with the day before the breakout as the most recent week (corresponding to 8 in this example). A volume pattern of 8, 6 and 7 would be down and up, as another example...

Results: Over 3 Months

I found almost no performance difference for patterns with volume above or below the prior month in the 3 months leading to an upward breakout. In other words, I split the 3 months before the breakout into three, one-month periods, and compared the volume between months 3 to 2, and 2 to 1. The best performance came when volume trended down (month 3 to 2) then up (month 2 to 1) leading to the breakout. The 2,306 samples fitting that scenario gained 34%. The other combinations of up and down volume showed gains of 32%. Yawn.

Results: Over 3 Weeks

Using the same method for 3 weeks before the breakout, I found the gains to be similar: 34% (up and up) to 31% (down and down). In other words, if volume trends upward from week 3 to week 2 and again from week 2 to week 1 before the breakout, the stock performed marginally better than the worst combination of volume trending lower over the three weeks. Another yawn.

Failures

If you consider that a rise of 10% or less to be a failure and a gain of 50% to be a winner, the best volume combination is up and up with 15% and 26% of samples showing losses/wins. A low loss number is best and a high win number is best. The numbers mean that 15% of the chart patterns failed to rise at least 10% after the breakout if the volume pattern was up and up. Twenty-six percent of the patterns met or exceeded a gain of 50% when the volume pattern was up and up. The following table summarizes the weekly results.

Volume	3=>2	2=>1	Gain	Wins	Losses
Best	Up	Up	34%	26%	15%
Worst	Down	Down	31%	24%	21%

If you read the study, here's what I conclude.

• The difference between rising and falling volume in the 3 months before a breakout from a chart pattern means a performance difference of 2 percentage points for upward and downward breakouts. A volume pattern of falling then rising leading to an upward breakout results in the best performance. For downward breakouts, rising volume leads to underperformance (a smaller drop).
• In the three weeks before a breakout, rising volume for upward breakouts outperforms the other variations but the difference between best and worst is just 3 percentage points. Downward breakouts show the worst performance if volume trends down then up, but the difference between best and worst is just 2 percentage points.
• Three days before a breakout, the best performance comes from a falling volume trend for upward breakouts. The results for downward breakouts are mixed. For upward breakouts, the performance difference is 2 percentage points, and for downward breakouts, it's 1 percentage point.
• Upward breakouts: Dramatic increases or decreases in volume over the two weeks before the breakout can lead to better post-breakout performance.
• Downward breakouts: As volume decreases over the 2 weeks before the breakout, so does the potential loss. Dramatically increasing volume tends to lessen the loss, but not to a large degree.

-- Thomas Bulkowski

Monday 3/15/10. Market Monday: The Week Ahead

A Brief Look Back

The following are economic reports that moved the markets last week. The numbers refer to the close-to-close move in the Dow industrials.

For the week...

Economic Reports

The following information is derived from yahoo!finance and sometimes Bloomberg.com with times local to the east coast.

Options Expiration

The following is courtesy of yahoo!finance and they stole it from The Options Industry Council.

Many options expire this week, so traders will be looking to close out their positions ahead of that, and that suggests increased volatility (large daily price swings).

CPI: Chart Pattern Indicator

As of 03/12/2010, the CPI had:

The chart pattern indicator is bullish with 2 of 3 full triangles showing (). Additional triangles are a measure of strength with solid triangles meaning a more reliable signal than half triangles.

Other

Earnings season is over.

The following industries, of 52 that I follow, were the best (1) and worst (52) performing.

My Prediction

The only index that shows anything of real interest this week is the transports. I picture it on the weekly scale.

The horizontal red line I drew from the most recent date backward in time to see where it lines up with prior peaks and valleys. It gives you an indication of where overhead resistance is hiding. Two shoulders of a head-and-shoulders bottom chart pattern I show at A and B. Those valleys mark overhead resistance as well as other valleys going back further in time (some not so close by).

I drew a blue trendline connecting more recent peaks. The junction of these two lines could spell a turn in the transports. After 5 straight weeks of consecutively higher peaks, you would think the index would want a rest.

What you might see forming in the coming weeks is a head-and-shoulders top. That's what I show as my guess to the right of the green line. Since this is on the weekly scale, my prediction is not for this week but for several months into the future. This week or next, I expect the transports to make another high, but to turn soon.

-- Thomas Bulkowski

Thursday 3/11/10. Escape A Bear Market. Here's How

Pictured above is a line chart of monthly closing prices of the S&P 500 index along with a 12-month moving average of those closes (shown in red).

Here are the trading rules.

1. Buy into the market when the S&P 500 index rises above the 12-month simple moving average of closing prices.
2. Sell when the index drops below the moving average.

Notice that during the start of the 2000 to 2002 bear market, the index drops below the moving average at A. This is a signal to sell and move into cash. In the 2007 to 2009 bear market, the index also drops below the moving average (at B). In both cases, the index remains below the moving average until recovery begins at C and D.

If you were to use the 10-month moving average instead of the 12, price would pierce the average in the circle and also along the CB move at the first touch. Those would have caused an unnecessary transaction (buy then sell, or the reverse), so a 12-month SMA works better. The slightly longer SMA will get you back into the market slightly later at C and D than would the 10-month SMA.

If you were to test this, make sure you use the monthly closing price and not the highs or lows intra-month. You'll find that the moving average reduces draw down and risk over buy-and-hold.

-- Thomas Bulkowski

Tuesday 3/9/10. Tutorial Tuesday: The Cloudbank Trading Setup

I changed a few of the statistics on the cloudbank page and then did more research for a setup based on cloudbank chart patterns. You can read about the setup by clicking on the link, but be sure to read both pages for a complete picture of the cloudbank pattern and then change the rules to suit your taste.

Here's a portion from the setup page...

The cloudbank setup has these rules. Reference the ideal cloudbank setup on the right (not drawn to scale) as I discuss the rules.

1. Find a cloudbank. Look for a cloudbank chart pattern on the weekly or monthly scale. The cloud should be long, often years in length, with a flat base that touches a horizontal or nearly horizontal trendline multiple times. The base of the cloudbank represents the fair value of the stock. It's where the stock would be trading if the economy or other factors didn't pull it down. By the length of the cloudbank, it should be clear to all that the cloud is where price should be trading and the drop after the cloudbank ends is just an anomaly.
2. Steep drop. Look for price to tumble, preferably in a straight-line decline but this is optional. The stock has to make a significant decline such that it represents not only a good value, but a steal. The decline is often the result of a bear market with problems common to many companies and not just this one. In fact, when selecting a cloudbank, look for a company that maintains its price until near the end of the bear market.
3. 30-week SMA crossover. Wait for price to touch or rise above the 30-week simple moving average. I have found that the simple moving average hugs price better than the exponential, so use the simple moving average. The 30-week is the same as the 150-day moving average. Other moving average values might work better, but I tested it with the 30-week. I show the crossover on the picture as Buy.
4. Profit potential. Measure from the base of the cloud (A) to the crossover price (the price where the stock rises above the 30-week moving average, shown as Buy on the chart). If the drop is less than 50% (such as the cloudbank at 10 and the crossover at 7), consider looking elsewhere for a more promising situation. Often if you wait, the stock will drop to a lower price, so sit back and wait if the profit potential is not high enough. In the 2007 to 2009 bear market, many stocks dropped more than 50%. The next bear market may not show such declines, so be flexible. This rule is to help minimize the paper loss when price resumes dropping if you buy too close to the cloudbank.
5. Sell. When the stock rises to the bottom of the cloudbank (A), sell. Why? Because once price enters the cloud, the rise slows dramatically. The rise from the lowest low (B) to the cloud base (A) will take about a year, on average, but so will the move from the bottom of the cloud (A) to the top (C) while covering about half the distance (129% rise to the cloud in 1.1 years vs 64% in 1.0 years in the cloud). Thus, momentum drops once price enters the cloud. If you can find another cloudbank forming, you'll have the opportunity to make more money quicker than holding while the stock traverses the cloud.

-- Thomas Bulkowski

Monday 03/08/2010. Market Monday: The Week Ahead

A Brief Look Back

The following are economic reports that moved the markets last week. The numbers refer to the close-to-close move in the Dow industrials.

For the week...

Economic Reports

The following information is derived from yahoo!finance and sometimes Bloomberg.com with times local to the east coast.

No options expire this week.

CPI: Chart Pattern Indicator

As of 03/05/2010, the CPI had:

The chart pattern indicator is bullish with 2 of 3 full triangles showing (). Additional triangles are a measure of strength with solid triangles meaning a more reliable signal than half triangles.

Other

Earnings season is over.

The following industries, of 52 that I follow, were the best (1) and worst (52) performing.

My Prediction

I show the Nasdaq composite (^IXIC) on the daily scale. Although price has been trending higher in a straight-line run up since D, I think it's time for the index to pause. My guess is that the gap up on Friday (gap 2) is an exhaustion gap. That follows the breakaway gap which I show as gap 1. Breakaway and exhaustion gaps are like twins, they come in pairs, not always, mind you, but if you see a second gap appear it's more likely to be an exhaustion gap than a continuation gap.

Point E also matches the congestion zone I show circled in red at A. That means overhead resistance.

There is also another chart pattern, shown as BCDE. It's a measured move up. The first leg, BC is supposed to match the length and duration of the second leg DE. After the measured move completes, price often retraces into the corrective phase CD. I don't know that the index will drop that far but for the coming week, I see it forming a consolidation region (the index moves sideways) before Friday's retail sales report announces that sales were worse than expected due to the winter weather. Everyone expects this, so maybe the markets will move up instead. I see the index easing lower.

I can be wrong about all of this. The Nasdaq has shown strength where the other indexes have not, so the index could continue to rise without pause. The Nasdaq was up 3.9% last week, beating the Dow industrial's 2.3% advance

-- Thomas Bulkowski

Thursday 3/4/10. Cloudbanks: A New Chart Pattern!

I discovered and cataloged a new chart pattern, which I call a cloudbank. Instead of discussing it here, since it's late Wednesday evening, I'll let you click on the link and discover how you can make big bucks by combining a 30-week moving average with overhead resistance. I'll be adding a new trading setup based on the cloudbank that will clarify how to invest/trade using this pattern, perhaps as soon as tomorrow.

-- Thomas Bulkowski

Tuesday 03/02/10. Market Monday on Tuesday: The Week Ahead

It took a day and a half to get my computer up and running again. The problem was that during boot, the system would reboot as if some Windows file was corrupted. System restore worked on Saturday but all restore points disappeared on Sunday. I had to reinstall Windows and suffer the time needed to upgrade to service packs 2 and 3 and a gazillion other files. Fortunately, I have a fiber optic connection to the internet to download the 150+ megabytes but the installation took lots of time. It seems to be an exercise I go through every year or two. Makes me yearn for a Mac.

A Brief Look Back

The following are economic reports that moved the markets last week. The numbers refer to the close-to-close move in the Dow industrials.

For the week...

Economic Reports

The following information is derived from yahoo!finance and sometimes Bloomberg.com with times local to the east coast.

Options Expiration

No options expire this week.

CPI: Chart Pattern Indicator

As of 02/26/2010, the CPI had:

The chart pattern indicator is bullish with 2 of 3 full triangles showing (). Additional triangles are a measure of strength with solid triangles meaning a more reliable signal than half triangles.

Other

Earnings season is over but it seems reports are still coming in.

I found 1 pipe bottoms last week, which is neutral. Large numbers of pipe bottoms often signal the start of a short to intermediate-term move up before price curls back down, forming an unconfirmed double bottom. I think of it as the signal for the first bottom of a potential double bottom.

The following industries, of 52 that I follow, were the best (1) and worst (52) performing.

My Prediction

I show a picture of the Dow industrials on the daily scale. The index has completed an simple ABC correction, which I show as A B C. The correction at B doesn't appear clearly on this picture, but it represents a pause between the two down legs, A and C.

If you are into Elliott wave, and I am not, then you will expect a resumption of the primary wave that comes as three upward moves and two corrections. I show the first three of those as 1, 2, and 3. Waves 1 and 3 take the index higher and wave 2 sucks it down.

If you like chart patterns, as I do, then you will recognize the measured move up chart pattern, also labeled 1, 2, and 3. Legs 1 and 3 power the index higher and the corrective phase, 2, takes it lower.

In other words, I think the upward move today (Monday, as I write this), suggests a resumption of the upward move in the index. If Elliott wavers are correct, then wave 3 should be the longest of the bunch. I expect the week will see the Dow higher at week's end than it was last Friday.

-- Thomas Bulkowski