Registered User
Joined: 4/21/2009
Posts: 24
|
An interesting article on TASC that talks about this. What is more surprising is that it was written by Thomas Bulkowski of whom I have read quite a bit on this board.
Well it seems that Bulkowski himself is contradicting what it has been said on some threads about upward patterns in a Bear market or viceversa.
Reading the article will enlighten you to my remarks, I am not going to go into detail here - but it is certainly clear in the article.
|
|
Registered User
Joined: 12/31/2005
Posts: 266
|
someone once said, "the trend is your friend."
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
well, if chart patterns don't work, would someone kindly explain to me how I have been doing so well?
Yes, of course, modifications must be made for varying conditions, but I believe chart patterns still work.
|
|
Registered User
Joined: 12/2/2004
Posts: 1,775
|
tobydad is dead-on right imo. My best picks broke out from well defined patterns, most recently TNDM, LFT, ARST. So absolutely chart patterns still work. But don't expect any bull pattern to work as well if markets are in a downtrend. Was just listening to the William O'Neil in a recent interview in which his team has just finished another indepth round of studying chart patterns, and he says it was the same chart patterns that worked best in the past year despite the bear market. This guy pulls up the most successful stocks for a one year period and studies their patterns in amazing detail and how their setups evolved before the big moves. He's done this literally every year for about fifty years and publishes them in detail. Probably most of the patterns fail at breakout, but that's not the point; the point is that the occasional breakouts that do lead to the really biggest gains within a year emerge from specific patterns that haven't changed in half a century. With strict risk management the patterns that fail are quickly sold for minimal loss (BEAT for example), while the ones that work are quickly added to within a relatively tight percentage range, and then sold only when the chart pattern gives a sell signal per specific rules.
I have read Bulkowski and imo his chart patterns are excessive in number. He's got like a hundred patterns, with many being of the short term variety, so maybe they have merit for very short term traders. For example I like his short flag pattern breakout, especially as an add-on signal. O'Neil on the other hand has only about 7 patterns that he says have consistently stood the test of time. Those being the flat base, cup without handle, cup with handle, W bottom with handle, ascending base, three-weeks-tight, saucer with handle. He's really big on the handle as it's a very telling marker for the last shakeout of weak hands. I think it's best to find three or four patterns that a trader becomes very familiar with and that dovetail with one's personality and comfort level.
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
Strange stuff.
I haven't seen the article yet.
Just the teaser on the website.
On his "upward breakout" patterns.
They stay pretty consistent from 1991 to 2003
(they did their best in 2003 which makes sense.
The start of the bull market)
Breakouts failed more in the 2000/2002 bear.
(which makes sense)
Then the upward breakouts fail more from 2004 to 2007.
(you would think they would do well in the bull market,
this appears to be the worst period since 1991)
Then they get better in 2008.
(you would think they would do worse in a bear market)
Thanks
diceman
|
|
Registered User
Joined: 8/15/2006
Posts: 132
|
I find any method of trading can be good or bad depending on the characteristics you adhere to. If you maintain objective rules within your method you can measure the effectiveness and modify any weak areas. This even goes for patterns, if you trade subjectively then you have no true measure and you could trade a pattern different than others so some will say it worked and some will say it did not. IMO people who write books and sell trading material are the last people you should listen to, they may have great ideas but that does not mean they can trade.
Scott
|
|
Registered User
Joined: 1/28/2005
Posts: 6,049
|
I think you guys are mischaracterizing Bulkowski.
I have never cared too much for patterns so paid little
attention to Bulkowski over the years. However I don't
believe the stuff said here is not true. This is what I
understand the situation to be:
(I could be wrong)
1) After his initial pattern research he retired from his
employment(some type of software work) due to trading
profits. So I think he knows something about trading.
2) He was frustrated no info was available on patterns. So he
decided to become a statistician of the markets. He defines patterns
and tells you their batting average.
(He doesn't care if they offer poor results or not)
3) He has 100's of patterns but it is not because he says they are
good, he follows 100's to be complete.
(someone will tell him he's missed one)
4) All his pattern statistics as well as software to find them.
Are available for free on his website.
(its hard to believe he is out to make money on this)
Think about it. If you were selling pattern information. Would you
bring up why they don't work?
Thanks
diceman
|
|
Registered User
Joined: 2/5/2006
Posts: 1,148
|
looking at the spy daily over the last 2 years, virtually every major turning point or continuaton, consists of a pennant, flag, h & s, double top or bottom, or wedge.
the trick is recognizing underlying technical strength, and a valid breakout.
|
|
Registered User
Joined: 2/5/2006
Posts: 1,148
|
i read the intro on s & c's website.
he says his research shows a higher failure rate for upward breakouts in chart patterns. i suspect that might be because the markets been in a secular bear for the last 9 years.
|
|
Registered User
Joined: 6/6/2005
Posts: 1,157
|
I just finnished the article. Bulkowski says that according to his research that patterns fail up tp 2x as much as they used to when traded with targets. He does studies on 10% and 20% target gains. He also says to trade inline with the trend -- i.e. upward breakouts in "bull markets" and downward breaks in "bear markets". -- but we don't need to go into that again. I don't know about anyone else but I personally haven't noticed a 2x increase in pattern failures. like funnymony pointed out -- the major turning points in the markets have been accompanied by recognizable (and IMO tradeable) patterns.
David John Hall
|
|
Registered User
Joined: 4/21/2009
Posts: 24
|
QUOTE (davidjohnhall) I just finnished the article. Bulkowski says that according to his research that patterns fail up tp 2x as much as they used to when traded with targets. He does studies on 10% and 20% target gains. He also says to trade inline with the trend -- i.e. upward breakouts in "bull markets" and downward breaks in "bear markets". -- but we don't need to go into that again. I don't know about anyone else but I personally haven't noticed a 2x increase in pattern failures. like funnymony pointed out -- the major turning points in the markets have been accompanied by recognizable (and IMO tradeable) patterns.
David John Hall
Doesn't that contradict what many said here before in regards to Bulkowski and trading against the grain (upward breakouts on a down trend)?
|
|
Registered User
Joined: 6/6/2005
Posts: 1,157
|
Maybe he decided to run a new test after we emailed him as he said "I can't believe that my research says that's correct" and then he got those results. I don't know.
On the same note: I just closed a trade in RAD for a 129% gain -- my highest % gain ever trade today in the middle of this bear market.
I really have no idea what to think any more and trade based on what my tests and my own trading show. If i feel a stock is rallying and I want to trade it, I trade it long. If I think the market is week then i trade an inverse etf.
Correlation exists, I know that. You can go long or short in any market but a majority of stocks correllate to the index direction, I don't think you can argue that. But i would rather take a 100% bounce than a 50% decline. But that's my own preference. But that's also not saying I can't do that while trading the major indexes or ETFs short based on a different system's parameters.
I guess it's all based on what you're comfportable and what works for you.
David John Hall
|
|
Registered User
Joined: 12/31/2005
Posts: 2,499
|
Well if the pattern is consistantly not working then just reverse the signal!
Just kidding... or am I?
Using chart patterns to find stocks is kind of like seperating the chaff from the wheat. During turbulent market conditions the separation is not as good, just as on blustery days the chaff doesn't separate from the wheat a much.
|
|
Registered User
Joined: 1/30/2009
Posts: 267
|
QUOTE (jas0501) Well if the pattern is consistantly not working then just reverse the signal!
Just kidding... or am I?
Using chart patterns to find stocks is kind of like seperating the chaff from the wheat. During turbulent market conditions the separation is not as good, just as on blustery days the chaff doesn't separate from the wheat a much.
And you have to grow wheat while the sun is shining.
|
|
Registered User
Joined: 3/6/2009
Posts: 78
|
QUOTE (jas0501) Well if the pattern is consistantly not working then just reverse the signal!
Just kidding... or am I?
If only it was that simple. The reason it is hard to make $ on trades is because of the overhead. Don't forget that your broker needs to get paid (we call that "commissions"), and so do the floor traders, market makers, and ECN operators who execute your trade (we call it "slippage").
|
|
Registered User
Joined: 12/31/2005
Posts: 2,499
|
Moving average crossovers is a chart pattern....
If you are looking for the "best" chart pattern for today's market conditions....
you'll have to wait a year or two and then backtest the heck out of the data to find it. Then you can sell the system pointing to the excellent performance of the system!
Hindsight is a wonderful thing.... It can convince you of the power of any approach given the right selection of data points. The problem I find in most articles in S&C is the limited set of data used. Both time spen tested and stocks tested. Too much curve fitting risk. Also little is ever mentioned of the performance of the article's approach after changing the parameters in small ways.
If you want to know the "truth" of any system backtest it yourself. Stockfinder is progressing toward becoming a very capable tool for answering this type of question suplying far better answers than supplied in this or any forum.
As already stated...
Think for yourself.
It is not easy and takes some work and some education but that's the fun part.
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
Of course, chart patterns work! Chart patterns are simply graphs of the feelings and emotions of people as they think about the markets. Because we are created beings, the basic nature of man does not change. So man will respond differently to different market conditions but it is all still based on the same human inclinations and desires. The basics do not change.
This is why a careful study of market turns is so important. Companies with decent fundamentals and reasons for growth (perceived or real) bounce well during mini-uptrends. Yes, the trend is your friend, but one can move a little more quickly as the trend makes small moves up during an overall down trend.
But, again, chart patterns definitely work and they always will. The real question is, do you work? Or are you wanting to prop your feet up and rely only upon what "has always worked".
Blessings all!
|
|
Registered User
Joined: 10/27/2004
Posts: 821
Location: Philly area
|
Two Sidebars:
* Bulkowski has a splendid free website with an ocean of pattern information and a pattern matching software program for free (it finds them).
* I have seen some commentary that some patterns are less useful than years ago (like when no-body knew about them). When a pattern is well known too many people jump onto it, perhaps spoiling the spoils to be had.
A general commentary is that NYSE stocks seem to be less good with pattern responses than Nasdaq, largely due to who trades on the two exchanges. That is more TA types are to be found on NASDAQ.
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
Bigblock;
nice to see you back, we've missed you...it's always so quiet around here without you.
|
|
Registered User
Joined: 2/21/2007
Posts: 797
|
QBB is BB? If so then i quess APS will show up soon. Tobydad, you shouldn't have let the cat out of the bag.
|
|
Registered User
Joined: 2/5/2006
Posts: 1,148
|
QUOTE (thyself) So has anyone really discover if the patterns work or not?
does this help?
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Chart patterns work in the sense that they give one a point of reference. You can play the break from said pattern one direction or another. Drop your anchor and play the probabilities. Patterns are a product of the trading sentiment (I believe that my friend Tobydad has spoken about this before) It is also what I believe.
The fact is that patterns fail. I like them because they let you know that the probability for a change is on the cusp. If managed properly you can make great profits from the patterns one direction or the other...
|
|
Registered User
Joined: 1/12/2009
Posts: 235
|
QUOTE (Apsll) Chart patterns work in the sense that they give one a point of reference. You can play the break from said pattern one direction or another. Drop your anchor and play the probabilities. Patterns are a product of the trading sentiment (I believe that my friend Tobydad has spoken about this before) It is also what I believe.
The fact is that patterns fail. I like them because they let you know that the probability for a change is on the cusp. If managed properly you can make great profits from the patterns one direction or the other...
Apsll I thought you left and were in the process of designing your own charting software and trading methods? How is that going?
|
|
Registered User
Joined: 6/6/2005
Posts: 1,157
|
Hi thyself,
I have just returned from a long journey to the nether reaches of the stock market. It was tiring and endless. I crawled through the valleys of the indicators, scaled the great volcano "Equity Curve" and have battled the evil Market Makers of the Nasdaq region. What a heroic quest it was! Luckily, it was not in vein because I have returned with an answer from the trading Gods up on Mount Dan Zanger. They have told me with 100% certainty that YES -- trading patterns work. They also told me something else interesting: that other methods also work.
To repeat: trading patterns...and other methods...work.
Now we can all go forth and prosper! Using trading patterns...or other methods.
:)
David John Hall
|
|
Registered User
Joined: 3/6/2009
Posts: 78
|
Of course chart patterns work. The group of traders that I belong to rely completely on patterns. Some patterns work better than others in different kinds of markets and in order to keep abreast of the changing markets and use the most profitable pattern in each market, we continually back test as the market changes. Some patterns seem to work well in certain markets and require very few changes.
Chart patterns are our bread and butter, so my vote is yes.
|
|
Registered User
Joined: 2/5/2006
Posts: 1,148
|
the trick is really in pattern confirmation, and getting good entries.
|
|
Registered User
Joined: 3/6/2009
Posts: 78
|
That's true funnymony. I always try to capture the pattern with the best price action that can obtain very high win ratios. For a long confirmation, I like to see a higher swing low and use the back test results of the number of bars since the higher swing low as an entry.
|
|
Registered User
Joined: 3/21/2006
Posts: 4,308
|
Traderm30; I did leave and the charting software that my friends are designing did not compare to the Wordens fine product. I am so used to TeleChart that I find that I cannot live without it. I have been swing trading as of late and not day trading. With the new shift in the markets I might just switch back over to Platinum so that I will day trade again, but we will see about that.
I have been busy with my real-estate ventures and starting a new business and that is why I have been recently absent. I love this forum and have been a long time member so it is hard to leave my friends. With the exit of certain folks from the forum I think that I can tone down my strong opinions and return to the "Apsll" that used to post a couple of years ago when this forum was fun.
I do not know how long that you have been trading but I hope that you and I will be able to share some good ideas together going forward. I am still busy with other things and my time will be limited here on the forum, but I will be around.....
Apsll.
|
|
Registered User
Joined: 12/8/2004
Posts: 15
|
Dryfess you said
“……we continually back test as the market changes. Some patterns seem to work well in certain markets and require very few changes.”
I’m always open to new ideas and have a few questions for you.
1. How far back do you backtest? Assuming there is a higher low three bars ago, and also at 10 bars and 30 bars ago. In other words days-weeks-months-years???
2. Do you use other indicators for verification?
3. Does volume effect your decision?
4. What patterns do you look for?
Perhaps you could post a typical chart and give your views. It would be most helpful even if it does not apply to the current market.
Although giving a link to your trading friends blog may not be allowed by the moderators. You could give it a try in separate post.
Thank you for joining the Worden group and any help you care to contribute.
Jack
|
|
Registered User
Joined: 3/6/2009
Posts: 78
|
Hi Jack, thanks for the welcome
I back test as far back as necessary depending on the time frame that I am using. At the present time I am trading the ES on a 1 minute time frame and I will back test 6 months prior and 1 month walk forward. If tick data was used to trade then only a few weeks would be used and if daily was used for trades, then several years of data would be used.
The pattern I mentioned above as requiring very few changes on a higher swing low is using price action without any indicators. The entry confirmation is a set number of bars after the higher swing low. This pattern also works very well on stocks with high volume that has well defined peaks and troughs. The most profitable is the beginning of a 3rd wave. To rough guess at how many bars since the higher swing low to use as a confirmation that a 3rd wave is verified, look at past price action of the 2nd wave and determine how many bars it took before the 3rd wave was under way. You should see a few head fakes, add 1 or 2 bars to that and you have a rough estimate. If you are day trading and watching a heavily traded stock you might see higher volume on that bar because most of the professional day traders will enter on the 3rd wave. Reverse everything I have said for shorts.
I have used several patterns that are profitable such as trend line breaks with volume as confirmation but have been kind of stuck on futures on a short time frame because with too many patterns the program can sometimes reach several thousand lines of code. Thanks to belonging to a group of talented traders, we put our heads together for the best ideas. Some of the guy’s high frequency trade and most of us use automated trading to eliminate human error and stress. As a matter of fact, one the guys used to post on Worden but I see he has not been active lately. That's how I heard of Worden.
I hope this answers your questions. I hastily wrote this and apologize for any errors as I wish to have dinner before the Lakers ball game
|
|
Registered User
Joined: 12/8/2004
Posts: 15
|
Dryfess. Thanks much for the prompt reply. Now I see where you are coming from. No charts needed, your explanation and trading style is crystal clear. Everything you said makes sense.
I never did ES contracts, but still have access to level II. You have piqued my interest once again. Will fire-up my old candle and stochastic program and take another look Tuesday. It must be nice to have all that software. Congrats to you and your group for developing it. I believe we have some day-traders here and you will be getting more than a few questions.
Incidentally, don’t worry about the bigblock thing. Most of us are in bb overload to varying degrees.
Welcome aboard again Dryfess and thanks for offering to help. Actually, this post was started about five hours ago but was continually interrupted by “Honey, how do you? Could you show me?” If I day-trade again I’m going to sound proof this room and put a lock on the door.
Go Lakers!
Jack
|
|
Registered User
Joined: 10/7/2004
Posts: 2,181
|
"If I have ever made any valuable discoveries, it has been owing more to patient attention than to any other talent."
Isaac Newton (1642-1727); English mathematician, astronomer and philosopher
|
|
Registered User
Joined: 3/6/2009
Posts: 78
|
Hi Jack, I'm glad that you understand the pattern. It is not really new, nor did we develop it. Trading the 3rd wave has been around for a long time and professional traders have used it for a long time as evident by the many articles and books that have been written on the subject. All we did was refining the entry timing and identify the most profitable securities to trade along with the best time frame. The hardest part was writing the code to back test and writing the auto trading code. You might be shocked to know the actual cost of "all that software" is very reasonable and consists of only one package. Good software does not have to be expensive. The best software is between the ears.
Actually Telechart is very capable of doing the same thing albeit different methods must be used. Considering the lack of back testing, the tedious task of finding a basket of stocks to trade with the best defined peaks and troughs could be done by eye. One thing to note is just because the waves do not look favorable on a daily time frame; take a look at smaller time frames. If you wish to trade on a 5 min time frame, then choose your stocks on the same time frame.
The well known method called LBR's 1st cross trade could be used to find the 3rd wave. It uses indicators only and with some clever programming from the Worden trainers, a scan could be written to uncover this pattern and the eyeball can then be used to verify the entry. The indicator used is the MACD with settings of 3-10-16.
Here is the sequence of events as I understand them. This is for longs and shorts.
1. The 16 line crosses zero (1st cross) alerting to a trend change. The 16 line must remain on this side.
2. The 3-10 histogram crosses to the opposite side of the zero line, which coincides to the first pullback after the trend change.
3. The 3-10 histogram then turns back to cross the zero line again. This tells me the pull back might be exhausted and the trend change is ready to continue.
The old axiom that indicators are lagging and cannot be reliable does not apply here as this method depends on the lag.
Oh, btw, the Lakers are not my favorite team. I just enjoy watching talent and the Lakers have plenty of that as does Cleveland.
Tobydad, is that a question or statement?
|
|
|
Guest-1 |