Thomas Bulkowski’s Pattern

Atilla said:
To be honest I have both his books and strongly recommend purchase.

Lots of stats on % probabilities of how these repeat patterns perform.

Not for FA traders. TA's only.
Click to expand...

Nothing wrong with him or his patterns. I haven't read his books but I have read a lot on the internet by him about chart patterns. Some good stuff indeed.

Peter
 
wackypete2 said:
Nothing wrong with him or his patterns. I haven't read his books but I have read a lot on the internet by him about chart patterns. Some good stuff indeed.

Peter
Click to expand...


Denial is first sign of truth imho... :cheesy:
 
I wonder how he got his stats. I have tried to program chart patterns in the past and it was extremely difficult. You have to make many assumptions in the code and recognition becomes kind of subjective. You tweak one variable or two and you have no pattern or you have a pattern. Do you simply take his stats for granted?
 
I haven't been to the website in a while. I wouldn't take the stats for granted but the info is good to use as a starting point if you want to trade patterns. Pick 3 or 4 that appear to perform best (he ranks them but use your own judgement) and try them out for yourself.

I agree that trying to code an ea to identify patterns is a nightmare. You can try using pattern recognition software from autochartist.com. They had a free trial period some time back, maybe they still do. There are a few others too. I've tried in on forex and I can tell you it does a very good job of identifying patterns you specify with some parameters. How you trade them is up to you though.

Peter
 
Albert Carey said:
I wonder how he got his stats. I have tried to program chart patterns in the past and it was extremely difficult. You have to make many assumptions in the code and recognition becomes kind of subjective. You tweak one variable or two and you have no pattern or you have a pattern. Do you simply take his stats for granted?
Click to expand...



By studying 10 charts and mulitplying by 10 to get 100 charts.

There is madness behind the stats. If you drop your cat from a 10 storey block of flats - how far it bounces having hit the ground is likely to depend on how fat your cat is coupled with how warm the weather is as one assumes one would pencil in an updraft factor...

I think you should buy the book.


Good investment imho...
 
I've been creating a candlestick recognition trading program from scratch for 8 months now. The program now has around 23 Meg of logic text, a huge program if you don't know what I mean. It was worth the effort... I based all my programming on visual recognition of the patterns. Candlesticks are everywhere, it is surrounding conditions that determine if you trade them or not.
 
numbertea said:
I've been creating a candlestick recognition trading program from scratch for 8 months now. The program now has around 23 Meg of logic text, a huge program if you don't know what I mean. It was worth the effort... I based all my programming on visual recognition of the patterns. Candlesticks are everywhere, it is surrounding conditions that determine if you trade them or not.
Click to expand...

Good work but huge program. Probably will be slow running. Michael Harris has created a program called Price Action Lab based on an older program of his that can detect high probability price patterns of up to 6 bars. Candlesticks are a very small subset of price patterns. If you can detect price patterns looking at candlesticks is a redundant exercise. His program does search, scan and several other things and it is only 2 megs of executable. It also generates code for TS and other programs. Here is an example from his blog:

QQQ and SPY: Inside Day Pattern After an Outside Day Pattern | Price Action Lab Blog
 
It's truly slow compared with other programs I've seen. I cycle through many conditionals to check for tradability also. I run through the last 250 daily bars, I trade with EOD data, of 2800 stocks in about 2 hours. It is fast enough I guess. If a trade is found it is listed in order of highest probability of being a good one. I would have been able to use a premade framework if I wanted but doing it my own way gives me flexibility that is unequal to anything I've seen out there. I've built it to suit my trading style. I agree that candles are a small part of the total picture but I look for conditions that increase their probability so they are my central signal. I trade reversals so I am always looking at tops and bottoms. Cut losses and let profits run.
 
PriceActionLab calls it 3 day inside outside price pattern. I call it an engulfing candlestick pattern with a bearish confirmation day. Potatoe patato. I find that the standard reversal candlestick patterns show that there is a trend change, not necessarily a reversal and not necessarily immediately. The surrounding price movements must be kept in context for the probability of the pattern to be useful to the stock operator. The formula for identifying the pattern that was given is extremely simplistic. I believe that the position of the bars relative to each other should be more precise or more limiting than their formula allows.
 
numbertea said:
PriceActionLab calls it 3 day inside outside price pattern. I call it an engulfing candlestick pattern with a bearish confirmation day. Potatoe patato. I find that the standard reversal candlestick patterns show that there is a trend change, not necessarily a reversal and not necessarily immediately. The surrounding price movements must be kept in context for the probability of the pattern to be useful to the stock operator. The formula for identifying the pattern that was given is extremely simplistic. I believe that the position of the bars relative to each other should be more precise or more limiting than their formula allows.
Click to expand...

Which formula are you talking about numbertea?
 
Hello Mr. Black,
I was referring to the link to the Price Lab blog above. The forth graphic down the page includes a formula they generated for backtesting an engulfing pattern for AmiBroker. Simple sure is better sometimes but I prefer profit to simplicity.
 
numbertea said:
Hello Mr. Black,
I was referring to the link to the Price Lab blog above. The forth graphic down the page includes a formula they generated for backtesting an engulfing pattern for AmiBroker. Simple sure is better sometimes but I prefer profit to simplicity.
Click to expand...


I don't understand you at all. The formula on that blog is an 1:1 mapping of that pattern.

What is your formula? If you have one then I may understand what you mean. Otherwise I cannot, sorry.
 
I'm using the candlesticks and finding that the visually superior instances of candle patterns have a greater probability of being successful instant reversals than all the instances that fit the general pattern template. I use the basic template as the first test in a series of other tests including the length of upper and lower shadows on the bars and the relative position of one bar to another. For instance, is the top and bottom of the first bar in an engulfing pattern between a certain percentage of second bar difference within the second bar. The extra complication yields higher probability success. I am taking the approach that if I can easily see the pattern on a chart then my program should trade on it, otherwise I would not want to trade it, for me the basic template of the pattern is not enough.
 
Ok folks,
I want get the focus on my first post.
Do anyone coded candle of:
1 candle
2 cancle
3 candle

to vave idea of most pattern on the market?
I suggest to read this:

Candlestick Corner | How Reliable Are Candlesticks? | Working Money

I know all this software that the user write in the threat.
I'd like to built in easylanguage and to be free to all.
SO every one can load indicator on tradestation-multicharts and see the results.

Do that we manage differet setup of buy ans sell.

I suggest a good point to take a lot of pattern already coded here:

http://www.trade2win.com/boards/tradestation/30152-easylanguage-candle-ident.html?highlight=raven
 
You must log in or register to reply here.

Similar threads

NVP
Thomas Bulkowski
Replies
2
Views
1K
NVP
NVP
T
'Patternz'
Replies
2
Views
2K
Trdr
T
R
Encyclopedia of Chart Patterns
Replies
5
Views
5K
Anonymous
A
D
Trading Systems that Work by THOMAS STRIDSMAN
Replies
1
Views
5K
FXCatty
F
S
Thomas DeMark Points
Replies
2
Views
2K
ssstangb0y
S
Top