Pattern testing

[cr6196]

I am trying to some statistical testing of the effectiveness of patterns in predicting future movements and having never done anything like this i am seriously struggling with defining each of the patterns.

Has anyone done this sort of testing? How would you define a double top/double bottom or a flag, in particular? I have a broad definition for both of them but they need refining quite badly, i got about 15/20 signals into trading when i decided that my definitions were causing serious problems, particularly with DTs/DBs.

Any help really on the subject of testing pattterns would appreicated.

Cheers

 

[danpayne]

I am trying to some statistical testing of the effectiveness of patterns in predicting future movements and having never done anything like this i am seriously struggling with defining each of the patterns.

Has anyone done this sort of testing? How would you define a double top/double bottom or a flag, in particular? I have a broad definition for both of them but they need refining quite badly, i got about 15/20 signals into trading when i decided that my definitions were causing serious problems, particularly with DTs/DBs.

Any help really on the subject of testing pattterns would appreicated.

Cheers

I would strongly recomend a book called The encyclopedia of chart patterns by Bulkowlski this book will answer all questions as far as im concerned regarding chart patterns including more statistics than you will know what to do with
not a cheap book book but a good investment long term
hope that helps
Dan

 

[megamuel]

...and his website The Pattern Site.com

 

[cr6196]

thanks, just flicking through it now, looks quite good, still need to think through how i am going to apply all this and get something useful out of it tho lol.

Thanks for the help

 

[danpayne]

thanks, just flicking through it now, looks quite good, still need to think through how i am going to apply all this and get something useful out of it tho lol.

Thanks for the help

the website is good but the book is more comprhensive I thought the same until I started using the book

 

[HaloTrader]

One of the things i've realised about chart patterns is that the conventional way to trade this isn't the only way. Its all about trader personality. I don't buy the breakout of a flag. I buy the bottom of the flag support level and sell at the top after a reversal pattern. This ensures i am in all the breakouts and in a no-risk trade. Because if the breakout were to fail and come back, i'd still be a profitable position to exit. So if a flag has a range. I am shorting the top after a defined reversal bottom and buying the low after a defined reversal pattern... When price breakouts of, i have already accumulated more that my risk through trading the risk and i catch the breakout, if it is a high-test bar, i am in from the low and therefore still profitable.

 

[new_trader]

One of the things i've realised about chart patterns is that the conventional way to trade this isn't the only way. Its all about trader personality. I don't buy the breakout of a flag. I buy the bottom of the flag support level and sell at the top after a reversal pattern. This ensures i am in all the breakouts and in a no-risk trade. Because if the breakout were to fail and come back, i'd still be a profitable position to exit. So if a flag has a range. I am shorting the top after a defined reversal bottom and buying the low after a defined reversal pattern... When price breakouts of, i have already accumulated more that my risk through trading the risk and i catch the breakout, if it is a high-test bar, i am in from the low and therefore still profitable.

What the..?

 

[new_trader]

I am trying to some statistical testing of the effectiveness of patterns in predicting future movements and having never done anything like this i am seriously struggling with defining each of the patterns.

Has anyone done this sort of testing? How would you define a double top/double bottom or a flag, in particular? I have a broad definition for both of them but they need refining quite badly, i got about 15/20 signals into trading when i decided that my definitions were causing serious problems, particularly with DTs/DBs.

Any help really on the subject of testing pattterns would appreicated.

Cheers

Don't waste your time...That 'Bulkowlski' book is fine for a door stop and that's about it.

 

[danpayne]

Don't waste your time...That 'Bulkowlski' book is fine for a door stop and that's about it.

So you know the book but you never read it is that right :?:

 

[cr6196]

Right as i said i have flicked through the book it is useful but it has its limitations...it is very thorough in going through the patterns and how they can be strictly defined so it is useful if your looking to build up stats on the patterns and what differences to look for between each setup...it is limited in that it refers only to us stocks, i trade forex but the core message is still useful.

If you think patterns can show future price direction, it could be useful...but each to his own.

To halotrader, i have been traded some of these patterns b4 and can say that trying to play flags that way is problematic, ultimately your scrabbling around for that extra few points and take a larger loss than expected if the pattern fails, also your prone to get caught switching about if it does a 1,2,3 type reversal against the continuation side of the flag. Just my opinion though.

 

[new_trader]

So you know the book but you never read it is that right :?:

No, I bought the book and I am trying to prevent people from making the same mistake. It is not the way to go about things.

 

[superspurs]

ffs nt, enlighten us or STFU

 

[Technically Fundamental]

trading patterns is like having a foreign alphabet and learning to write a language but not having a clue about the phonetics. I believe I will become infinitely more knowledgable in my trading if I learn more about sentiment and PA behind patterns etc.

 

[danpayne]

No, I bought the book and I am trying to prevent people from making the same mistake. It is not the way to go about things.

evening
CR6196 Is looking for a book that involves chart statistics , then in my opinion thats probably the book, Im not saying anyone should base there whole trading of patterns on that book but if its statixtics your after, and cr6196 is then thats the book there is probably others but none ive read
Or n
maybe send him your copy if you dont use it lol

 

[HaloTrader]

What the..?

Oppps, i should really read what i write to ensure it makes sense. I can't be bothered to explain in this instance.

 

[new_trader]

evening
CR6196 Is looking for a book that involves chart statistics , then in my opinion thats probably the book, Im not saying anyone should base there whole trading of patterns on that book but if its statixtics your after, and cr6196 is then thats the book there is probably others but none ive read
Or n
maybe send him your copy if you dont use it lol

Ok, good point, if it is statistics you want. Let me try to illustrate what I am saying using the weather as an analogy. Person 'A' might look outside and see that it is sunny now just like it has been for the last 3 days. Then, in order to determine if they should have a BBQ tomorrow they look up weather statistics and find that in June there is a 57% chance of rain on the 5th day whenever there have been 4 sunny days in a row.

Person 'B' has a barometer, hygrometer and thermometer and has actually made their own weather observations for the last 2-3 years and kept a careful record. They understand what cloud formations are like when it is about to rain and what conditions were like leading up to it. They were planning a BBQ for the same day and from experience they know that it is highly likely that it will rain heavily tomorrow.

Now, you want to plan your own BBQ but you haven't got a clue about the weather, so you ask person 'A' and they say "From my experience of studying statistics I say that there is a 57% chance of rain tomorrow", you ask person 'B' and they say "From my experience of studying the weather I say that it is highly likely that it will rain heavily tomorrow".

Which forecast would you use to make your decision?

 

[alexander]

Micro-patterns can be more effective than chart patterns in some ways but you have to figure out how to use them properly. This company has done a lot of work in this area:

Discover trading systems automatically with APS Automatic Pattern Search trading software

 

[Technically Fundamental]

Ok, good point, if it is statistics you want. Let me try to illustrate what I am saying using the weather as an analogy. Person 'A' might look outside and see that it is sunny now just like it has been for the last 3 days. Then, in order to determine if they should have a BBQ tomorrow they look up weather statistics and find that in June there is a 57% chance of rain on the 5th day whenever there have been 4 sunny days in a row.

Person 'B' has a barometer, hygrometer and thermometer and has actually made their own weather observations for the last 2-3 years and kept a careful record. They understand what cloud formations are like when it is about to rain and what conditions were like leading up to it. They were planning a BBQ for the same day and from experience they know that it is highly likely that it will rain heavily tomorrow.

Now, you want to plan your own BBQ but you haven't got a clue about the weather, so you ask person 'A' and they say "From my experience of studying statistics I say that there is a 57% chance of rain tomorrow", you ask person 'B' and they say "From my experience of studying the weather I say that it is highly likely that it will rain heavily tomorrow".

Which forecast would you use to make your decision?

Thats what I said.

 

[Technically Fundamental]

Micro-patterns can be more effective than chart patterns in some ways but you have to figure out how to use them properly. This company has done a lot of work in this area:

Discover trading systems automatically with APS Automatic Pattern Search trading software

Thanks for the info. My profits should now increase by 3000%

 

[Jason101]

Thats what I said.

lol , its funny because it's true.

I was impressed with NTs' word smith's reply until I read this, then spat my beer out with laughter.

I gotta get out more.