Automatic Pattern Search

I have upgraded to the p-indicator version and the signals I get are very good. Keep in mind that the p-Indicator is not like your usual indicator. You must set the risk/reward criteria and adhere to those when you take the signals. I also think this indicator is best for position trading. I have not tried it with intraday data. Excellent work by Mr. Harris.
 
Hi Alex,

What markets are you trading ?

Best,

Chris

I currently trade ETFs, very liquid US stocks, energy futures and forex. My trading relies solely on patterns I get from APS. I also started using the p-indicator signals as a confirmation. When I get a signal from a pattern I wait until I get another one. If I get another one within the next two days I wait for an intraday correction and I also check the value of the p-indicator from next day's close. This strategy pays but you must have discipline to follow. APS is the best tool around but to use it properly you must first figure out a few things and this is my secret. :D
 
thanks for this detailled explanation.another question: what type of computer are you using for aps ?
 
thanks for this detailled explanation.another question: what type of computer are you using for aps ?

Desktop with Intel Core i7 running Windows 7.

By the way, I just realized in my previous reply I meant to say that "I also check the value of the p-indicator from previous day's close".
 
Skanker see my previous posts. All of Harris' products and books are worth their money. He is one of the most sensible, practical and knowledgeable trader around. For details on the p-Indicator try this

http://www.tradingpatterns.com/The_p-Indicator.pdf

IMO it is not a good practice trying to illegally purchase software from the internet. You never know to whom you are talking. You may get software that contains sophisticated spyware that it is virtually undetectable and can steal trading passwords, cc numbers and do all sorts of nasty things to your life. Be very careful.
 
What is everyone's thoughts on the p-indicator? Worth the money?

What do you mean by "worth the money"? Any tool out there is as good as its user. A trader can have the best system and still screw up and lose all of his money.

APS is one of the most powerful trading programs I have ever used but at the end of the day it won't help anyone who does not know how to use the power properly. I think I have figured out how to use APS properly and it has worked nicely up to this point. Please do not ask for details because I am not going to say anything over and above what I already have said in other posts.
 
What do you mean by "worth the money"? Any tool out there is as good as its user. A trader can have the best system and still screw up and lose all of his money.

APS is one of the most powerful trading programs I have ever used but at the end of the day it won't help anyone who does not know how to use the power properly. I think I have figured out how to use APS properly and it has worked nicely up to this point. Please do not ask for details because I am not going to say anything over and above what I already have said in other posts.

Nothing can help a trader who does not understand probability theory and risk management. Even the best systems fail when not properly traded. As far as APS, I do not think it is for everyone. Its proper application relies on a good understanding of probability and statistics. I do not think it can help traders who do not understand well these concepts. It is remarkable that this program can find all these patterns but using the output in profitable ways is not as straight forward as some may think.
 

Thanks for the link but for $2k you get nowhere close the power of APS. The code generation seems ad-hoc. From the example on that website

"Code Sample
An example of the EasyLanguage strategy code generated by Builder is shown below.* This code represents a day trading system where the logic for long and short trades is different (Symmetry option set to false). The strategy enters long on a breakout from the prior close, where the breakout amount is a fraction of the average true range (ATR)."

It is an indicator based search that looks for an optimal system after a number of permutations. This is very naive type of search. Strata Search does the same thing a lot better.

Even worse, they do not offer a demo version. I would consider buying it if I could do extensive testing before, like I did with APS.
 
I upgraded to the new version with the p-indicator a month ago. I was not sure about the indicator capabilitites but my profits from my recent three trades covered the cost. I am still not sure what the best way is to use this indicator. For example, in a T/S file, do you input just one line of parameters or multiple lines, what is better?. Which is the best target/stop to exit the actual position when multiple lines are present in a T/S file? I use the smallest value of the target to exit the position but maybe using the average makes more sense.
 
I upgraded to the new version with the p-indicator a month ago. I was not sure about the indicator capabilitites but my profits from my recent three trades covered the cost. I am still not sure what the best way is to use this indicator. For example, in a T/S file, do you input just one line of parameters or multiple lines, what is better?. Which is the best target/stop to exit the actual position when multiple lines are present in a T/S file? I use the smallest value of the target to exit the position but maybe using the average makes more sense.

How do you backtest the p-Indicator? This is my problem since I upgraded too. It works nicely, I have gotten a few good signals and I think I will be concentrating on that portion of APS.

For the backtest, I am trying to develop a script that will do the following:

1. Start with a historical data file.
2. Run APS p-Indicator
3. Save the results in Excel
4. Extract the following from the excel file and historical data file and store in a text file: Date, O,H,L,C, PIL, PIS, S
5. Delete the last line in the historical data file
6. Go to step 2
7. Invert the text file to have the most recent data in the last line
8. Use the text file to backtest p-indicator

I think it should work in principle. Any thoughts?
 
Top