Developing a Trading Strategy Part 2

This is a discussion on Developing a Trading Strategy Part 2 within the Trading Systems forums, part of the Methods category; We've just published a new T2W article called " Developing a Trading Strategy Part 2 " by Tim Wreford . ...

Reply
 
LinkBack Thread Tools Search this Thread
Old Jan 11, 2005, 7:15pm   #1
The Staff are paid members employed by T2W to perform some specific role such as editorial, advertising or technical work.
 
Sharky's Avatar
Joined Nov 2000
Developing a Trading Strategy Part 2

We've just published a new T2W article called "Developing a Trading Strategy Part 2" by Tim Wreford.

Quick Summary: The second part of Tim's popular guide to creating a trading strategy.

PS. Don't forget to rate the article after you've read it and share your comments on this thread.
Sharky is offline   Reply With Quote
Old Jan 11, 2005, 7:15pm   #2
 
JonnyT's Avatar
Joined Dec 2001
Excellent.

JonnyT
JonnyT is offline   Reply With Quote
Old Jan 11, 2005, 11:53pm   #3
Joined Jan 2004
Impressed with the author's rigour and application. Not only that, the content is first class.
Many thanks, very useful.
alpha_monkey is offline   Reply With Quote
Old Jan 12, 2005, 9:45am   #4
Joined Oct 2003
Very good article. I appreciate the effort that has gone into writing it.

The only thing I was uncomfortable with in the example was finding significance in Thursdays not being profitable. One could argue this is over-fitting as in the test period the decision to cut out Thursdays was based on 22 trades (not really enough in my book) and based on the fact it wasn't profitable in the test period. However, this could be confirmed or not by looking at the out of sample data OR/AND asking if there is any logic behind cutting Thursdays (I don't know of any, if it were a Monday one may be able to put an logical argument together).
Tuffty is offline   Reply With Quote
Old Jan 12, 2005, 12:09pm   #5
 
sidinuk's Avatar
Joined Dec 2002
Thanks for your comments!

Tuffy, the Thursday thing really surprised me as well. My sample period for the purpose of the example in the article was only 6 months which as you say only gives 22 Thursdays to base the rule on. The point was, though, that we base our system on data in the test period and in the test period Thursday's were extremely poor. If we then look at the out of sample data and change the rule then our out of sample data then becomes part of our test data - invalidating the out of sample test! Your right, though, a larger test period in the first place would be better, this was only an example for the article.

Friday is the second worse performing day so we could conclude that the beginning of the week is better for breakouts than the end. Logically, many investors will make trading decisions over the weekend and implement them at the beginning of the week, as the week grinds on they wait until the weekend to decide what to do next. Thus, the market makes it's move in the beginning of the week and consolidates towards the end of the week.
sidinuk is offline   Reply With Quote
Thanks! The following members like this post: johnlvs2run
Old Jan 12, 2005, 12:25pm   #6
 
G-Man's Avatar
Joined Feb 2003
Have to agree with other posters. This is one of the best articles I have read for a while.

With respect to the Thursday factor. I had been intending to create a thread regarding this as I also find Thursday a pretty poor day for day-trading my FX (primarily Cable) strategies. Again the sample size is a little small for a definitive conclusion, (max 16 months data), but I still think that this is a factor to big to ignore.

Has anyone else noticed this or a similar phenomenon with their instruments, and if this is a recurring theme, does anyone trade counter to their strategy on the poor performing day.

Comments appreciated.
__________________
He who never had a loss, never made a profit
G-Man is offline   Reply With Quote
Old Jan 12, 2005, 1:28pm   #7
Joined Jan 2004
http://www.seykota.com/tribe/risk/index.htm

Quite a useful piece that complements the article under discussion.
alpha_monkey is offline   Reply With Quote
Old Jan 12, 2005, 1:33pm   #8
Joined Jul 2001
Missed part one!

Quote:
Originally Posted by Sharky
This is a forum thread for discussing the Knowledge Lab Article, "Developing a Trading Strategy Part 2".
Interesting stuff, where is part one. I looked in the Lab??
Gardan is offline   Reply With Quote
Old Jan 12, 2005, 1:33pm   #9
TWI
 
TWI's Avatar
Joined Jan 2004
Very good article. Thanks.
Will there be a part 3 ? If so, maybe you could address execution risk. For example, system in question would in reality have end results adjusted for cost and slippage. This actually becomes significant over 200+ trades and leaves little room for poor execution. Assuming 1tick slippage and $1.25 bro per side, runing it through TS we get an average trade expectation of less than 3pts vs. over 5 pts before. This still works but it is tight.
Incidentally, if anybody wants code for TS I will post it.
__________________
"I refuse to tiptoe quietly through life
only to arrive safely at death"
TWI is offline   Reply With Quote
Old Jan 12, 2005, 2:22pm   #10
 
rglenn's Avatar
Joined Dec 2002
Excellent Article.

Quite an eye opener!
rglenn is offline   Reply With Quote
Old Jan 12, 2005, 2:43pm   #11
 
JonnyT's Avatar
Joined Dec 2001
The article did allow 3 points for slippage and commission.

My own experience autotrading a slightly different version of this is that maybe 4 points should be used for slippage and commission. However I'm working on getting this down.

JonnyT
__________________
Trading Mentor
JonnyT is offline   Reply With Quote
Old Jan 12, 2005, 3:45pm   #12
Joined Jan 2003
Quote:
Have to agree with other posters. This is one of the best articles I have read for a while.

With respect to the Thursday factor. I had been intending to create a thread regarding this as I also find Thursday a pretty poor day for day-trading my FX (primarily Cable) strategies. Again the sample size is a little small for a definitive conclusion, (max 16 months data), but I still think that this is a factor to big to ignore.

Has anyone else noticed this or a similar phenomenon with their instruments, and if this is a recurring theme, does anyone trade counter to their strategy on the poor performing day.

Comments appreciated.
I trade fx, primarily cable at the mo. A number of fx traders avoid Mondays as they can be quite choppy. I am not quite as extreme as this but through observations of market action I have concluded that I should avoid trading on the Monday following the Non farm payroll data which generally comes out on the first friday of the month. This usually results in a volatile friday, generally with a big move one way or the other (or both!) on Friday PM. It is quite logical that the Monday session is likely to consolidate and my observations seem to confirm this
darrenf is offline   Reply With Quote
Thanks! The following members like this post: wasp
Old Jan 12, 2005, 5:57pm   #13
 
G-Man's Avatar
Joined Feb 2003
Darrenf,

Thanks for your observation about Monday trading after payroll data. I hadn't especially noticed reduced performance on those days so I will look out for that.

As regards other Mondays, although they tend to offer better performance, for me, than Thursdays, they still lag well behind Tue, Wed and Fri. As I think it is well known that some traders avoid Mondays I always thought that could explain the quieter Monday performance, and the normally stellar Tuesday performance but I could never fathom why Thursdays were so poor. Any ideas to explain this very welcome.
__________________
He who never had a loss, never made a profit
G-Man is offline   Reply With Quote
Thanks! The following members like this post: wasp
Old Jan 12, 2005, 6:59pm   #14
TWI
 
TWI's Avatar
Joined Jan 2004
Quote:
The article did allow 3 points for slippage and commission
Yeah, I saw that but the results based on my own data corresponded with the final results in the article before I introduced any slippage and Bro. Once I had, the per trade expectation was considerably less. I rechecked the code and audited trades to be sure I was correct.
__________________
"I refuse to tiptoe quietly through life
only to arrive safely at death"
TWI is offline   Reply With Quote
Old Jan 13, 2005, 8:54am   #15
 
sidinuk's Avatar
Joined Dec 2002
My own day by day back test results for 2003-2004 are now available at www.online-futurestrading.com . I will also keep this up to date each week with current performance.
sidinuk is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Similar Threads
Thread Thread Starter Forum Replies Last Post
Article: Developing a Trading Strategy T2W Bot Educational Resources 15 Sep 26, 2013 3:54pm
Advice on developing a strategy adzy First Steps 8 Aug 17, 2005 8:43pm
Developing Trading Systems for the E-Mini josbarr Trading Systems 0 Jun 30, 2005 2:32pm
Part-time strategy SAINT First Steps 1 Jun 3, 2004 8:32pm

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)