Article Developing a Trading Strategy Part 2

Sharky

Staff member
Nov 4, 2000
5,402
253
143
42
www.trade2win.com
#1
In the first part of this article, which can be read here, we looked at choosing an instrument and timeframe to trade, as well as establishing the set-up and entry rules. In the second and final part we will consider how to establish exit rules as well as various filters and money management rules to maximise the profitability of the system.
6. Stop Loss Rules.
Our strategy already has a natural stop loss in the stop order that does not get filled. The objective of the strategy is to capitalise on those days where the high or low for the day is in place early (9.30-11.45am). If we enter a trade on a breakout of either the high or the low and then the market subsequently hits the other stop we know that our trade is invalid. We know from our testing earlier that this only occurred 10% of the time.
We could add additional rules for the stop loss such as:

Moving the stop to breakeven when we are a...
Continue reading...
 
Last edited:

Tuffty

Active member
Oct 15, 2003
442
8
28
#4
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).
 

sidinuk

Active member
Dec 4, 2002
624
4
28
Surrey
www.online-futurestrading.com
#5
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.
 
Likes: johnlvs2run

G-Man

Active member
Feb 26, 2003
243
3
28
48
London
#6
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.
 

TWI

Well-known member
Jan 22, 2004
2,516
245
73
London
#9
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.
 

JonnyT

Well-known member
Dec 19, 2001
2,560
22
48
54
Notts
www.spotontrading.co.uk
#11
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
 

darrenf

Active member
Jan 28, 2003
481
3
28
#12
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
 
Likes: wasp

G-Man

Active member
Feb 26, 2003
243
3
28
48
London
#13
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.
 
Likes: wasp

TWI

Well-known member
Jan 22, 2004
2,516
245
73
London
#14
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.