Article Developing a Trading Strategy Part 2

May 15, 2002
23
3
13
#16
Excellent article but I'm a bit confusedby the different results on your website. Is the higher points total just due to no commissions being taken into account?
Also, the Thursday filter seems to have little effect on the 2004 results

2004 points Thursday Filter added
Q1 489 545
Q2 380 467
Q3 327 192
Q4 469 329
 
Likes: wasp

sidinuk

Active member
Dec 4, 2002
624
4
28
Surrey
www.online-futurestrading.com
#17
shorty,

The figures in the article are after allowing 3 points/trade for slippage and commission, on the website they are gross.

The Thursday filter does reduce the total points made in 2004 from 1,665 to 1,533 but it also reduces the number of trades from 195 to 161 and so increases the points/trade from 8.5 to 9.5. That is an important consideration.
 
May 15, 2002
23
3
13
#18
ok, thanks for the reply. I take your point about the reduced trades (and therefore reduced costs) but I'm trying to get my head round the Thursday factor. Most market days the important news comes out at 8:30 am or 10.00am so there is plenty of time for the market to digest it before the system's 11:45 start time. I was thinking along the lines of important data after 11:45am might cause the plan to falter on a specific day. The main announcements I can think of after 11:45am are Beige Book, Fed Interest Rate Decisions, Philly Fed ...but only Philly Fed is ususally on a Thurs I think, so prob not enough on it's own to effect the plan
 
#19
As a system developer I must say excellent articles, love the larry willams trade day of the week filter.

The most popular s&p500 trading systems have a breakout with countertrend component which I'm glad you mentioned! They are tested by futures truth website. (If anyone is interested)
 
Likes: wasp

Tuffty

Active member
Oct 15, 2003
442
8
28
#20
sidinuk, Is the optimisation strategy you used in the example the one you use for ‘live’ systems (i.e. ones you trade)? The strategy as I understand it being to optimise part of the data set, test it on unseen data and then go ‘live’ if you’re happy with the results without re-optimising.
 
#21
Other filters you can use our intermarket relationships such as bonds, gold being up or down several days ago. Also their has been some interesting work done in the area of where the closing price is in relation to the days range !!!

Tom Demark did extensive studies in the area of daily range expansion when he worked for paul tudor jones in the 1980's, they found a good entry around a four (4) day high. So the 5 day filter was a good choice.
 

sidinuk

Active member
Dec 4, 2002
624
4
28
Surrey
www.online-futurestrading.com
#22
Tuffy, yes but I would use more data for testing - around 2 years.

knucklehead, I have tested (not for this system but another breakout system) the relationship between the closing price and the day's range but I found it less effective than just the size of the range.

Last weeks results for the system are available here: week ending 14 Jan 2005. The Thursday filter is working against us at the moment!
 
Nov 25, 2004
6
0
11
Dublin
#23
Exit Time

sidinuk,

Did you backtest your strategy by exiting your open positions at 21:00 hrs or 21:15 hrs London Time?
Probably, it would not make a big difference but I just like to know.

Regards,
Mr. H
 
Last edited:

rog1111

Active member
Jan 30, 2004
673
10
28
#25
This is far better material than found in many commercially available trading books. Well done !

Roger
 
Nov 25, 2004
6
0
11
Dublin
#29
sidinuk said:
Mr H, I used 9.00pm London as the close as that is when the underlying market closes. I've just done a quick test and over the 24 months (Jan03-Dec04) a close at 9.00 makes 3,556 pts and 9.15 makes 2,978. That is including Thursdays!
Sidinuk,

Thank you very much for your reply.
I am quoting the following from your article:
It should be noted that over these 2 years the Dow Jones index has experienced very low volatility when compared to previous years making it a fairly difficult time for day trading systems. However, the currency markets have been volatile over this period and we could have chosen to develop a system to trade dollar/euro or dollar/pound futures, which would have been far more profitable.
I hope you would soon develop a similar strategy for trading the Currency Futures.

Regards,
Mr. H