Minimum number of markets traded

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
Jwebster
Roundtable Fellow
Roundtable Fellow
Posts: 52
Joined: Fri Jul 04, 2003 10:23 am
Location: London, UK

Minimum number of markets traded

Post by Jwebster » Wed Oct 22, 2003 9:19 am

Does anyone have a view on the minimum number of markets a system should be able to trade profitably before you can call it a valid system? One I am playing around with returns great results on a portfolio comprised of swiss franc, yen, Nikkei225 index and 10 yr note. If I add other markets such as orange juice and copper results decline as the system is either marginally profitable or incurs losses in these latter markets. I realise that in optimizing a system it is best to do this over a number of markets so as to improve the validity of the indicators, I’m just not sure if having a system that works well over four seemingly diverse markets is enough. The high correlation between these markets also concerns me, even though the 10 year note has a strong negative correlation to the swiss franc.

redbullpeter
Roundtable Fellow
Roundtable Fellow
Posts: 82
Joined: Mon Apr 28, 2003 9:12 am
Location: London, UK

Post by redbullpeter » Wed Oct 22, 2003 9:47 am

Funny thing I was about to post an identical question.

I'm working on a system that is highly profitable on Nikkei 225, Hang Seng Index and FTSE 250. Using excel the correlation between the markets on a daily, weekly and monthly scale are between 0.33 and -0.33 for Nikkei/Hang Seng and Nikkei/FTSE250 but around 0.62 for Hang Seng/FTSE 250. I adjust position sizing to compensate for correlation. The idea is that if I am currently in a trade for Hang Seng Index and a signal for FTSE 250 shows up, the amount of equity risked is scaled down for the new trade. I need to do more testing to see if this is a "safe" thing to do.

Does anyone have any insights?

Peter

Ted Annemann
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Tue Apr 15, 2003 7:44 pm
Location: Arizona

Post by Ted Annemann » Wed Oct 22, 2003 11:32 am

The price data might have low correlation yet the trading system equity curves might have high correlation. Since you've only got 3 markets "F", "H", and "N", it only takes a few experiments to test this out fully and completely.

Right now you are trading portfolio #1 below: (1/3) F, (1/3) H, and (1/3) N. You could test to see if its trading results are noticeably better than any of the other nine portfolios. If so then the equity curves have beneficial non-correlation. If not then you can trade fewer instruments without degrading performance.

1. F=(1/3) H=(1/3) N=(1/3)
2. F=(3/3) H=0 N=0
3. F=0 H=(3/3) N=0
4. F=0 H=0 N=(3/3)
5. F=(1/3) H=(2/3) N=0
6. F=(2/3) H=(1/3) N=0
7. F=0 H=(1/3) N=(2/3)
8. F=0 H=(2/3) N=(1/3)
9. F=(1/3) H=0 N=(2/3)
10. F=(2/3) H=0 N=(1/3)

redbullpeter
Roundtable Fellow
Roundtable Fellow
Posts: 82
Joined: Mon Apr 28, 2003 9:12 am
Location: London, UK

Post by redbullpeter » Thu Oct 30, 2003 3:58 am

Ted,

Thanks for your thoughts. It turns out that that particular market set is the minimium set possible while keeping the performance parameters within my required limits. Gives me a reference point as I add more markets.

Thanks

Peter

Post Reply