Open Old Positions When Starting to Trade a System?

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
JohnTHill
Contributing Member
Contributing Member
Posts: 7
Joined: Wed Mar 11, 2009 6:51 pm
Location: United States

Open Old Positions When Starting to Trade a System?

Post by JohnTHill » Mon Mar 16, 2009 11:27 pm

I am interested in beginning to trade stocks in a dual moving average system. It is a moderately long term system. When I generate orders for day one my current software gives me orders to open today that really would have been generated at some time in the past had I been in the market the whole time. Diversification aside should I be opening all these positions on day one or just ignore them until new signals are generated and trade them?

I know this has been mentioned before here but this is so important to me I wanted to hear peoples most modern stance on it. I can test this some and of course I will. That said I am looking for the guidance and experience that only people who have been there before can provide me with in addition to my test results. Your advice and assistance are greatly appreciated. Thank you.

JT

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue Mar 17, 2009 4:08 am

The first question is whether you want to draft and test a version of the DMA which only trades immediately after a crossover and compare it to the performance of that which you have already tested. Even so of course you will still be faced with the question of whether to assume all existing positions when you start up.

People have argued (convincingly in my opinion) that new investors in a trend following programme run by an outside third party expect to participate in the equity curve immediately they hand their money over. Such people may well attempt to time their initial investment so as to coincide with a drawdown rather than an equity high.

A drawdown from initial capital is far more de-stabilising than a retrenchment from a new equity high. You face the problem not only on initial start-up but also on an increase in trading capital.

That sort of timing is a question, largely, of luck. Whatever the CTAs do and whatever the back tests may show, you may be more comfortable taking only new signals and working your way slowly into the programme. Only you can decide.

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Tue Mar 17, 2009 6:30 am

Method 1.
Suppose someone else had started trading your exact system(s) and your exact portfolio(s), three years ago. Also suppose they had started with a small amount of capital and, after three years of trading, had grown it nicely. Further suppose their total account equity today is exactly equal to the amount of money you will use to start trading.

An interesting question to ask is: After three years and a large number of trades, what positions does this person have TODAY? They've got the same amount of money as you, they're trading the same systems as you, they're trading the same markets as you. Maybe you should immediately put yourself into their shoes and immediately enter the exact same positions at the exact same sizes?

Method 2.
Systematic traders acknowledge that emotions can interfere with decisionmaking, and so they strive to separate emotions from trading where possible. (In fact, trading systems themselves are structures that codify and enforce emotionless decisionmaking. At least in theory.) However, Start-Up is a time of particularly intense emotion; you struggle with two different kinds of remorse. The first kind is the remorse of not immediately jumping into an existing trade, only to watch that trade go on to make big fat profits without you. The second kind of remorse is when you do jump immediately into an existing trade, only to have it make a loss (perhaps even a great big loss) right at the beginning of your trading experience.

The solution? Intentionally & deliberately create equal portions of both kinds of remorse. With half your money, immediately jump into all existing trades. With the other half of your money, don't enter into any existing trades. (The net effect is: you enter into all positions at half-size). This is easiest to do in instruments with very small minimum lotsize, such as stocks, mutual funds, CFD's, spreadbets, and miniForex. Notice the result of this Method: whatever happens, you know in advance that you will only experience half-size remorse. If a trade goes really well, you've got half a position, so you will feel remorse at not taking the trade at full size. If a trade goes really badly, you've got half a position, so you will feel remorse at not skipping the trade entirely.

After the Start-Up phase is complete, you can gradually become a fully systematic trader who disdains emotions like remorse, finding them irrelevant to trading.

Method 3.
Ask a guru what to do and follow their advice. If the result is pleasant, you are happy. If the result is unpleasant, it's not your fault: the guru steered you wrong.

nzbryant
Senior Member
Senior Member
Posts: 29
Joined: Mon Dec 13, 2004 10:43 pm

Post

Post by nzbryant » Wed Mar 18, 2009 6:36 am

Great post Sluggo, as always.

-w.
Roundtable Fellow
Roundtable Fellow
Posts: 95
Joined: Wed Nov 19, 2003 11:02 am

Post by -w. » Wed Mar 18, 2009 12:13 pm

I'm totally with sluggo's Method 1, however there are some pitfalls you might want to consider. I've been working with moving average (crossovers) as well and so have had the same problem but not only in the starting phase of the system.

I also use filters, so it happens that trades in some markets are not taken because of the filters' restrictions but when other positions are closed, these markets become available for trading, sometimes quite late after the original signal would have entered.

If the averages are not too long, a crossover usually enters you in the direction of the market. Entering basically randomly just because a slot became available by closing another market often lets you enter in a short term countertrend, triggering stops that would not be there had the position been taken on the first signal.

So what I did was adding some sort of confirming entry signal, such as enter long if short MA > long MA AND instrument.close > short MA or something. That reduced the whipsaws a lot.

-wojo

JohnTHill
Contributing Member
Contributing Member
Posts: 7
Joined: Wed Mar 11, 2009 6:51 pm
Location: United States

Post by JohnTHill » Wed Mar 18, 2009 7:27 pm

Thanks, I appreciate the feedback and help from everyone!

I am working on some tests and also tend to favor Sluggo’s Method 1 at the moment.

Bottom line is some fear is coming up and at a certain point in time you need to decide if you are going to dive off the high dive or not! Of course all this reminds me of why I want to do system trading in the first place and leave my emotions behind. Thanks again everyone!

Feel free to send me a bill for psychological services rendered. - LOL

RedRock
Roundtable Knight
Roundtable Knight
Posts: 941
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

rainbow cones

Post by RedRock » Thu Mar 19, 2009 12:15 am

I like the one about asking the client which flavor of risk he/she prefers up front, given that there may be no correct answer. perhaps a bit of each is a tasty compromise.

Post Reply