BioComp Dakota with Swamp Technology

[jalsck]

I think the way to go is to adjust the size of the stop loss based on the current volatility. Also, the number of contracts traded should be determined by working out the risk per trade that you want to take. Therefore, the more volatile the market the wider the stop loss and the fewer contracts traded.

Another idea that might payoff is to halve the number of contracts traded when the position is against the trend as determined by the 50 day moving average minus the 200 day moving average. Generally, it is more difficult to make a profit against the main trend.

I have it on my to-do list to build a trading system along those lines using AmiBroker. At the moment I am working on some R&D on the market timing signals.

Best Regards,

James

 

[jalsck]

Signal for Tue Feb 24, 2009 MOC:
Long

Trading Instructions for Tue Feb 24, 2009
None

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[jalsck]

Signal for Tue Feb 24, 2009 MOC:
Long

Trading Instructions for Tue Feb 24, 2009
None

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[jalsck]

Signal for Fri Feb 27, 2009 MOC:
Long

Trading Instructions for Fri Feb 27, 2009
None

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[jalsck]

Signal for Mon Mar 2, 2009 MOC:
Long

Trading Instructions for Mon Mar 2, 2009
None

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[jalsck]

Signal for Wed Mar 4, 2009 MOC:
Long

Trading Instructions for Wed Mar 4, 2009
None

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[jalsck]

Signal for Fri Mar 6, 2009 MOC:
Short

Trading Instructions for Fri Mar 6, 2009
Sell 2 ESH9 MOC

Open Position
Long 1 ESH9 @ 836.00

Closed Net Profit / Loss: $6,322.50

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[fundjunkie]

Quote from Humpy:

Have to do something about these big drawdowns imho James

Currently at over 50 S&P points again ?

How does your system look with a 10 point stop ?

Just my observation but if Biocomp was more helpful to struggling programmers developing better robots, then it would be beneficial to company and customers alike

This is a valid observation but, i would suggest, a somewhat unfair criticism of James' system. I'll try to explain why.

As you've realized this system is a counter trader and attempts to adapt to variations in market conditions. The notional capacity of a biocomp system to adapt is dependant in part upon the size of the parameter space and ability of the bots to understand and react appropriately to changes in that space.

It's generally acknowledged that swarm efficiency drops off if the parameter spaces is windened too much. However, in volatile conditions like these it's entirely possible that the configurd parameter space is too narrow to give the swarm any chance of avoiding drawdown. There are ideas out there that might overcome that hurdle but they require under the hood changes within Dakota to support them,ie. there's not much James can do right now (IMHO) to directly address them from a Dakota point of view.

So, given that, one could look to do some kind of post processing overlay. You suggest a stop loss but they tend to turn too many profitable trades into stopped losses. Instead, I favour a rule based stop & reverse where under certain criteria a stop triggers a position reversal.

These are all stop gaps though until more intelligent swarms can be coded that collectively take informed decisions on risk control. Right now, they could be coded to take those kind of decisions but the problem is they wouldn't be suitably "informed".

Hence James' system is probably doing about as well as it can given the extreme market conditions. And remember, what he's publishing here is the naked signal without the support of any supporting trade logic...


Thx
D

 

[jalsck]

Signal for Mon Mar 9, 2009 MOC:
Short

Trading Instructions for Mon Mar 9, 2009
None

Open Position
Short 1 ESH9 @ 687.25

Closed Net Profit / Loss: ($1,115.00)

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

 

[Humpy]

Quote from jalsck:

Signal for Mon Mar 9, 2009 MOC:
Short

Trading Instructions for Mon Mar 9, 2009
None

Open Position
Short 1 ESH9 @ 687.25

Closed Net Profit / Loss: ($1,115.00)

Note: ($amount) indicates a loss.

Hypothetical trading history is attached.

Kind Regards,

James

At last it's gone short !!

Your account has been blown away completely !

Excuse me for saying so, but I'm would consider going long on a contrarian view of your system !
Brave effort but you have only demonstrated that a lot needs to be done to make the system viable !

---------------------------------------------------------------------------------

So, given that, one could look to do some kind of post processing overlay. You suggest a stop loss but they tend to turn too many profitable trades into stopped losses. Instead, I favour a rule based stop & reverse where under certain criteria a stop triggers a position reversal.

----------------------------------------------------------------------------------


Certainly needs something