Multiple System Allocation
- Thread starter trend456
- Start date
I think I see what you're getting at now trend.
Are you looking for ways to have your systems interact such that position size in each system is dynamically adjusted to achieve an overall portfolio risk level?
That would be pretty cool. Basically the way the risk management function of an IB controls the overall exposure of the firm by placing limits on the individual trading desks' activities maybe?
Are you looking for ways to have your systems interact such that position size in each system is dynamically adjusted to achieve an overall portfolio risk level?
That would be pretty cool. Basically the way the risk management function of an IB controls the overall exposure of the firm by placing limits on the individual trading desks' activities maybe?
trend,
the "New Money Management" is much deeper than optimal F. Optimal F is about position sizing for one system/market. The NMM is about finding the efficient frontier for multiple systems, markets, strategies under your risk constraints.
It's fortunately not difficult to understand, and the math is easy. Can all be done in e.g. Excel.
the "New Money Management" is much deeper than optimal F. Optimal F is about position sizing for one system/market. The NMM is about finding the efficient frontier for multiple systems, markets, strategies under your risk constraints.
It's fortunately not difficult to understand, and the math is easy. Can all be done in e.g. Excel.
trend456,
I use multiple time frame systems to help me scale in and out when trading the same contract.
It works like this.
You have four systems each operating in a different time frame on the same contract.
You back test each individually to establish the largest drawdown for each.
You then multiply the largest drawdown in dollar terms by five. This will give you the amount required to trade the system with a theoretical drawdown of 20%.
You them trade each system as if is independant of the other with the total capital required for the four systems.
Your total position at any one time with either be 4 contacts, 2 contracts or flat.
This is because if you a long 3 and short 1 you are net long two.
Or if you are short 3 and long one you are net short two.
If you are long 2 and short 2 you are net flat.
And of course if you're long 4 and short none you are net long 4.
Or you can just be flat 4.
What this allocation model does is it keeps you out of the market when your time frames are conflicting.
When the market is trending, you'll find that all time frames line up and you can be fully exposed.
I like this system because it:
A. Prevents you from over trading (conflicting time frames act as a trade filter.)
B. Prevents you from betting too much money when the market just aint paying out.
C. Dramatically smooths the equity curve.
Runningbear
I use multiple time frame systems to help me scale in and out when trading the same contract.
It works like this.
You have four systems each operating in a different time frame on the same contract.
You back test each individually to establish the largest drawdown for each.
You then multiply the largest drawdown in dollar terms by five. This will give you the amount required to trade the system with a theoretical drawdown of 20%.
You them trade each system as if is independant of the other with the total capital required for the four systems.
Your total position at any one time with either be 4 contacts, 2 contracts or flat.
This is because if you a long 3 and short 1 you are net long two.
Or if you are short 3 and long one you are net short two.
If you are long 2 and short 2 you are net flat.
And of course if you're long 4 and short none you are net long 4.
Or you can just be flat 4.
What this allocation model does is it keeps you out of the market when your time frames are conflicting.
When the market is trending, you'll find that all time frames line up and you can be fully exposed.
I like this system because it:
A. Prevents you from over trading (conflicting time frames act as a trade filter.)
B. Prevents you from betting too much money when the market just aint paying out.
C. Dramatically smooths the equity curve.
Runningbear
I trade 6 strategies on the E-Mini, Dax, US, Bund and Euro currency. The six systems are non-correlated but in the same time frame. I can be 0, +/- 2, +/-4 or +/-6. The only problem was entering the orders on a timely basis but Tradebolt solved that problem. Drawdowns are reduced, equity curve smoother and all stress gone!
Obviously there are strong correlations between ES & Dax and US and Bund but the multi strategy approach generally keeps you out of whipsaws but kicks in nicely on the swings.
Obviously there are strong correlations between ES & Dax and US and Bund but the multi strategy approach generally keeps you out of whipsaws but kicks in nicely on the swings.