Trade Sizing and Money Mgmt

minx

Well-known member
Messages
328
Likes
10
Ok I'm interested in the pros and cons of these methods of trade sizing/MM:
Oh, and I'm talking about futures contracts here so we arent as scalable as stocks (lumpy margin scaling and all that)

1) Max Historical DrawDown + % Buffer:
Basically you have starting capital, say £10k and you find that on a 10-lot you have an historical DD of £15k, then add your buffer, say 25% to this to make £18,750. Divide this into your £10k and you get 0.53333, this means you can now only do a 5-lot with your £10k.

2) Equity %: With your £10k you still start with your 10-lot and you adjust as your capital increases or decreases, quite simple.

3) You use the Max Historical DD + Buffer% and scale this to Equity %. This is good as you have two systems in use to protect your capital.

Cons:

Method 1): Statistically you are likely to have a max DD in the future that is greater than you have witnessed in the past. This means you could blow up on even a modest increase in max DD.

Method 2): This method means oyu have to decide on your max loss, which affects your sizing which can then affect how you can adjust to volatility.

Method 3): Hmmm, am thinking that this is a robust system of risk management.

I'd really appreciate peoples thoughts, their own pros and cons to these ideas. If you think one is particularly good/bad then please also say why.

I would also add that I have another sizing formula I use but thats on top of the base size (in method 1this base is 5 lots, method 2 its 10 lots, etc...) based on a number of factors.

Thanks!
 
To get an accurate likelihood of DD I run a Monte Carlo analysis over the trades. Effectivley I put them all in a hat. Pull one out record it and put it back in the hat. If you do this 000's of times you get a distribution likleihood of various DD limits. I then use the DD at the 95% level, and still add a saftey margin.

Of course the trades must be representative of what you may expect in the future so you have to not use any over-optimised parameters to produce them.

I also ensure any total position size will not destroy me given a catastrophy (i.e. assuming the market closes mid session thus destroying the protection of any market stop losses). This normally limits how big I can run a system rather than anything else.
 
Top