Don't know what to do when account < $40 million

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
marriot
Roundtable Knight
Roundtable Knight
Posts: 351
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot » Mon Feb 14, 2011 1:50 pm

After months (years?) of tests I finally created the perfect suite.
A great mix of different systems that work on a portfolio of
hundreds futures.
It is a long time that i gave up on find the best MAR, R3 or R Sharp, since
these carry DD to unacceptable levels.
The percentage of risk is now very low on each system.
Let's say that having 40 million i would know what to do.
Now for the brutal reality ':
Although the "portfolio" is composed of
hundreds of markets, the low-risk set excludes me from
many signals, it makes me go check if i have forget to insert that future
on which i am not active and,
Damn, is making a killing.
So 'I try to start tests on different dates, I try to test five
rather than 30 years and it seems that i can risk more,
at least until the capital has not reached " X".

X 's my nightmare.
I have no idea what kind of tests can show me the ligth.
For the moment, I am looking at Margins / Equity Ratio and this seems a bit '
strange as it is 'one of the few things "not really testable' .

stancramer
Roundtable Fellow
Roundtable Fellow
Posts: 60
Joined: Sat Feb 07, 2004 10:34 am
Location: Washington DC

Post by stancramer » Mon Feb 14, 2011 3:59 pm

Several people discuss their solutions to the problem here: viewtopic.php?t=5991

marriot
Roundtable Knight
Roundtable Knight
Posts: 351
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot » Tue Feb 15, 2011 3:30 am

Thank you Stan,
really my point was discussed long time ago and lead to this:
viewtopic.php?t=2663&highlight=live+fast+die+young.
Maybe after six years can we add differents point of view?
This test has a starting account of 1.5 millions.
In the first graph,whith a test starting date on 1970, we had a 30 % DD on Jan 2000.
Second graph, everything equal exept starting date.
DD on Jan 2000 is now only 20 %.
If starting with larger account, let's say 40 millions, there are non differences in DD what ever date you start.
While the reasons are obvious, how to find best time to cut down risk is the field on which i would like to listen some opinion.
Attachments
1997 start.jpg
1997 start.jpg (35.9 KiB) Viewed 2522 times
1970 start.jpg
1970 start.jpg (44.77 KiB) Viewed 2522 times

Paul King
Roundtable Knight
Roundtable Knight
Posts: 207
Joined: Mon Feb 23, 2004 9:13 am
Location: Vermont, USA
Contact:

Post by Paul King » Tue Feb 15, 2011 8:54 am

There is no simple solution to your problem since it's a fact that futures trading systems do not scale down well because it's impossible to trade fractional contracts. Variability of return due to start date is an effect of the scalability problem i.e. you need a certain size of capital at a certain point in time to be able to take all the signals your system is generating.

In my opinion, you have 3 possible solutions:

1 Redesign your system to use instruments that do scale down well (e.g. FX, equities)
2 Accept a worse/more volatile risk-adjusted return when trading with a smaller amount of capital (assuming it's still acceptable)
3 Don't trade until you have at least $40 million ;-)

Hope this helps

Paul

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

Post by sluggo » Tue Feb 15, 2011 9:33 am

CTAs who offer managed accounts work hard to ensure that all accounts perform alike, even though some accounts are much larger than the minimum account size.

One approach I've seen used, is to quantize the account equity. Usually 1 quantum equals 1.0 x the minimum account size. So for example if the minimum account size is $400K and customer Z's account equity is currently $950K, then customer Z is considered to have 2 quanta of equity.

Position sizes for new trades are calculated for the minimum account size (i.e. for a 1 quantum account). Then each customer gets some integer multiple of the minimum account's position size. Customer Z, who has 2 quanta of equity, gets position size 2x the position size of the minimum account.

What's good about this: all accounts get the same trades. If a trade is too risky for a $400K account, NOBODY gets that trade, not even the $90 million account. This tends to keep the performance of the various different managed accounts tightly grouped together, and oh by the way, it also simplifies the day to day operations and bookkeeping of the CTA.

What's bad about this: Nobody gets the diversification benefits of having a huge account, not even the customers with huge accounts.

One way to implement this idea in Trading Blox, appears below. "riskPerTrade" and "Equity_Quantum" are user supplied parameters, which is why they appear in red.

The exact opposite of this idea would be: EVERYBODY gets the benefits of having a huge account, even the customers with tiny accounts. Everybody gets the same trades because everybody takes all trades, even the trades that require a $40 million account size. Of course I'm talking about a Fund (also called a Commodity Pool), rather than managed accounts.
Attachments
quantized_equity.png
example code, MM script
quantized_equity.png (52.23 KiB) Viewed 2469 times

Post Reply