Re: [amibroker] PositionSize based on Equity and Trade Risk, AmiBroker Email List Archive

David,

I haven't gotten to the point of selecting position size, however, if you know your entry price and initial stop, you can use something like the following to determine initial dollar risk:

Prange = PHighPrice - PLowPrice;
MaxLoss = PointValue * PRange;
MaxLossLimit = Optimize ("MaxLoss", 1000, 150, 4000, 25);

In this case I am going long at PhighPrice with an initial stop a PLowPrice and not taking the trade if the MaxLoss exceeds a certain level. From this point you could calculate the MaxLoss as a % of equity. Does this help?

James

From: dbwyatt_1999 <dbw451@xxxxxxxxxxxxx>
To: amibroker@xxxxxxxxxxxxxxx
Sent: Saturday, May 30, 2009 7:39:47 AM
Subject: [amibroker] PositionSize based on Equity and Trade Risk

I know setting PositionSize based on Equity and Trade Risk has been the topic of numerous threads and writeups, but I'm missing some understanding. I'm backtesting a portfolio of futures contracts and would like to set the Position size of each trade based on each trade's risk (i.e. abs(entryPrice - stopPrice)) and a percent of the current portfolio Equity value. I understand that to get the current portfolio Equity, I need to use the custom backtester. The custom backtester also gives me access to all the data I need to calculate a new PosSize except my stopPrice values.

What is the easiest way to access my backtester Phase 1 stopPrice array in the custom Phase 2 backtester for each trade so that I can calculate a new PosSize? Using AddToComposite( ) and Foreign() does not seem like an option because I want my individual stop values for each instrument, not a portfolio composite of all stop values. Or maybe I would have to create a separate composite for each instrument. Or do I have to resort to something like writing and reading files or developing a custom DLL to store the stopPrices between the backtester Phase 1 and the custom backtester Phase 2? Hopefully, I'm just not understanding something and there is an easier way. I would think that accessing Phase 1 backtester variables in Phase 2 would be pretty common; especially for calculating custom metrics.

Thanks for any suggestions or insights.

Regards,

David