Saturday, 13 November 2010

Another trading week

The last week or so I have been looking at equity-curve risk management. This can be an odd topic as it is based on a meta principle approach. I remember years ago raising something like this with one of my hedge fund shareholders and he was very puzzled. The stumbling block is that at the "meta-level", one might trade in ways that are actually in conflict with the way one trades in the actual markets. For instance, we tend to lose money when a solid trend occurs intra-day. As counter-trend traders we add to our position as things go against us and so are maxed out in a trend and lose heavily if that trend continues. Yet equity risk management would chop that position just when the underlying trading would expect a rebound. The Sharpe ratio impact of this is so dramatic that it is worth doing though, no matter how hard it can be to exit positions for equity curve risk management purposes

Also this approach suggests a limit exit when things go well - which is also rather counter-intuitive. But research suggests that a limit exit at +2.5% intra-day is the soundest policy.

So I have been preparing MAE and MFE data for the trading system equity (something that can take me 2 hours per day to work out as we haven't got adequate information to work it out more easily). And we have a tentative equity risk-management policy coming out from this.

Interesting this also suggests that some parts of the trading system aren't pulling their weight, one of which is the core trading signals that power the Martingales. So oddly enough, we might end up trading a martingale system but without the systems from which the Martingales are derived.

No comments: