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.
A recent issue of Futures magazine quoted a CTA saying, "The imminent war with Iraq presented issues of volatility and correlation that no system regardless of robustness could adequately measure."
Do you agree with that quote? Is it your understanding that a proper system should have the correct failsafes inherently designed into it to prevent "issues of volatility" or is a little bit of override discretion required in all systems for the extreme circumstances? Maybe the CTA's thinking wasn't robust enough to design a robust enough system.
Regards,
Derek
MODERATOR'S NOTE: This thread was split from the thread "Mistakes c.f. and Others have made"
The above-mentioned CTA wrote:"The imminent war with Iraq presented issues of volatility and correlation that no system regardless of robustness could adequately measure."
This sounds like an excuse for poor performance to me.
I don't think the pre-war volatility was that bad. Yes, last year had some tough drawdowns but nothing outside the envelope for most systems.
If you feel the need for discretionary overrides, I suggest strongly thinking about the circumstances ahead of time and then developing an extra rule. That way when the time comes you won't be reacting emotionally.
Unfortunately, it's hard to test these sorts of rules since there may be only a handful of historical events that might trigger such a rule. October 1987 and September 11 are the only ones that come to mind since I started trading.
Personally, I prefer to let the stops and the systems take care of themselves. A well-diversified portfolio of systems will let you make money off the volatility that this CTA describes.
For a different viewpoint, I think that a systematic trading system is much more discretionary than most traders recognize. A trader first uses ultimate discretion in building and testing the system, then in choosing markets to trade, allocating capital to the system, and choosing when to begin trading the system. These are a few examples of discretionary aspects of systems trading, I am sure there are more.
A system is properly used to do wrote tasks without mistakes, not to think so that a trader does not have to. It happens that one of the keys to profitability in the markets is implementing rather simple ideas consistently without too much outsmarting ourselves. Systems are great at doing this.
However, a system is not built to make the most profitable / least risky decision in any individual instance. It is also not aware of material facts that a disciplined trader may be able to use to produce better results than the system alone. The war in Iraq is a good example of one of these rare circumstances. The trader is aware, beforehand, that an unusual situation that his system is not optimized for is at hand. Most "black swan" events are not so easily dealt with (9/11 for instance) because they cannot truly be prepared for in advance. In these cases, a trader has to either make some rules beforehand that dictate the actions to take in such a case or rely on his historical testing and system robustness to perform.
All in all, my opinion is, if a trader can see a situation coming that he knows his models are not prepared for, he is justified in using discretion just as he is justified in using discretion in the design and implementation of other parts of the system. It is a matter of knowing exactly how and when to break the rules. It is much more difficult to know how to do this in the aftermath of an unforseen event. I view the way a trader might deal with the Iraq war as more akin to the way one might decide to exclude a market that has undergone a major fundamental structural change from a trading portfolio, rather than as a "panic" situation.
It is a fine line, but I believe that the more prone we are as traders to "never touch the system, under any circumstances" the more likely it is that we have created a crutch for ourselves instead of a tool. Of course, as many contingencies as possible should be planned for in advance in the system design phase, but we can't plan for everything. If this inability to plan for every possiblity restricts our range of choices, I don't view that as a good thing.
BTW, one of the CTAs qouted in that article that reduced leverage during March 2003 was Salem Abraham. He returned 74% last year. He is also an example (like everyone cited in the Futures article) of a trader that did not really alter his basic strategy, but rather reduced leverage in the face of uncertainty. When managing money professionally for others, this is often a more prudent course of action than "blindly" following the system even when you can see that your volatility exposure may rise to a dangerous level. As a private trader, one has a bit more leeway to be the guy that followed the system through WWIII and ended up making a killing because of it, despite the huge swings.
