Discretion

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
J D Canning
Roundtable Fellow
Roundtable Fellow
Posts: 74
Joined: Thu Jul 03, 2008 6:35 am
Location: Surrey, UK
Contact:

Discretion

Post by J D Canning » Fri Jan 23, 2009 2:02 pm

I would like to bring up the topic of allowing for a certain degree of discretion when trading a mechanical trend following strategy. I know that these two approaches should be mutually exclusive, but we are currently seeing some pretty extreme movements across number of markets, some of which I feel could potentially prohibit the opening of new trades.

Example 1: 3-month Eurodollar. Here, we have seen the March contract touch 9920 recently before coming off slightly. Using a Donchian Breakout system, it may come to pass that a buy signal is generated at or around these levels at some stage. Can anybody honestly justify taking this trade? The upside potential has to be limited and, as far as I am aware, one of the precepts of mechanical teend following is to use trailing stops to allow for a potenially unlimited upside.

Example 2: Crude Oil. I know that there have been a number of false bottoms over the past nine months, but who could honestly justify shorting crude oil at $36? Maybe what I'm driving at here is that there is a difference between short and long trades with respect to the profit potential.

I would welcome members' thoughts on these matters.

LeviF
Roundtable Knight
Roundtable Knight
Posts: 1428
Joined: Mon Dec 22, 2003 12:24 pm
Location: Des Moines, IA
Contact:

Post by LeviF » Fri Jan 23, 2009 2:42 pm

I can see the potential for not buying ED although:

http://www.bloomberg.com/apps/news?pid= ... refer=home

Why can't crude go to $30....$25....$20....?

ps - I hope your systems were long ED and short CL a long time ago.

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

Post by sluggo » Fri Jan 23, 2009 3:03 pm

The geniuses at Standard & Poors agree. They won't be going long Eurodollars, or going short Crude either.

S&P distributes a "Diversified Trend Indicator" that is a fully disclosed mechanical trading system, applied to a basket of futures contracts. (ref1), (ref2), (ref3). However they have made a conscious decision to exclude Eurodollars from their portfolio -- so they won't be going long ED at these high prices. (Or any others). The DTI does trade Crude Oil but it ignores mechanical signals to short Crude. There's an exception, an override, a special case which says "We can go long Crude but we can't go short". They can short all the other futures in the portfolio; Crude is the only one with this special no-shorting rule.

They froze these rules more than two years ago.

J D Canning
Roundtable Fellow
Roundtable Fellow
Posts: 74
Joined: Thu Jul 03, 2008 6:35 am
Location: Surrey, UK
Contact:

Post by J D Canning » Fri Jan 23, 2009 3:43 pm

My systems did indeed go long ED and short CL some time ago, although I trade on a shorter time frame than many, so have had quite a number of trades in these instruments over the past few months.

My Crude Oil example was somewhat trite - of course it could go lower - many people were using "can't go lower/bargain basement" arguments at $60 and look what happened.

Perhaps a better example would be if it went down to $15 or $10 a barrel... I'm guessing there is an asymptotic relationship between price and potential (diminishing) returns on short trades as commodity prices tend towards zero? Upside payoffs on long trades generally have more potential than short trades - although the short side has obviously yielded some pretty extraordinary returns for trend followers during Q4 08.

drm7
Roundtable Fellow
Roundtable Fellow
Posts: 96
Joined: Sun Apr 20, 2003 9:02 pm
Location: Richmond, VA

Too high/low?

Post by drm7 » Fri Jan 23, 2009 4:39 pm

I think this issue was addressed in the Way of the Turtle book, where he shorted Sugar at some single-digit price and made truckloads of money.

How could he have made money, when Sugar was selling for less than the production cost?

He had the "roll" in his favor. I don't remember the exact prices, but the example goes something like this: near-month is at $5 ("too low"), but next month is $6. $5 goes to $4.50, next month goes from $6 to $5.50. Roll into next month at $5.50, which then goes to $5.00. Roll into next month which is $6, which goes to $5.50. Lather, rinse, repeat.

Oil is extremely contango right now, so shorts are earning a huge roll yield. ED is probably backward, so longs get the roll yield. I'm not saying that ED at close to par is "cheap," but you need to look at the term structure before you make any snap decisions.

alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Re: Discretion

Post by alp » Sat Jan 24, 2009 7:59 am

J D Canning wrote:I would like to bring up the topic of allowing for a certain degree of discretion when trading a mechanical trend following strategy. I know that these two approaches should be mutually exclusive, but we are currently seeing some pretty extreme movements across number of markets, some of which I feel could potentially prohibit the opening of new trades.
I think that there should be very objective criteria to allow discretionary intervention in the system's signals. So I think you might try these approaches until you find the best compromise:

1) improve your system's rules so as to manage the risk of, say, going long ED and short CL;
2) define objective criteria for those instances which cannot be translated into computer code, such as reoptimization time intervals, "exceptional" or very unusual price behaviour, etc.;
3) be very careful about step 2.

Your criteria for intervention have to be very objective. Otherwise you'll never know whether your intervention is truly justified, or you're just not being able to acknowledge the feelings that come up with following your system's signals. If the latter is the case, try to work with your feeligns instead, celebrate and acknowledge them. If not you're very likely to abandon discipline, override signals or drop your system in the moments you should be courageous and nonattached enough not to do so.


Weissman in "Mechanical Trading Systems" does an excellent discussion of the issue of discretionary intervention in mechanical trading systems.

bobsyd
Roundtable Knight
Roundtable Knight
Posts: 373
Joined: Wed Nov 05, 2008 2:49 pm

Post by bobsyd » Sat Jan 24, 2009 4:32 pm

Alp and everyone

Would be very interested in your objective criteria for reoptimisation time intervals. My current thinking is as short as possible taking into account 90% remaining degrees of freedom and having at least 50 trades. Of course the reoptimisation process itself may yield results with less than 50 trades.

Also, if you do Walkforward analysis, how do you determine the length of the optimisation windows and the subsequent walk forward (out of sample) windows? I find Pardo (chapter 11) to be frustratingly vague on this point. Finally, if you trade Intraday as well as EOD, any differences in your protocols?

alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Post by alp » Sun Jan 25, 2009 7:42 am

From Mechanical Trading Systems, by R. L. Weissman:

[quote]Traders and system developers alike must be ever mindful of paradigm shifts in market dynamics. Because markets are rarely stagnant, what worked in the past may not be robust enough to survive dramatic shifts in the dynamics of market behavior as exemplified by our study of natural gas in 2000. Despite our diligence in backtesting and forward (out-of-sample) testing of a wide variety of asset classes and market environments, sometimes unprecedented shifts in market dynamics are too dramatic to enable utilization of previously successful trading systems (examples of such shifts occurred in agricultural markets in the 1970s and in metals markets in 1979 to 1980). In such instances, those who can quickly identify the paradigm shift in market behavior and make the necessary adjustments will outperform the remainder of the pack. This is why the ability to analyze the historical performance of trading systems is invaluable in distinguishing between an equity drawdown within “normalâ€

fourfman
Contributor
Contributor
Posts: 4
Joined: Wed Dec 31, 1969 7:00 pm

Post by fourfman » Wed Feb 04, 2009 12:50 pm

You should have tested your system and now have the trust to trade it.
Few can tread these waters, if it was easy those who can wouldn't make the kind of money we make.
You should trade you system with unflinching trust and when you don't trust the system you should not be trading. Your risk management will take care of you in bad times and make you rich in good.
Test and verify then trust and trade!
fourfman

Post Reply