Drawdown Reduction Threshold:

Discussions about Money Management and Risk Control.
Post Reply
adamant
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Mon Aug 25, 2008 1:20 am
Location: Norway

Drawdown Reduction Threshold:

Post by adamant » Thu Feb 12, 2009 5:05 am


Could you expand on the false premises of the drawdown reduction threshold?

Alp answered:

As I see it, the drawdown reduction threshold is based on the assumption that one can predict the clustering of winning and losing trades which, by the way, is the same assumption of the "trade if last is a winner" rule. These rules contradict the basic trading system's philosophy of not being able to predict the future. In other words, for me they look like cognitive biases translated into system rules.
I thought this was an interesting question for further disussion. Is this the consensus view of the drawdown reduction threshold? What are counterarguments, if any? Do the benefits of using the DRT outweigh the downside Alp points out? The DRT makes my own, primitive systems more palatable, so I think there is something there. But what I would really like to know are the thoughts of some of the more experienced traders and system designers in here. Any feedback is appreciated.

A

Demon

Post by Demon » Thu Feb 12, 2009 6:52 am

In my opinion DRT reduces overall risk-adjusted performance of most systems but can significantly reduce risk of ruin. As always its swings and roundabouts, there's no right or wrong, but whatever makes you most comfortable with sticking with your system through thick and thin is in my opinion the most important factor.

DeanoT
Roundtable Fellow
Roundtable Fellow
Posts: 52
Joined: Sat Apr 15, 2006 7:37 am
Location: Melbourne, Australia

Post by DeanoT » Thu Feb 12, 2009 7:26 am

You can use the drawdown reduction threshold to do one of two things: 1) increase your position size as a drawdown emerges (like doubling your bets when you lose), or 2) you can trade smaller to avoid an even larger drawdown.

The first scenario significantly increases your risk of ruin, while the second method is a bit like shutting the gate after the horse has bolted. Additionally, trading smaller after a drawdown is likely to stunt your equity growth when more favourable trading conditions return, preventing you from achieving new equity highs.

Either method is a departure from a fixed fractional risk management strategy, which maintains that we should not vary the bet size upon our assumptions about future trading conditions.

If you don't like the drawdown your system is giving you, a better strategy, in my opinion, would be to trade with smaller position size using a fixed fractional strategy, rather than adjusting position sizes based upon current drawdowns.

Another strategy might be to liquidate all positions, and cease trading until you develop a system which does not cause you to change your trading habits based on drawdowns.

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

Post by sluggo » Thu Feb 12, 2009 8:37 am

Francis Bacon wrote:In the year of our Lord 1432, there arose a grievous quarrel among the philosophers over the number of teeth in the mouth of a horse. For 13 days the disputation raged without ceasing. All the ancient books and chronicles were fetched out, and wonderful and ponderous erudition, such as was never before heard of in this region, was made manifest. At the beginning of the 14th day, a youthful friar of goodly bearing asked his learned superiors for permission to add a word, and straightway, to the wonderment of the disputants, whose wisdom he sore vexed, he beseeched them to unbend in a manner coarse and unheard-of, and to look in the open mouth of a horse and find answer to their questionings. At this, their dignity being grievously hurt, they waxed exceedingly wroth; and, joining in a mighty uproar, they flew upon him and smote him hip and thigh, and cast him out forthwith. For, said they, surely Satan hath tempted this bold neophyte to declare unholy and unheard-of ways of finding the truth contrary to all the teachings of the fathers. After many days more of grievous strife the dove of peace sat on the assembly, and they as one man, declaring the problem to be an everlasting mystery because of a grievous dearth of historical and theological evidence thereof, so ordered the same writ down.
Perhaps a good idea even today. Rather than postulating and imagining and suggesting the effects that Drawdown Reduction Threshold should have, why not run some experiments and measure the impact it actually does have? If this is really important to you, fire up a backtesting simulator and run some tests, possibly along the lines sketched below.
Attachments
counting_the_horses_teeth.png
counting_the_horses_teeth.png (10.39 KiB) Viewed 10510 times

adamant
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Mon Aug 25, 2008 1:20 am
Location: Norway

Post by adamant » Thu Feb 12, 2009 8:49 am

I am doing exactly that Sluggo. Also, I am not trading any systems in any market, btw. Merely trying to learn at the moment.

I am not blind to the fact that in order to learn, you must test! However, as I am new to system design and trading, I value also discussions of principles. Insights and a better understanding of principles as applied to trading may steepen my learning curve and help me avoid (some) stupid mistakes and dead ends. Maybe you will agree that something productive may come from discussion of principles also, I don't know. I do see your point though.

As I mentioned, DRT has a positive effect on my systems so far, in my opinion, these systems become easier to trade, on paper (I have run the test you propose, obviously).

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

Post by sluggo » Thu Feb 12, 2009 9:28 am

Excellent! Congratulations and best wishes.

For discussion, here is a postulate which may or may not be true (I don't know which):
  • With various positionsizing techniques, you can manipulate drawdown depth (measured in %), or you can manipulate drawdown duration (measured in days), but you can't manipulate both at the same time. In particular, the area under the drawdown curve is approximately constant. Therefore, anything that reduces drawdown depth will increase drawdown duration. And anything that reduces drawdown duration will increase drawdown depth.
You could write code to measure the area under the drawdown curve (in Blox, an aux blok using test.AddStatistic, in Wealth-Lab, a perfscript) on various trading systems or suites of trading systems, while you apply positionsizing techniques such as DRT, anti-DRT, and so on. You could see for yourself whether the area is ~~ constant, or not. If so, that result would be a bit discouraging, for it would say that pain is conserved. (You can suffer less, for longer); or (you can suffer more, for less time); but those are the only choices.

Your code would identify the piece of the equity curve that includes the maximum DD, and calculate the sum of DD% for each day in that drawdown, from beginning of DD to end (the day before a new equity high). Voila, you've got the area. Congratulations, a calculus professor would call it "The Riemann sum approximation to the integral." See whether it remains ~~ constant as you run the system many times with many different parameter values of the DRT, anti-DRT, or whatever rude medicine is being administered.

gunter
Roundtable Fellow
Roundtable Fellow
Posts: 83
Joined: Wed Aug 24, 2005 7:25 am
Location: George Town, Cayman Islands

Post by gunter » Thu Feb 12, 2009 11:00 am

Great post Sluggo,

I have never viewed the area under the drawdown curve, nor attempted to calculate it. I might do so for future backtests.

However, in all my live trading, I found that trying to use DRTs just led to the drawdown becoming longer. Shortly after the drawdown reached its maximum point, some nice trends would appear and I would trade them with greatly reduced size.

In backtests, doing this usually looked appealing, however living through extended drawdowns became extremely frustrating. Others might be happy accepting longer drawdown periods if this reduces their drawdowns.

Perhaps another way of looking at this might be: What would I hope to accomplish with using DRTs and how does this fit in with my own investment strategy?

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Thu Feb 12, 2009 12:07 pm

Francis Bacon wrote:In the year of our Lord 1432
What a simply wonderful quote. Theologians are very silly monkies. I suppose you get that way if you endlessly conjure up meaningless "proofs" of the existence of God. At least the likes of Hildegard of Bingen believed they had touched the metaphorical robe and thus, to them at least, their proof was empirical.

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

Post by alp » Thu Feb 12, 2009 8:39 pm

I see a dialetical, often complementing process between reason and faith, evidence and theory. Moreover I suppose that empiricism also implies learning from other's mistakes.

adamant
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Mon Aug 25, 2008 1:20 am
Location: Norway

Post by adamant » Fri Feb 13, 2009 2:47 am

Moreover I suppose that empiricism also implies learning from other's mistakes.
I hope you're right Alp. This would certainly save me some time...I am sure I will make plenty on my own.

Roscoe
Roundtable Knight
Roundtable Knight
Posts: 250
Joined: Sat Jan 24, 2004 2:06 am
Location: Houston TX

Post by Roscoe » Sun Feb 15, 2009 6:44 pm

I don't know if this helps any but I have observed a linear relationship between drawdown and return and that relationship has remained fairly consistent across many hours of testing. The attached chart shows the effect of increasing risk on return and drawdown in a test portfolio. It should be noted that all MarketSystems in the portfolio have a positive expectation when the largest winner has been discounted. Your mileage may vary.

Taking the other view, any and all attempts to reduce drawdown consistently result in reduced return in my testing.

No grand theory to speak of, merely the result of looking in the mouths of many horses.
Attachments
!Heat_001.gif
!Heat_001.gif (15.96 KiB) Viewed 10280 times

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

Post by sluggo » Sun Feb 15, 2009 9:39 pm

Lovely curves Roscoe, thanks for sharing them.

On a tiny scale, I did a study that was sort of similar, looking only at the Blox presupplied "Donchian" system across a large range of different Heat values, from ridiculously low to fatally high. Experimental setup at >link1< and plot of results (curves a bit like yours), at >link2<

Asamat
Roundtable Knight
Roundtable Knight
Posts: 175
Joined: Fri Jun 03, 2005 7:50 am
Location: Walldorf, Germany

Post by Asamat » Mon Feb 16, 2009 7:39 am

Hi Roscoe,

nice curves indeed. Surely you did inquire into the nature of those two bumps at 70 an 85, didn't you? Can you relate any possible findings? Maybe additional markets being accessible?

Regards,
Asamat

Roscoe
Roundtable Knight
Roundtable Knight
Posts: 250
Joined: Sat Jan 24, 2004 2:06 am
Location: Houston TX

Post by Roscoe » Mon Feb 16, 2009 5:22 pm

Asamat wrote:Surely you did inquire into the nature of those two bumps at 70 an 85
No, I was really just interested in the relationship between DD and return and a few other things (the chart in my post has been trimmed to show just DD and return) and I don't tend to dissect test data to any fine degree as I am looking for a broader picture, but your guess is probably the right answer.

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

Post by alp » Mon Feb 16, 2009 10:25 pm

Roscoe wrote:No, I was really just interested in the relationship between DD and return and a few other things.
Roscoe, did you test the drawndown reduction threshold rule or instead only the relationship between bet size and return? The original Turtle system concept works like a changing bet size, i.e., as the drawdown gets bigger, the percentage bet size is decreased (indirectly by limiting the account equity upon which to size the bet).

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

Post by alp » Mon Feb 16, 2009 10:29 pm

gunter wrote:However, in all my live trading, I found that trying to use DRTs just led to the drawdown becoming longer. Shortly after the drawdown reached its maximum point, some nice trends would appear and I would trade them with greatly reduced size.
That's my point.

Roscoe
Roundtable Knight
Roundtable Knight
Posts: 250
Joined: Sat Jan 24, 2004 2:06 am
Location: Houston TX

Post by Roscoe » Mon Feb 16, 2009 10:39 pm

alp wrote:did you test the drawndown reduction threshold rule or instead only the relationship between bet size and return?
Just the relationship in the chart shown. I agree with Gunter.

adamant
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Mon Aug 25, 2008 1:20 am
Location: Norway

Post by adamant » Tue Feb 17, 2009 5:19 am

Thanks to everyone for sharing their experiences/thoughts with/on DRT.

gunter
Roundtable Fellow
Roundtable Fellow
Posts: 83
Joined: Wed Aug 24, 2005 7:25 am
Location: George Town, Cayman Islands

Post by gunter » Tue Feb 17, 2009 2:56 pm

One of the insights I got from my experience was that if I feel uncomfortable with the drawdowns, I should rather look at diversifying more and reducing the overall portfolio risk. Using DRTs to try to trade aggressively, while still keeping drawdowns low did not work as well as I had hoped.

Perhaps learning to become more comfortable with large drawdowns might work! :shock:

Post Reply