Dealing with a drawdown

Discussions about personal psychology for the individual trader.
ksberg
Roundtable Knight
Roundtable Knight
Posts: 208
Joined: Fri Jan 23, 2004 1:39 am
Location: San Diego

Psychology

Post by ksberg »

I note that this is a thread under Trading Psychology, yet we're probably treading on a thread that might belong under System Testing. I know other threads have addressed system measures better than MaxDD. However, for better or worse, I think MaxDD as a measure is here to stay because of the powerful psychological impact of losing money.

There are some mental tricks within LTTF to understand MaxDD in a broader context.

First, MaxDD can be, but most often is not the same thing as drawdown against initial capital. I graph what I call ICapDD (initial capital drawdown), and the values, quantiles, and distribution are distinct from MaxDD, and always less or equal MaxDD. For example, while the prior MaxDD showed 3 in 4 (75%) chance of not seeing a drawdown greater than 42.3%, the ICapDD for the same system stood a 3 in 4 chance of not being greater than 34.5%.In fact, ICapDD can trail off nicely to 0%; the possibility of not ever seeing a drawdown against your initial stake. Hey, that's one piece of good news!

Second, MaxDD is always measured from peak. One trade-off from riding the trend as far as possible is giving the price enough room to expand or contract, which usually translates into giving back substantial profits. So, in LTTF most of MaxDD is the price of riding the trend.

Of course, life pretty much sucks when after the appropriate due dilligence you started trading a new system and it immediately goes into a tailspin. This is where Chris openned the thread.

I think there is another piece of analysis that can help plan and set expectations. I trade by looking for an edge and exploiting it. The edge is a statistical beast, not a certainty. The more I can understand about the nature of that edge, the better I can use it. So, the question is: how long does it take before I can expect the edge to work in my favor, and with what certainty? My action is to commit to that period without fail or hesitation. BTW: research on my current LTTF systems show the high-confidence period to be between 2 to 3 years!!! Better not have ADD! :-)

Cheers,

Kevin

ps. For me, "without hesitation" needs to include pre-arranged staging and kill-switches. Where possible I backtest both with and without those policies in place.
Bernd
Roundtable Knight
Roundtable Knight
Posts: 126
Joined: Wed Apr 30, 2003 6:39 am

Post by Bernd »

:wink:
Last edited by Bernd on Fri Apr 18, 2008 6:24 am, edited 1 time in total.
Asterix
Senior Member
Senior Member
Posts: 44
Joined: Mon Apr 05, 2004 11:16 pm
Location: San Diego

Post by Asterix »

I believe ksberg is pointing out a common error that occurs thru a misunderstanding of statistical analysis. If you run a test on a system and find that a certain size of drawdown only occurs 20% of the time, it does not follow that this steep of a drawdown won't emerge until you've traded the system for a long time.

For example, if you take a coin and flip it 10 times, it is most likely that you will get 5 heads and 5 tails. However, there is nothing to prevent all 10 coin tosses coming up heads. This is not a very likely result, but just because a result is not likely to occur does not mean that it won't occur.

Along the same lines, if you were to flip a coin 10 times and repeat this sequence a 1000 times, probability theory indicates that you should only get one sequence in a 1000 (actually 1024) where you get 10 heads. But there is nothing in the probability theory to indicate when this sequence will occur. It might happen on the first set of 10 coin tosses! In addition, you might actually get as many as 3 to 5 sequences where you get 10 heads rather than the 1 predicted by probability theory.

The same logic applies to your trading results except that things get way more complicated due to the fact that price distributions tend to have fat tails which means that the likelyhood of a major loss is higher than predicted by standard Gaussian statistics.
ksberg
Roundtable Knight
Roundtable Knight
Posts: 208
Joined: Fri Jan 23, 2004 1:39 am
Location: San Diego

Post by ksberg »

Spot on Asterix. And since we're talking about psychology, I believe level-setting expectations is a big part of being able to deal with the results.

Cheers,

Kevin
Chris67
Roundtable Knight
Roundtable Knight
Posts: 1052
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 »

As I opened this thread I thought I better chuck my 2 cents worth in again ..

Firstly .. thanks for some great , level headed and thoroughly helpful comments / analysis.

Unfortunatley for me this drawdown is initial cap. drawdown but thats not really that important as I have the capital to fund it. Its just a bit daunting at the moment with the huge array of so called smart people saying that there are just too many system traders out there.. theres nothing like a 15 % down month to make you question your own theories and thought processes.

So what are we to make of this currently held belief that trendfollowing is going to struggle. Well typically I would say its a great opportunity to chuck more money in as people have been saying this for years.. buy I have 15 years experience trading at a high level and I do think that just about everyone I know now is running a system and it DOES worry me.

However markets will always trend I have no doubt so I believe you need a system that can capture this and I also believe that it will be the players who can accept the 30 % plus drawdowns that will survive since the majority of current system traders are always looking for ways to minimise the drawdowns by being cute and overlaying systems and I believe this will be the real danger area.

Good example is Oil .. if you bought it around 30 $ per barrel as many lttf systems would have indicated then a move from 48 to 42 will not take you out .. many trendfollowers will be out on that move however and this is the danger area...
As rich Dennis said .. either trade very long term or very short term but avoid the middle like the plague

Regards
Chris
leonardo
Roundtable Knight
Roundtable Knight
Posts: 100
Joined: Wed Apr 30, 2003 2:27 pm
Location: Minneapolis, MN
Contact:

Post by leonardo »

Chris67 wrote: However markets will always trend I have no doubt so I believe you need a system that can capture this and I also believe that it will be the players who can accept the 30 % plus drawdowns that will survive since the majority of current system traders are always looking for ways to minimise the drawdowns by being cute and overlaying systems and I believe this will be the real danger area.
I couldn't agree more. Every successful futures trader I've ever known (even including the mature CTAs like John Henry) have had a number of 50% or greater drawdowns which they came back from. I doubt they enjoyed the ride much.

Tenacity is a prized skill.

Leonardo---
FrankWala
Full Member
Full Member
Posts: 22
Joined: Sun Apr 11, 2004 12:32 pm
Location: San Diego, CA

drawdowns

Post by FrankWala »

Most of the work on any system is statistical. How did previous occurences measure up to todays occurences. I tried various systems (from TS) in a portfolio of Holdrs/ETF/index for 1 ,2 ,5,10 years . The systems (momentum, CBO, Volatility, 3 Mov.Avg) hold up in a portfolio well. The NET profit is there but sometimes I go through a DRAWDOWN of 2 to 4 months in historical testing.

Drawdown's are accepting the reality. But how do I know when to switch off / reduce / exit the bet size on the system.

I tried reducing the bet size during the drawdown, std. dev. of equity, (they dont work ). In 95 % of cases the ROI / Sharpe is higher if the system is left untouched.

In a real world..... I know a few people who bought commercial system for $ 1 to 5 k, and they stopped trading, taking signals in a drawdown phase. In some cases the drawdown was in excess of 6 months with a 50 % dip in equity.

In LT testing, the net profit is always there, since the system caught the trend in the early 80's and 90's. There is a steady erosion of returns every year since 1980.

In short, HOW do I know when to know when the system is broke .... a 1 or 2, 6, 9 month drawdown. The flip side is the system is working when higher equity peakes are made, daily, weekly or monthly; -OR- when to hold and when to fold .


Frank, San Diego
*******************************************************
The question remains: what do you do with the information that the probability of MaxDD greater than 42.3% is 25%? Does it boost your confidence in the system? Does it make you believe that the MaxDD of 60% is highly improbable?

I remember myself performing once a Monte Carlo simulation of a LTTF system results using data up until the end of 2002. The result I got was that the chance that MaxDD would reach 60% or more was below 10% (if my memory serves me well). Sure it felt good, as if it just couldn't happen. Then I tested the system on the 2003 data and, sure enough, the equity dropped by 60%. I have not performed a Monte Carlo simulation ever since. But then, I consider myself a newbie in the Monte Carlo world.[/quote]
leonardo
Roundtable Knight
Roundtable Knight
Posts: 100
Joined: Wed Apr 30, 2003 2:27 pm
Location: Minneapolis, MN
Contact:

Re: drawdowns

Post by leonardo »

FrankWala wrote: .... In short, HOW do I know when to know when the system is broke .... a 1 or 2, 6, 9 month drawdown. The flip side is the system is working when higher equity peakes are made, daily, weekly or monthly; -OR- when to hold and when to fold.


Frank, San Diego
A very good question that many have had to deal with.

Logic does have a good answer.

Systems are nothing but a protocol that can be followed if certain conditions exist. If the trading universe which the system is based on changes, or if the system is found to be based on a premise which was thought to be a constant and then is found out to be otherwise--- then stop trading the system. (As an aside, the scariest thing I could imagine would be for the system to continue working for some other unknown reason, perhaps chance; with more and more funds dedicated to it--while the original premise ceases to be valid.)

The reason why long term trading systems have been arguably consistent over many years is because they tend to be based on changes that have a temporal periodicity which closely matches that of the gradual but major changes dictated by the strongest impluses of human behavior.

Shorter term systems applied to markets only operate effectively because of more corrective elements of the human condition.

These corrective elements can take on a variety of faces, and cause confusion to system traders who depend on a recognized recurring market behavior which only serves a purpose for a while and then gets sated by a succeeding behavior.

There are shorter term systems that have worked consistently for decades. They are hardy but relatively low profitability systems based on "knee-jerk" market responses that I doubt will change until humans learn to control themselves from acting impulsively from fear and greed.

Leonardo---
bevok
Full Member
Full Member
Posts: 20
Joined: Tue Oct 20, 2009 5:38 pm
Location: Northland, New Zealand
Contact:

Post by bevok »

Interesting thread, good to see that Aberration Equity Curve from 2004. Although I don't know the composition of that portfolio, I'll bet it would have seen pretty stratospheric growth since then. Theres life in the old dog yet!
Gadoli
Roundtable Fellow
Roundtable Fellow
Posts: 63
Joined: Wed Jul 06, 2005 5:00 pm
Location: Jerusalem

6 years later...follow up

Post by Gadoli »

Hi Chris. you did generate this very interesting thread indeed. I have read quite a few of your more recents threads and gathered that you 're still well alive and kicking in the challenging world of system trading! so you must have survived that particular tough immediate DD you were refering to... However I was curious as to inquire about what happened to this particular system , and whether you decided to keep on trading it? What happened to the real life Max DD compared to your back-tested figures?
Best,
Gadoli.
Post Reply