How much drawdown is too much?

[Strw23]

Hello All

How much Drawdown is too much, absolute dollar drawdown and drawdown as percentage of proffit?

I am currently adjusting a system I am developing to try and get the best result I can. I am adjusting the system by changing formula parammeters but leaving the money management alone. The problem I am facing is that I can get huge profits but this also increases my drawdown. This brings me to my question

*How much drawdown is too much?
*Is there some ratio that people like to use?
*At What point do people say I will cap my potential profits in order to limit my drawdown.

Would be interested in other peoples thoughts and ideas.

Scott

 

[bullmarket]

Hi Strw23

That's a difficult question to answer with actual numbers using the info you supply in your post.

We don't drawdown on our capital but if we had to then as a starting point we would model various drawdown scenarios to ensure that after any drawdown the remaining capital would still be able to generate sufficient returns, within our risk tolerances, to at least meet our future income requirements and to also hopefully regain the drawdown eventually.

Obviously the last thing we would want to do is draw down to the point where we would have to take uncomfortable risks with the remaining capital to generate our required returns.

Now how you model the drawdown scenarios and what criteria and parameters you use boils down to individual circumstances.

Hopefully maybe someone who draws down will be able to give a better indication based on what they do.

Anyway, just food for thought

cheers

bullmarket

 

[happytrader]

I was taught to closely observe a 3% move either up or down as a high probability indication of trend change. This is assuming you are trading with the trend. Of course this is general.

Maybe you could look at your stocks behaviour with a bit more than the usual attention. Be a Specialist.

Cheers
Happytrader

 

[tech/a]

strw 23

Initial drawdown is that which a method can expect to use whilst setting itself with trades which trade beyond their purchase price and away from drawdown territory.

If we are trading a portfolio of 10 stocks if they were each bought on the same day with an expected 50% success rate IE winners v losers 5 could well be losers,however string of consecutive losses is also very important as you could win the lottery and ofcourse get that string in your initial start up period.If thats 9 (and I have seen more in methods I use) straight then that must always be considered.
So you may risk 1% on your stop setting each trade.
I use this IE 10 trades 10% of purchase price is my stop,Equal parcel sizes = 1% initial risk on each trade.


Techtrader has an initail drawdown of 7% using this method and string of losses is 9.So if my initial drawdown ever was 10% then I would have a system that was out of sample.
It actually only ever made about 4% but had open profit well in excess of the stopped trades.
Remember that a 5% loss in initail capital equates to a 10% required profit from there to breakeven with only 90% of capital.
If your trading margin or CFD's then thats multiplied by the leverage against initial capital.

So if trading CFD's with T/T I would have had a drawdown of 40% at 10:1 and only 60% of initial capital to pull myself out of the fire.
If I had the 9 losses straight then 90% loss of capital---basically un workable--not due to the method but the leverage used.

What is acceptable. Well of all the methods I have tested and the 3 I use longterm under 10% is the aim.Less is better.

Peak to Valley Drawdown is the maximum drawdown from the highest peak to its corresponding lowest trough following that peak.
This one is a balance as to achieve maximum gains you need to give a method enough room to stay in good trades during periods of correction without giving back the majority of your profit.
Testing gives you confidence in the smoothness of your equity curve.You wouldnt want a curve with drawdowns most months.Smoother the better.
Techtrader is around 15% and it came close to that in April last year.
It has proven to be within parameters and the method has kept powering along and has reclaimed all drawdown and more.Trading outside the maximum drawdown would be instant closure of the trading method and its portfolio.

As your in the process of testing I hope these parameters are helpful.

 

[Strw23]

Thanks to all those that replied.

The problem I currently have is that my latest system is making on average 1000% return over the last 4.5 years after costs without compounding but the initial drawdown in the first few months is about 65%. NB this is using leveraged products. Clearly this is too much. Its nice to say that looking at it on paper I could handle that because I know what profits are coming but if I was trading this they would be some very sleepless months. I guess I will have to try and lose a large chunk of profit to minimise my drawdown. I have played around with the money management but then the system in its current form is not consistant enough.

Thanks again

Scott

 

[bullmarket]

Hi Strw23

I see you say that you have tested your system over the last 4.5 years.

But as I'm sure you are aware, since March 2003 we have essentially been in a roaring bullmarket. That means your system is likely to be successful in bull markets.

But, just as a thought, have you had a chance to test your system over market data during periods where there have been significant negative market performances over various periods - say in 2000 - 2003 for example or any other period that suits.

Anyway, just an idea and food for thought.

cheers

bullmarket

 

[tech/a]

Scott

Its ugly.

 

[Strw23]

BullMarket, the reason I have only tested it over the last 4.5 years is that it is for sectors and there is only decent data for that time. Before 22/6/2001 its just OCHL all at the one point. They way it is configured at present it gives 265 trades and 38 of them are short. I can get just the short trades upto 700% but still with 50% DD.

Tech/a I know its ugly and untradeable. If tonight wasnt a good TV night I wouldnt get off the computer.

Scott

 

[bullmarket]

ok no problem Strw23

 

[Strw23]

tech/a
got drawdown down to 13% while still remaining profitable, just have to play around with different time periods over the last 4.5 years to see if it holds up and is not to curve fitted.

Bullmarket
on a completely different subject your name would be a good name for a boat as aussie rule is already taken.

Scott

 

[lesm]

Scott,

You may like to take a look at Chapter 6 from 'Trading as a Business', I have included the URL to an online version below:

http://www.elitetrader.com/tr/index.cfm?s=17&t=77

You may find some the principles put forward in this chapter of assistance.

Have you looked at the charts to identify the trades/chractersitics that have resulted in the large drawdowns?

Looking at the charts can assist in developing an approach that is tradeable and acceptable.

Cheers.

 

[tech/a]

tech/a
got drawdown down to 13% while still remaining profitable, just have to play around with different time periods over the last 4.5 years to see if it holds up and is not to curve fitted.

Scott

Scott thats fine for an un leveraged system but at 10x leverage you're dead.

Thats the trouble with Hometrader---what they dont teach you.
Their intentions and basic idea is sound but in practice as many find its not easy. Many blow up.

Ever wondered why the same names never appear twice for winning their comp?

Anyway you're getting there.

 

[Strw23]

Tech/a
I will be trading sectors with the system so its 100:1 leverage. The 13% drawdown is also after leverage so it shouldnt be a problem. The coaches go through all these things with you. As for not seeing the same names up on the winning list that is not correct. Alot of the strong performers are up there on a regular basis, if you want a copy of the last few years results I can get them for you.

Scott

 

[Strw23]

lesm, thanks for the link I will sit down later and have a good read of that.

Scott

 

[tech/a]

Tech/a
I will be trading sectors with the system so its 100:1 leverage. The 13% drawdown is also after leverage so it shouldnt be a problem. The coaches go through all these things with you. As for not seeing the same names up on the winning list that is not correct. Alot of the strong performers are up there on a regular basis, if you want a copy of the last few years results I can get them for you.

Scott

Well that wasnt the case when I checked into the Adelaide operation.
But have to say,although not a student and with no interest in them they seemed genuine enough.

Main thing I guess is that your consistently on the list.
Good luck and interested in your long term results.

 

[bullmarket]

hi Scott

Bullmarket
on a completely different subject your name would be a good name for a boat as aussie rule is already taken.

Scott

hmmmmm.............I see what you mean

Greg Norman's boat was named Aussie Rules - I'm not sure if the name was retained after it was sold.


cheers

bullmarket

ps...it's not my boat

 

[Strw23]

F@#ken typos.
No "I told you so's" please Tech/a but I found a coding error in my tradesime code. I had written a C instead of an H for part of my formula for my intraday short trailing stop. This killed all profits. I have sinced changed to a different stop and got the profit back to more than bafore and reduced my drawdown to 3%. The only problem I have noticed when going through my trade log is that the fixed position size money management I was using is giving me to much risk per trade. If I didnt have a heap of winning trades at the start I would blow up my account. Hence again it is untradeable. Back to the drawing board to reduce profits.

Scott

 

[tech/a]

Scott

You wont get I told you so from me.
Thats what testing is all about.
Its happened to all of us who design trading methods regularly.
Being able to discuss and play with ideas with like minded people is what its all about.

Due to the short time frame you need more wins than losses as profits dont have a chance to run,so just bear this in mind when designing.