NexusFi: Find Your Edge


Home Menu

 





How to compare drawdown metrics?


Discussion in Traders Hideout

Updated
      Top Posters
    1. looks_one Koepisch with 3 posts (3 thanks)
    2. looks_two kevinkdog with 3 posts (5 thanks)
    3. looks_3 Quick Summary with 1 posts (0 thanks)
    4. looks_4 vvhg with 1 posts (2 thanks)
    1. trending_up 4,815 views
    2. thumb_up 10 thanks given
    3. group 7 followers
    1. forum 7 posts
    2. attach_file 0 attachments




 
Search this Thread

How to compare drawdown metrics?

  #1 (permalink)
 Koepisch 
@ Germany
 
Experience: Beginner
Platform: NinjaTrader
Broker: Mirus Futures/Zen-Fire
Trading: FDAX
Posts: 569 since Nov 2011
Thanks Given: 440
Thanks Received: 518

Hi,

i want to compare strategies in terms of drawdowns. Currently i have build an underwater chart which visualizes the max. Drawdown in % at a given point of time. If i have a large set of strategy results i can't compare the charts manually anymore.

So here are my final questions:
Is there any common formula or set of a few numbers which can quantify the "risk" from an underwater charts data?
How do you measure drawdowns?

"Risk" should be larger if:
* max. drawdown is higher
* recovering time from drawdown is higher
* the occurence (count) of drawdown peaks are higher
* the distribution of drawdown peaks is erratic

Thanks,
Koepisch

Started this thread Reply With Quote
Thanked by:

Can you help answer these questions
from other members on NexusFi?
ZombieSqueeze
Platforms and Indicators
Deepmoney LLM
Elite Quantitative GenAI/LLM
Are there any eval firms that allow you to sink to your …
Traders Hideout
My NT8 Volume Profile Split by Asian/Euro/Open
NinjaTrader
Better Renko Gaps
The Elite Circle
 
Best Threads (Most Thanked)
in the last 7 days on NexusFi
Get funded firms 2023/2024 - Any recommendations or word …
61 thanks
Funded Trader platforms
39 thanks
Battlestations: Show us your trading desks!
26 thanks
NexusFi site changelog and issues/problem reporting
25 thanks
The Program
17 thanks
  #3 (permalink)
 kevinkdog   is a Vendor
 
Posts: 3,645 since Jul 2012
Thanks Given: 1,890
Thanks Received: 7,338



Koepisch View Post
Hi,

i want to compare strategies in terms of drawdowns. Currently i have build an underwater chart which visualizes the max. Drawdown in % at a given point of time. If i have a large set of strategy results i can't compare the charts manually anymore.

So here are my final questions:
Is there any common formula or set of a few numbers which can quantify the "risk" from an underwater charts data?
How do you measure drawdowns?

"Risk" should be larger if:
* max. drawdown is higher
* recovering time from drawdown is higher
* the occurence (count) of drawdown peaks are higher
* the distribution of drawdown peaks is erratic

Thanks,
Koepisch


You could use something like the Ulcer Index. A good article on that: https://freeman.tulane.edu/trading/pdf/UlcerIndexExplained.pdf

Or, calculating the area of the drawdown curve would incorporate time and magnitude.

A few years back Futures Magazine had a good article about measuring and quantifying drawdowns. I don;t remember exact date or article name though.

The problem, in my experience, is that no one number incorporates everything that may matter to you. For example, is a 1 day drawdown of 10% better/worse than 5 day drawdown of 2%? Both would have the same area under the drawdown curve. Some people care more about magnitude, some care more about time.

Personally, I have found the MAR ratio (annual % return / max % drawdown) to be the best metric that uses drawdown. That's just me of course.

Follow me on Twitter Reply With Quote
Thanked by:
  #4 (permalink)
 Koepisch 
@ Germany
 
Experience: Beginner
Platform: NinjaTrader
Broker: Mirus Futures/Zen-Fire
Trading: FDAX
Posts: 569 since Nov 2011
Thanks Given: 440
Thanks Received: 518

Thanks @kevinkdog,

it's true that a single number can't describe drawdown completely. But you give me enough stuff to think about. It seems that i can enhance the Ulcer index to fit my needs.

Because i want to see the max. risk over a period, i start the strategy run multiple times with the same initial amount at EVERY day in the period (for instance 1 year = 365 runs) and save the max. drawdown in % at every day. Is that the normal way to build an underwater chart? The other way is to let it run only once and calculate the drawdown from the cumulative last high of the account.

Whats a "good" MAR ratio?

Koepisch

Started this thread Reply With Quote
Thanked by:
  #5 (permalink)
 kevinkdog   is a Vendor
 
Posts: 3,645 since Jul 2012
Thanks Given: 1,890
Thanks Received: 7,338


Koepisch View Post
Thanks @kevinkdog,

it's true that a single number can't describe drawdown completely. But you give me enough stuff to think about. It seems that i can enhance the Ulcer index to fit my needs.

Because i want to see the max. risk over a period, i start the strategy run multiple times with the same initial amount at EVERY day in the period (for instance 1 year = 365 runs) and save the max. drawdown in % at every day. Is that the normal way to build an underwater chart? The other way is to let it run only once and calculate the drawdown from the cumulative last high of the account.

Whats a "good" MAR ratio?

Koepisch


I have never calculated dd the way you describe (running it multiple times at every day), so I can't really comment on that method.

Many times I'll use a Monte Carlo analysis to look at it - if I start with $X, and trade for a year, what are the odds that I'll have a maximum drawdown of %Y sometime during the year?

For MAR Ratio, I usually use a 3 to 5 year period for analysis. By going to a reporting service, you can see what CTAs (the "pros") do. Of course, their results are after fees. Nevertheless, comparing yourself to pros is a good benchmark.

1:1 is pretty decent. This would mean that over a 3 year period, I averaged a 30% compounded annual rate of return, with a 30% maximum drawdown. 2:1 would be really good over that same time period. There are some people who get over 3:1.

One fly in the MAR ointment is that it can't predict future drawdowns. Some people with really high MAR ratios might be that may because of hidden risk they are taking that just hasn't been realized. Eventually, they will probably blow out (unless they stop trading first!).

Follow me on Twitter Reply With Quote
Thanked by:
  #6 (permalink)
 
vvhg's Avatar
 vvhg 
Northern Germany
 
Experience: Intermediate
Platform: NT
Trading: FDAX, CL
Posts: 1,583 since Mar 2011
Thanks Given: 1,016
Thanks Received: 2,824

There are many metrics that can be used to describe a system's performance or risk. To name just a few more, there are Sharpe ratio CALMAR ratio SQN by Van Tharp and many more...
Personally I think they all have their value, but it is alays dangerous to look at just one set of data (the original trades), so it is always a good idea to run a Monte Carlo over it. Unfortunately that in itself is a science and an art
A point that @kevinkdog already has made, but can't be stressed enough is to look at hidden risks.

vvhg

Hic Rhodos, hic salta.
Reply With Quote
Thanked by:
  #7 (permalink)
 Koepisch 
@ Germany
 
Experience: Beginner
Platform: NinjaTrader
Broker: Mirus Futures/Zen-Fire
Trading: FDAX
Posts: 569 since Nov 2011
Thanks Given: 440
Thanks Received: 518

Thanks, for you comments. Because i'm not a naitive english speaker it's a little harder for me to describe what i'm doing. "My" method is like monte carlo simulation. But instead of mixing the trades i START the strategy run at different dates (per day granularity) on the exact same set of trades (real trade sequence is still valid). The results are held in ONE drawdown series and the highest value is saved. So i can see the max risk (drawdown) in that period without having the profits distort the picture. A monte carlo version of "my" method would blast my pc. I (think to) know how to use of monte carlo analysis, but because my current strategy (results) is (are) very cyclical it isn't the first choice currently. Cyclical drawdown periods are the best to have - i can easily detect these periods and can filter out a huge portion of it.

"Hidden Risk" are the risk which be hidden if i use only a few numbers as metrics? My initial issue was to filter down the set of strategy results regarding the drawdown. Finally i look always at my underwater chart. I hope there are no hidden risks.

Started this thread Reply With Quote
Thanked by:
  #8 (permalink)
 kevinkdog   is a Vendor
 
Posts: 3,645 since Jul 2012
Thanks Given: 1,890
Thanks Received: 7,338


Koepisch View Post
Thanks, for you comments. Because i'm not a naitive english speaker it's a little harder for me to describe what i'm doing. "My" method is like monte carlo simulation. But instead of mixing the trades i START the strategy run at different dates (per day granularity) on the exact same set of trades (real trade sequence is still valid). The results are held in ONE drawdown series and the highest value is saved. So i can see the max risk (drawdown) in that period without having the profits distort the picture. A monte carlo version of "my" method would blast my pc. I (think to) know how to use of monte carlo analysis, but because my current strategy (results) is (are) very cyclical it isn't the first choice currently. Cyclical drawdown periods are the best to have - i can easily detect these periods and can filter out a huge portion of it.

"Hidden Risk" are the risk which be hidden if i use only a few numbers as metrics? My initial issue was to filter down the set of strategy results regarding the drawdown. Finally i look always at my underwater chart. I hope there are no hidden risks.

When I refer to hidden risk, it means events/situations that haven't occurred yet, and that you might be at risk if they occur...

Example 1: Let's say you are betting martingale style. You start out betting 1 unit, and if you lose, you double your next bet again and again until you win, and then you go back to betting 1 unit. Depending how things go, you might have a nice equity curve, with small drawdowns. BUT, eventually with this method the bet size will get so big you will blow out your account. So, the risk (of ruin) is there all along, you just have not experienced it yet.

Example 2: You write naked options for premium. Depending how you do it, you might make a few percent per month, every month. Equity curve looks nice. Then, bam! - a black swan event happens. You get wiped out. No way you could have seen that coming from just the equity curve or any statistics. Every market crash claims many victims this way - people who thought they had an ATM machine that consistently gave out cash.


So, the point I am trying to make is that there are risk measures which can be used on backtest or historical results. At the same time though, there is a market axiom that says "your biggest drawdown is always in front of you."

Kevin

Follow me on Twitter Reply With Quote
Thanked by:




Last Updated on November 13, 2012


© 2024 NexusFi™, s.a., All Rights Reserved.
Av Ricardo J. Alfaro, Century Tower, Panama City, Panama, Ph: +507 833-9432 (Panama and Intl), +1 888-312-3001 (USA and Canada)
All information is for educational use only and is not investment advice. There is a substantial risk of loss in trading commodity futures, stocks, options and foreign exchange products. Past performance is not indicative of future results.
About Us - Contact Us - Site Rules, Acceptable Use, and Terms and Conditions - Privacy Policy - Downloads - Top
no new posts