Advertisement
Calculating the systemic risk of a EA before demo/live launch
Apr 14, 2013 at 13:49
1,024 Views
9 Replies
Apr 14, 2013 at 13:49
Hi everyone,
So, i am currently working on multiple automated strategies and will launch some of them in the near future. My problem is :
How can you assess the systemic risk of 1 EA performing over multiple currencies when in your backtests you can only see what would have happened on 1 currency?
Let's say my EA trades on 10 currencies. How can i know that the total drawdown at X time will not be superior than my total profit at X time?
If in the backtests it's saying for ex: 20% drawdown and 100% profit. If you apply the consecutive loosing trades multiplied by the number of currencies you are trading, so basically adding all the drawdown and measuring if profits are still gonna be available to not let your system crash.
If you understand what i mean, please let me know on how we could calculate this more than IMPORTANT variable for any strategy.
Thanks !
So, i am currently working on multiple automated strategies and will launch some of them in the near future. My problem is :
How can you assess the systemic risk of 1 EA performing over multiple currencies when in your backtests you can only see what would have happened on 1 currency?
Let's say my EA trades on 10 currencies. How can i know that the total drawdown at X time will not be superior than my total profit at X time?
If in the backtests it's saying for ex: 20% drawdown and 100% profit. If you apply the consecutive loosing trades multiplied by the number of currencies you are trading, so basically adding all the drawdown and measuring if profits are still gonna be available to not let your system crash.
If you understand what i mean, please let me know on how we could calculate this more than IMPORTANT variable for any strategy.
Thanks !
Member Since May 11, 2012
16 posts
Apr 15, 2013 at 11:15
(edited Apr 15, 2013 at 11:21)
See http://www.strategyquant.com/eaanalyzer/ , good free tool for analysis.
Member Since Jun 18, 2012
9 posts
Apr 24, 2013 at 07:19
Thats your job. Just run backtests for each currency and see what you would have netted for each month - Run backtest on all 10 currencies. import those tests into myfxbook then add net gain/loss for each month. (ex: add all april results together) - then you'' have a pretty decent idea of how many will be in DD any particular month and what to expect as a net result.
Being Cheesy Ain't Easy...
Apr 24, 2013 at 15:18
You said import into myfxbook ? Is that the strategy tester or how can myfxbook help me i'm not aware of this function..
Also, adding drawdown with drawdown and profits with profits is not correct. SO your method can not be used if wanting to have a profesionnal backtest. Let's say you have 2 currencies in drawdown at the same time, you cannot just simply add the profits that wil result afterwards because those results are based on single currency drawdown and not taking in consideration the other currency's drawdown which means your money management should have changed but here in the calculation you don't take that in consideration which therefore is incomplete...
This is very important because as a full system, we need to know the bahaviour of the other factors such as the money management .
Also, adding drawdown with drawdown and profits with profits is not correct. SO your method can not be used if wanting to have a profesionnal backtest. Let's say you have 2 currencies in drawdown at the same time, you cannot just simply add the profits that wil result afterwards because those results are based on single currency drawdown and not taking in consideration the other currency's drawdown which means your money management should have changed but here in the calculation you don't take that in consideration which therefore is incomplete...
This is very important because as a full system, we need to know the bahaviour of the other factors such as the money management .
Member Since Nov 21, 2011
1601 posts
Apr 28, 2013 at 18:24
This is very simple, It takes 2 minutes to do it.
Step 1
Backtest 1 pair: you get the first drawdown.
Use this pair as reference.
Step 2
Use average correlation to that pair to calculate estimated drawdrown
Step 3
Do it for every pair you wish to trade and add them to the drawdown you backtested.
Step 1
Backtest 1 pair: you get the first drawdown.
Use this pair as reference.
Step 2
Use average correlation to that pair to calculate estimated drawdrown
Step 3
Do it for every pair you wish to trade and add them to the drawdown you backtested.
Member Since Nov 21, 2011
1601 posts
Apr 29, 2013 at 18:05
wahjay posted:
SO you use the average corrolation function in order to get the drawdown of other currencies on your strategy?
Well, this can give you a quick estimated drawdown for your case.
Also, note that a EA is expected to perform for a particular pair and I will not recommend to use it for 10 currencies at a time.
Member Since Nov 21, 2011
1601 posts
Apr 30, 2013 at 21:16
Well, it's up to you off course.
If you think trading 20 pairs at a time is good strategy.. that's your point of view, depending on what does your EA.
By the way, I don't trade 20 pairs at a time, I'd rather to focus a specific currency when one is getting interesting.
When you have focus the good horse... then trade it at a right time. : )
If you think trading 20 pairs at a time is good strategy.. that's your point of view, depending on what does your EA.
By the way, I don't trade 20 pairs at a time, I'd rather to focus a specific currency when one is getting interesting.
When you have focus the good horse... then trade it at a right time. : )
*Commercial use and spam will not be tolerated, and may result in account termination.
Tip: Posting an image/youtube url will automatically embed it in your post!
Tip: Type the @ sign to auto complete a username participating in this discussion.
