Sounding Board and Random Brainstorming

arsh22g

Well-Known Member
#91
Here is the backtest from Jan 2016 - Mar 2016 (3 months). 75 Nifty shares (1 lot) for 1 Lakh capital. I am worried about the K-ratio. I read about it and literature suggests it should be >2. I haven't seen any system in Traderji or outside which has >0.3. Does Ami calculate it differently from standard ones, so that their scale is different from others?



First day of live testing on different TFs. Here are the results till now for 1 lot NF (12:10 PM):

1-min : Net = -4500 | Tr Cost = 9667 , 13 trades
3-min : Net = +1192 | Tr Cost = 4297, 6 trades
5-min : Net = +1500 | Tr Cost = 715, 2 trades
15-min : Net = +2250 | Tr Cost = 715, 2 trades

I am assuming 0.03% as tr cost of average trade value.

Net is after deducting tr cost. So, in 1-min, tr costs are more than the profit.
 

arsh22g

Well-Known Member
#92
One query regarding Amibroker backtesting.

Backtesting has two parameters by the name Max Trade Drawdown and Max System Drawdown. Official Amibroker definition is as follows:

HTML:
Max. trade drawdown - The largest peak to valley decline experienced in any single trade. The lower the better

Max. trade % drawdown - The largest peak to valley percentage decline experienced in any single trade. The lower the better

Max. system drawdown - The largest peak to valley decline experienced in portfolio equity. The lower the better

Max. system % drawdown - The largest peak to valley percentage decline experienced in portfolio equity. The lower the better
I was manually checking these backtest report paramaters against the trades. But they are not tallying. Where can I be wrong?
 

suri112000

Well-Known Member
#93
One query regarding Amibroker backtesting.

Backtesting has two parameters by the name Max Trade Drawdown and Max System Drawdown. Official Amibroker definition is as follows:

HTML:
Max. trade drawdown - The largest peak to valley decline experienced in any single trade. The lower the better

Max. trade % drawdown - The largest peak to valley percentage decline experienced in any single trade. The lower the better

Max. system drawdown - The largest peak to valley decline experienced in portfolio equity. The lower the better

Max. system % drawdown - The largest peak to valley percentage decline experienced in portfolio equity. The lower the better
I was manually checking these backtest report paramaters against the trades. But they are not tallying. Where can I be wrong?
Amibroker backtest report is not reliable, it seems. Unless we know what the code is meant to do, it is very difficult to rely on backtest reports. There is lot of difference between manual testing and machine testing especially when we want to use some filters and commission reducing etc.

We cannot risk our money purely on backtest report of amibroker.:D
 

arsh22g

Well-Known Member
#94
Amibroker backtest report is not reliable, it seems. Unless we know what the code is meant to do, it is very difficult to rely on backtest reports. There is lot of difference between manual testing and machine testing especially when we want to use some filters and commission reducing etc.

We cannot risk our money purely on backtest report of amibroker.:D
Completely agree with you, that's why I always take a sample of trades to test manually.

Coming back to my specific query, is it possible that the drawdown shown by Amibroker is what any specific position has experienced during the duration of backtesting, but that position eventually got closed at a different level?

If such is the case, then the variation is huge.
 

mechtrader

Well-Known Member
#95
Completely agree with you, that's why I always take a sample of trades to test manually.

Coming back to my specific query, is it possible that the drawdown shown by Amibroker is what any specific position has experienced during the duration of backtesting, but that position eventually got closed at a different level?

If such is the case, then the variation is huge.
The DD shown by amibroker is not on trade close to close basis. It is on MTM basis. eg: your position goes 50 pts against you and finally closes 30 pts against you. Ami will show trade DD as 50 pts, whereas you realized 30 pts.

MT
 

mechtrader

Well-Known Member
#97
Mechtrader, As you have a lot of experience in backtesting and algo trading.
Can you please point out some points what we should take care when we do amibroker backtesting.

thanks
This is a very general query. I will be happy to help for any specific query. You can also read out my thread if you have not done so yet. I have explained alot of questions there.

MT
 

vijkris

Learner and Follower
#98
The DD shown by amibroker is not on trade close to close basis. It is on MTM basis. eg: your position goes 50 pts against you and finally closes 30 pts against you. Ami will show trade DD as 50 pts, whereas you realized 30 pts.

MT
thanks mechtrader for visiting this thread.:rofl:
your experience will be quite helpful for us.
 
#99
This is a very general query. I will be happy to help for any specific query. You can also read out my thread if you have not done so yet. I have explained alot of questions there.

MT
:)
I am a regular follower of your thread and whenever you post something I read very deeply as I know there is something to learn like u posted about optimization about parameters
.:thumb:

Ok... Will ask if any specific query.

thanks
 
Last edited:

vijkris

Learner and Follower
:)
I am a regular follower of your thread and whenever you post something I read very deeply as I know there is something to learn like u posted about optimization about parameters
.:thumb:

Ok... Will ask if any specific query.

thanks
bro, today u got lot of profits kya? it seems u r very very happy..u r not telling the reason ;)
 
Last edited:

Similar threads