Max DrawDown

Created at 05 Nov 2012, 15:02
How’s your experience with the cTrader Platform?
Your feedback is crucial to cTrader's development. Please take a few seconds to share your opinion and help us improve your trading experience. Thanks!
RK

rkokerti

Joined 28.06.2012

Max DrawDown
05 Nov 2012, 15:02


Dear Support,

Could you please explain me how Max DrawDown is calculating in cAlgo backtest table?

I have not been reproduce it in MSExcel based on data. 

Thanks!


@rkokerti
Replies

admin
05 Nov 2012, 18:12

The max drawdown is calculated based on the following:

Max Drop of Long Consequtive Losing Trades after a winning trade(long or short)

((Balance after winning trade(long or short) - Balance after last loosing long trade)/Balance after winning trade(long or short) )*100

Max Drop of Short Consequtive Losing Trades after a winning trade(long or short)

((Balance after winning trade(long or short) - Balance after last loosing short trade)/Balance after winning trade(long or short) )*100

For example:

Number time Type Volume Commission S/L T/P Price Profit Balance
1 19.01.12 - 20:30 Buy 100000 6 1.53411 1.53411 1.45231 144 10144
2 19.01.12 - 20:31 Sell 100000 6 1.53411 1.53411 1.45231 -106 10038
3 19.01.12 - 20:32 Buy 100000 6 1.53411 1.53411 1.45231 -56 9982
4 19.01.12 - 20:33 Buy 100000 6 1.53411 1.53411 1.45231 -6 9976
5 19.02.12 - 20:30 Sell 100000 6 1.53411 1.53411 1.45231 344 10320
6 19.02.12 - 20:31 Buy 100000 6 1.53411 1.53411 1.45231 1494 11814
7 19.02.12 - 20:32 Buy 100000 6 1.53411 1.53411 1.45231 294 12108
8 19.03.12 - 20:30 Sell 100000 6 1.53411 1.53411 1.45231 -2006 10102
9 19.03.12 - 20:31 Sell 100000 6 1.53411 1.53411 1.45231 -156 9946
  Max Drop of Long Consequtive Losing Trades after a winning trade(long or short) Max Drop of Short Consequtive Losing Trades after a winning trade(long or short) Max Drop of all Consequtive Losing Trades after a winning trade(long or short)
Max DrawDown 1.66 17.86 17.86
Max. Drop of balance in % after a winning trade  ((10144 - 9976)/10144)*100 ((12108-9946)/12108)*100  

 

 

 


@admin

rkokerti
06 Nov 2012, 01:51 ( Updated at: 21 Dec 2023, 09:20 )

Dear Support,

Thanks for your quick answer. 

After analyzed your example, I understood why I could not reproduced the Max DrawDown in Excel. The reason was the approach.

According to definition as far as I know, the Drawdown is the measure of the decline from a historical peak. The key word is the historical. I think it’s very important, that trader doesn’t loose money that has earned in a given period of time (backtest period).

But, cAlgo continually restart the calculation of Drawdown. So it’s like a block-calculation and not a historical one. And it’s very confusing and misleading for me.

I attached a table below that shows the difference between the result of my calculation and cAlgo. Unfortunately there are significant differences.

Max DD - cAlgo calculation for all: 6.06%

Max DD - My calculation for all: 12,47%

 

What is your opinion about it?

Thanks


@rkokerti

admin
06 Nov 2012, 12:46

Hello,

 

The max drawdown is defined as the difference of the value at the peak before largest drop minus the lowest value before a new high divided by the value at the peak before largest drop.

So, a drawdown is established at a trough right before a new peak.

What you are describing is the percentage difference of the maximum value minus the minimum value.

We will have to think about implementing something like what you are describing but that will be somehing different that max drawdown.

 


@admin

rkokerti
06 Nov 2012, 15:17 ( Updated at: 21 Dec 2023, 09:20 )

Hello,

Thanks for your answer!

But, as I mentioned this is a very important index-number in a risk management point of view. So, I would like to know what is the measure of the decline, from a historical peak (when I reached a new higher-high in account balance)

see below the last bold arrow on picture:

If it’s possible please implement it into the backtest result table.

Thanks for your help!


@rkokerti

admin
06 Nov 2012, 17:03

Hello,

 

Thank you for your suggestion. We understand the importance of this and we are taking it into serious consideration to add such a calculation in the backtesting environment. Most probably it will be added very soon.

 

 

 


@admin

rkokerti
30 Nov 2012, 17:16

Thanks for the correction. Now it looks good!

Thanks again!


@rkokerti