Dukascopy Support Board :: View topic - Signal Provider Contest Calculations

The problem is not how the average is calculated at Dukascopy contest.
The problem is that an average method of calculation will not help determine and rank accurately consistent trading systems.

I agree with your suggestion.

according to Dukascopy's current methodology [(25+100+75+150+50+800)/6], his "average profitability" will be 200%, only due to one very lucky month!

Average Profitability (%) = (Total P&L / (100,000 * n)) * 100%, where n is the number of months in the respective period, including the current one.

in the real world, we don't get a fresh $100000 at the start of every month

What I suggest for the future is to modify the performance calculations. Perhaps a possibility would be replacing the average monthly profitability with a monthly median profitability.

to alltrade:

Quote:

according to Dukascopy's current methodology [(25+100+75+150+50+800)/6], his "average profitability" will be 200%, only due to one very lucky month!

your calculation method is not right indeed. The correct one can be found in the Contest rules.

Quote:

Average Profitability (%) = (Total P&L / (100,000 * n)) * 100%, where n is the number of months in the respective period, including the current one.

As you see the PnL is the key element of the formula but not the equity.

Quote:

in the real world, we don't get a fresh $100000 at the start of every month

that is right unless you are a trader of the big financial institution. Assume that you are the trader in the big hedge fund or bank and 100,000 USD is your loss limit. You have to reach maximum profit using this capital. For more details please read the Trader contest philosophy.

Author:	AndyWeiss [ Thu 05 Dec, 2013, 04:33 ]
Post subject:	Signal Provider Contest Calculations
Hello all, I was browsing the contests section and saw the signal provider contest, I was glad to finally find a contest that promotes consistency and rewards the best long term performance. The objective of this contest stated by Dukascopy " is to determine the most reliable and consistent Signal Provider of Dukascopy Forex Community." But after looking at many of the top traders, I've noticed that a lot of them had a single good month or wining streak that put them in that position. Although those traders show the ability to make high profits, they do not show the ability to reproduce such performance. In other terms they do not show consistency, which is one of key element for a signal follower to consider. Therefore the ranking may then be misleading for some followers. What I suggest for the future is to modify the performance calculations. Perhaps a possibility would be replacing the average monthly profitability with a monthly median profitability. This change would promote consistent high performance and eliminate flukes in monthly performance. Please show your support if you believe in a fairer calculation method. Andy

Author:	alltrade [ Sun 08 Dec, 2013, 13:02 ]
Post subject:	Re: Signal Provider Contest Calculations
I'm glad somebody else other than me noticed the MASSIVE issue with how the so called "profitability" is calculated in Signal Provider Contest, which is supposed to reward reliable & consistent trading! Let's look at an example, let's say a trader finished his 6 months in the Signal Provider Contest with balances of 25k, 100k, 75k, 150k, 50k, 800k. Of course, according to Dukascopy's current methodology [(25+100+75+150+50+800)/6], his "average profitability" will be 200%, only due to one very lucky month! But the REAL profitability of the said trader can be calculated through a simple calculation like this one : [(100000 * 0.25 * 1 * 0.75 * 1.5 * 0.5 * 8) - 100000] / (1000 * 6) = 2.08% So, as we can see, the assumed trader, who shows an average profitability of 200% according to Dukascopy's methodology but his REAL average profitability is a mere 2.08%

Author:	AndyWeiss [ Mon 09 Dec, 2013, 01:40 ]
Post subject:	Re: Signal Provider Contest Calculations
Thank you for your reply alltrade. I agree with you when you said.. due to one lucky month the average profitability is boosted. But, let me reiterate my point. The problem is not how the average is calculated at Dukascopy contest. The problem is that an average method of calculation will not help determine and rank accurately consistent trading systems. (Also, I have to tell you that your proposed calculations are wrong. Ask me in PM if you need me to explain.) Regards,

Author:	alltrade [ Mon 09 Dec, 2013, 11:09 ]
Post subject:	Re: Signal Provider Contest Calculations
AndyWeiss wrote: The problem is not how the average is calculated at Dukascopy contest. The problem is that an average method of calculation will not help determine and rank accurately consistent trading systems. The two sentences contradict each other. If the objective is to reward reliable & consistent signal providers & the current calculation isn't achieving it then obviously, the current method of calculation IS THE problem. Nope, calculations aren't wrong, may be you haven't quite grasped what the calculations intend to accomplish. Their purpose is to take the 100000 of the first month of the 6/12-month-period & then treat the profits/losses during subsequent months as if the trader had continued trading with the original capital. That's the only way consistency & profitability of a good signal provider can be judged because in the real world, we don't get a fresh $100000 at the start of every month.

Author:	alifari [ Mon 09 Dec, 2013, 14:05 ]
Post subject:	Re: Signal Provider Contest Calculations
Quote: Nope, calculations aren't wrong, may be you haven't quite grasped what the calculations intend to accomplish. Their purpose is to take the 100000 of the first month of the 6/12-month-period & then treat the profits/losses during subsequent months as if the trader had continued trading with the original capital. That's the only way consistency & profitability of a good signal provider can be judged because in the real world, we don't get a fresh $100000 at the start of every month.[/quote] I agree with your suggestion.

Dukascopy Support Board http://www.dukascopy.com/swiss/english/forex/jforex/forum/

Signal Provider Contest Calculations http://www.dukascopy.com/swiss/english/forex/jforex/forum/viewtopic.php?f=48&t=50464	Page 1 of 1

Author:	alltrade [ Tue 10 Dec, 2013, 05:51 ]
Post subject:	Re: Signal Provider Contest Calculations
alifari wrote: I agree with your suggestion. Thanks, brother!

Author:	oleksandr_n [ Wed 18 Dec, 2013, 17:35 ]
Post subject:	Re: Signal Provider Contest Calculations
to alltrade: Quote: according to Dukascopy's current methodology [(25+100+75+150+50+800)/6], his "average profitability" will be 200%, only due to one very lucky month! your calculation method is not right indeed. The correct one can be found in the Contest rules. Quote: Average Profitability (%) = (Total P&L / (100,000 * n)) * 100%, where n is the number of months in the respective period, including the current one. As you see the PnL is the key element of the formula but not the equity. Quote: in the real world, we don't get a fresh $100000 at the start of every month that is right unless you are a trader of the big financial institution. Assume that you are the trader in the big hedge fund or bank and 100,000 USD is your loss limit. You have to reach maximum profit using this capital. For more details please read the Trader contest philosophy. to AndyWeiss: Quote: What I suggest for the future is to modify the performance calculations. Perhaps a possibility would be replacing the average monthly profitability with a monthly median profitability. please show us the formula of you calculation method.

Author:	alltrade [ Fri 20 Dec, 2013, 06:29 ]
Post subject:	Re: Signal Provider Contest Calculations
oleksandr_n wrote: to alltrade: Quote: according to Dukascopy's current methodology [(25+100+75+150+50+800)/6], his "average profitability" will be 200%, only due to one very lucky month! your calculation method is not right indeed. The correct one can be found in the Contest rules. Quote: Average Profitability (%) = (Total P&L / (100,000 * n)) * 100%, where n is the number of months in the respective period, including the current one. As you see the PnL is the key element of the formula but not the equity. You are right, the average profitability would be 100% instead of the stated 200%! But is that really relevant to the point I have made? No. My point still stands that the way calculation is done doesn't tell us about the true profitability of the trader as a "signal provider" as considered by most people. oleksandr_n wrote: Quote: in the real world, we don't get a fresh $100000 at the start of every month that is right unless you are a trader of the big financial institution. Assume that you are the trader in the big hedge fund or bank and 100,000 USD is your loss limit. You have to reach maximum profit using this capital. For more details please read the Trader contest philosophy. I have already read it but that just leads to rewarding rampant gambling as can already be seen & not "reliable & consistent signal provider" as the stated Objective on the contest's page; while the method I have described will reward signal providers who are showing consistent profitability over a period of time. Not to mention, traders at big financial institutions have a whole host of rules to follow in order to control risk & gambling which isn't the case with demo-money so that's like comparing apples & oranges. If the contest were titled Institutional Trader Contest then one might see some sense in the current methodology but as it is called Signal Provider Grand Prix, the average "profitability" figure is highly misleading considering how a typical "Signal Provider" would be judged.