CFDs are complex instruments and come with a high risk of losing money rapidly due to leverage. 66 % of retail investor accounts lose money when trading CFDs with this provider. You should consider whether you understand how CFDs work and whether you can afford to take the high risk of losing your money.

Experience Investable Attribute

@OGFX
If you read my old comments you will find I was suggesting a max D-period duration (40 or 50 days), following the same logic that there is a min D-period duration of 15 trading days.

Unfortunately Darwinex isn’t so “apples-to-apples”.
After one year of slow trendtrading you could get an Ex of 2.
This is not the same as 1 year of scalping and a Cp of 2 … :wink:

1 Like

Good morning everyone!

The experience attribute is calculated based on D-Periods of experience. 1 D-Period of experience equals to 1 month of full time trading. In order to get the highest grade in this attribute any trader needs to accrue 12 D-periods of experience = 12 months of full time trading.

For any trader to collect 1 D-period of experience, they´d need to meet 2 criteria:

-15 TRADING DAYS => Days in which the trader has operated (buy, sell, partial sell, etc)

-18 REPRESENTATIVE TRADING DECISIONS. We measure how representative a decision is based on 2 factors: D-LEVERAGE and DURATION. In fact, we calculate every single position representativeness with the following formula: MARKET EXPOSURE = D-LEVERAGE (of Darwin) * SQRT TIME

In Darwinex, we grant 1 full representative decision to the position with the highest exposure within the last 15 trading days. With this one acting as a benchmark, we calculate how representative the other positions, opened in the last 15 trading days, are.

@OGFX In your case, the most representative of all your positions opened within the last 15 trading days would be the one lasting 7 calendar days. This is going to act as a benchmark for the rest. One position lasting 7 days is 7 times more representative than one lasting 1 day or 14 times more than one lasting 12 hours. Remember that we have to get to 18 to gain a D-Period of experience. I can assume draw the conclusion that the main factor preventing your Experience attribute from improving at a faster rate is the fact that some of your positions remain open much longer than others. That means that the market exposure (representativeness) of the ones that last much longer than others -remember that one variable to calculate this is the sqrt of time- overshadows the ones lasting less.

I encourage everyone to watch or rewatch our EXPERIENCE webinar again since it is not an easy concept to grasp.

Have a great one!!

Jesús

7 Likes

I simply love this clear answer from @JesusDarwinex :heart_eyes: Well done!

5 Likes

That still is not quite clear to me. You need 15 days of trading, 18 decisions made and also 22 days of trading to collect 1Dperiod. SO it is not 15 days of trading. It is that in 22 days, 15 has been trading days with 18 decicions done over those 22 days.

There is still a confusion for me. Dperiod is anyways clear described but Experience grade is like a black box. It never matches the D-period, then why? when there are 30 trades in the month but only 18 are considered representative? The d-period has all the possible penalizations, so then why the experience grade is smaller?

Anyways… is a lossing trade a experience trade?
Has the same experience someone trading for 6 months lossing than someone trading for 3 months wining?
Yes you learn from lossing but someone may need 10 trades, other 20, other 1. So the experience learned is in the first winning trader after the lossing (learning trades ) so then let just count the winning ones. Or at least split the experience in some others ( as it is done with Risk and Timiing)

Something like learning rate. Because we one to follow traders with high experience, but also how that expeirence has achieved matters. Rather follow a Fast learner than a Slow learner (more lossing trades per winning trade (learned traded ) Is is like measure the streak loss or win of trades.

I hope the following webinar is useful: https://www.youtube.com/watch?v=MAz3aoKonmE

It is a bit old, but I hope it helps to get the main idea behind the Experience attribute.

Best,
Ignacio

2 Likes

@nanotrade

Experience has nothing to do with learning.

It is about the amount of data available to estimate investable attribute.

If you have not enought data, IA estimation is not relevant (I.E. Var estimated could be 10, but may be it’s 40, who knoes ?)

Furthermore if the trader realy need learning… hum I dont give a damn about is DScore. :sunglasses:

4 Likes

Good morning @nanotrade,

I highly recommend you to read my response regarding the attribure Ex where everything related to this IA is explained.

From your post, I interpret that your main doubt is not understanding why the D-Period metric and the Ex grade does not coincide. Fortunately, this has a very simple explanation. In order for any strategy to accrue 10 out of 10 in the Ex investable attribute, a strategy needs 12 D-periods and that´s why the Ex grade and the D-periods never coincide.

Here´s also an article from our FAQ section :wink:

Enjoy your day!

4 Likes

@JesusDarwinex
Maybe Experience should be renamed to Significance, so people would understand that the score has nothing to do with education or maturity of the trader.

7 Likes

Yes I agree with this and also I would suggest having the projected D-Score in brackets e.g. D-Score 37.2 (Est 75.3)

6 Likes

Experience score 10 = 12 D-periods.

In order to collect 1 D-period of experience, any trader needs to meet 2 criteria:

  • 15 trading days => Days in which the trader has operated (buy, sell, partial sell, etc.)
  • 18 representative trading decisions.

We measure how representative a decision is based on 2 factors: D-leverage and Duration. In fact, we calculate every single position’s representativeness with the following formula: Market Exposure = D-leverage * Sqrt Time.

We grant 1 full representative decision to the position with the highest exposure within the last 15 trading days. With this one acting as a benchmark, we calculate how representative the other positions, opened in the last 15 trading days, are.

That part can be found in ‘Education’ section of the website. But I needed a little bit clearer explanation. Here goes my communication with @JesusDarwinex.

My question:
Market Exposure = D-leverage * Sqrt Time…for example when D-leverage is 21.96 and position duration is 51min…what would be exact Market Exposure calculation? I need this to understand so I can better understand what 1 full representative trading decision represents and how the calculation works so I can see how my positions influence D-period calculation.

Answer:
Good afternoon Lukas!
Based on the data provided, the market exposure of this particular position will be 156.82 (21.96*sqr 51). We then, calculate the market exposure of all the positions closed over the last 15 trading days -duration in minutes- and use the one with the highest market exposure as a benchmark. This position -the most exposed- will be granted 1 representative trading decision and we will calculate the “representativeness” for the rest.
Let´s imagine that the position with the highest exposure has yielded a value of 1000. We grant this position 1 representative decision. In order to calculate the “representativeness” of the position in the above-mentioned example, we need to divide it by the most exposed one: 156.82/1000 = 0.15682. Therefore, these 2 positions make 1.15682 representative decisions.
In order to get a D-period, you´ll need 18 in more than 15 trading days as explained in this article.
May I wish you a great afternoon!
Jesús

After this I think everyone should be clear how Experience is calculated :slight_smile:

@JesusDarwinex maybe you could add the explanation you gave me with the example in some form to the education section to make the explanation even clearer :wink:

5 Likes

Thanks for pointing this out @Livenemaxx!

Sure, I´ll add this paragraph to the Ex article :wink:

THX again!

3 Likes

I agree about the key point you figure out : experience is not the right word because it’is all about statistical significance, not about the trader.

4 Likes

One option about so called Experience, would be a Real experience as a Darwin. This way everybody starts at ground zero when it comes to Ex score.

That means one who chose to start Darwin from scratch is in the same position redgards Ex score as one who transferred results from other broker.

Right now we see some Darwins who are here many years getting “nothing”, while new comers with history can get great positions at instant in Darwinia.

Reality is that Darwin is behaving slightly different than underlying strategy. While somebody who has account open at Darwinex for many years knows how to control Darwin while transferred newbie Darwin with good track record must still adopt this environment.

3 Likes

This was my idea one year ago.
Now I changed slightly my opinion.
Let’s suppose a trader trades just once per month, 24 trades would be a ridicolous sample even if the timespan is 2 years.

1 Like

100% agree with this part, but such a distiction is up to investors, for example I pretend a decent lenght of native trackrecord, but it has nothing to do with Ex and DScore, algos rate a trackrecord not the education of the trader.

The longer track record the better.
The longer track record as an active Darwin even better.

The longer track record gives a clue about the trader.
The longer track record as an active Darwin gives a clue about the trader, how is he/she/it adopted to Darwinex environment and algos.

I would keep the track record. It has great value by itself!
I would score Ex from the time when the Darwin was created. Score as an experience in Darwinex environment. This way everybody is newbie in Ex score when Darwin is created.

It is just an idea.
At the end we will all eventually get there.
Ex is the only score that does not fall.:wink:

5 Likes

Hi Jesus, apologies for the delayed response.

What you have said above (in emphasis) is exactly the problem with this attribute: it unfairly penalises long term strategies that use a trailing stop. If I’m in a position for 7 days, a month etc it is because the position is profitable and the market is moving in my favour. If I’m in a position for 12 hours it is because the market has moved against me and I have exited. This is good trading 101 - “Cut your losers, let your winners run” - it shouldn’t be penalised.

IMO this whole “representative trading decision” thing is incredibly abstract. Because my strategy let’s profitable positions run, the Darwinex algos arbitrarily discount the vast majority of my trades. In order to please this attribute, I would have to risk more on my losing trades and risk less on my winning trades … I think I missed that chapter of good trading 101 :joy:

It will take me close to 5 years to clock 10 experience at this rate. Even though my current parameters are by far the most profitable, if I want to make Darwinia before my hair falls out, I’m going to have to downgrade to a substantially less profitable set of parameters with tighter targets or time-capped exits.

A much fairer and simpler solution for d-periods would be:

  • 1 D-Period requires at least 15 trades opened and one calendar month

This would mean all strategies would require a sample size of 180 trades and at least 12 months of track record. Can’t be gamed through HFT, doesn’t overly penalise long term strategies.

Would still take me 2 years+ to clock full experience, but that’s better than 5!

1 Like

If you think you current parameters are profitable it means you testesd them on a sample of trades.
My parameters are tested on a sample of more than 200 trades, when I tested on smaller samples parameters didn’t work.
You invest if you have reliable parameters, they invest if they have reliable scores.

[quote=“JesusDarwinex, post:42, topic:541”]
@OGFX In your case, the most representative of all your positions opened within the last 15 trading days would be the one lasting 7 calendar days. This is going to act as a benchmark for the rest. One position lasting 7 days is 7 times more representative than one lasting 1 day or 14 times more than one lasting 12 hours.[/quote]

The problem is that Darwinex think that one position lasting 7 days is 7 times more representative…

It’s what Darwinex think, but purhaps it’s not true. Personaly I suppose it is mathsqrt(7) times more representative - may be less. You may have 10 trades winning 30 pips during 12 hours in average and one during four days. One cant say that there is a strong decision (the 4 days one) and nine other whitch dont realy count.

In fact it’s the market witch decide of what the futur is. Control from skilled quants is an illusion. It’s the same problem with Os IA. A bad score because of -10%… -50% better, mean try to improve the strategy using the time machine.

6 Likes