Is martin so bad? Or do you have to know how to cook it? - page 14

[Maxim Zaguzov]

tol64:

Examine each drawdown in visualisation mode to find out what was really going on there. A detailed analysis will give you ideas on how to reduce drawdowns. And you may also discover errors, after fixing them, the result may also improve. ))

I am familiar with such cases. I have seen a losing position. I have an idea how to get rid of it. I have added a new function to my Expert Advisor, ran a new test. And "Oh, my God!" This loss trade has disappeared, as well as some other ones! But some of my other trades were lagging and some of them were filtered out! :DDD

And you start to analyse the report generated after the test and think which variant is better. And it will be better only for the history interval we've performed the test on... What a mess! ;)))))))

P.S.: Sorry to interrupt your dialogue...

[Anatoli Kazharski]

MaxZ:

I know this kind of thing. Saw a losing trade. Got an idea how to get rid of it. I added a new function to the Expert Advisor, ran a new test. And "Oh, my God!" This losing trade was gone, as well as some other ones! But some of my other trades were lagging and some of them were filtered out! :DDD

And you start to analyse the report generated after the test and think which variant is better. And it will be better only for the history interval we've performed the test on... What a mess! ;)))))))

P.S.: Sorry to interrupt your dialogue...

Oh, come on. Everybody can give their opinion. )) And the option is better the one with less losing series and consequently less drawdowns and number of drawdowns. The final test on, for example, a ten year history with a varied volatility landscape.

[Roman Shiredchenko]

tol64:
Oh, come on. Everybody can give their opinion. )) The better variant is the one that has fewer losing series and consequently fewer drawdowns. The final test on, for example, a ten year history with a diverse volatility landscape.

I don't get it - what is the scheme for winding volumes: "Thus TP = 2 * SL. When one of the orders triggers, the second is modified by the volume of the open position in the scheme 1 2 2 4 4 8 8 16 16 32 32, etc.".

Is it realistic to make a profit on series of trades?

May I chew a little, it's not clear to me... I haven't converted this MM variant to my code yet and I haven't tried it with my Avalanche...

[Anatoli Kazharski]

R0MAN:

I don't get it - what is the scheme for winding volumes: "Thus TP = 2 * SL. When one of the orders triggers, the second is modified by the volume of the open position in the scheme 1 2 2 4 4 8 8 16 16 32 32, etc.".

Is it realistic to make a profit on series of trades?

May I chew a little, it's not clear to me... I have not converted this MM variant to my code yet and I haven't tried it on my Avalanche...

I haven't tried this scheme yet, but I'll definitely try it. In those results I gave above SL == TP, and on losing trades there is a doubling. Just developed a decision block that answers the question, "What to do if something goes wrong?". And it can only go wrong if there is a prolonged series of losing trades. And if there is a limiter and a reset, you can still think what to do if one series is immediately followed by another one. You can also switch algorithms. Only one algorithm does not have to work all the time.

I do this when I start a new project. I'm not trying to figure out how this or that idea that comes to my mind might work, or I've read somewhere an interesting algorithm that someone has described in words. Trying to figure out what might come out of it can lead to the wrong conclusions. It may or may not seem to work. And in the end it will turn out to be the opposite. That's why I start coding straight away, the tests will show everything.

[Alexey]

R0MAN:

I don't get it - what is the scheme for winding volumes: "So TP = 2*SL. When one of the orders triggers, the second is modified by the volume of the open position in the scheme 1 2 2 4 4 8 8 16 16 32 32, etc.".

Is it realistic to make a profit on series of trades?

May I chew a little, it's not clear to me... I haven't tried this MM variant yet and I haven't tried it with my Avalanche...

Here you can find a detailed description of ithttps://www.mql5.com/ru/forum/138220

It's worth noting that no martin scheme gives positive mathematical expectation, on a random walk. Sinking on a Martin in any of his scheme is inevitable it's just a matter of time, and this time interval may be much shorter than when trading a fixed lot.

Now let's get to the point. The real price series differs from random fluctuations in volatility. If we consider intraday periods volatility is periodic and related to the Earth's rotation. This is what I am trying to base my model on.

I will not answer about the real profit yet, I need to complete tests. At the moment I have the following preliminary results. Max drawdown in money on constant lot 0.1 with 100 leverage.

[Roman Shiredchenko]

ivandurak:

It's all explained in detail herehttps://www.mql5.com/ru/forum/138220

It is worth noting that no martin scheme gives a positive mathematical expectation on a random walk. Losing on a martin in any of his scheme is inevitable it's just a matter of time, and this time interval may be much shorter than when trading with a fixed lot.

Now let's get to the point. The real price series differs from random fluctuations in volatility. If we consider intraday periods volatility is periodic and related to the Earth's rotation. This is what I am trying to base my model on.

I will not answer about the real profit yet, I need to complete tests. At the moment I have the following preliminary results. Max drawdown in money on constant lot 0.1 with 100 leverage.

О! Thank you... Familiar links - I'll read them, freshen up the information... The topic on the spider was raised a long time ago: "The regularities of the euro's hourly movement"... The topic was related to volatility fluctuations.

I downloaded the owl on mcl4 - I have it in my tasks, I haven't looked at it yet.

[JohnyPipa]

There is a parameter in the test results called Z-Score. This is what will tell you whether you can or cannot use a martin. Here is a clear and straightforward example http://miotai.info/5070-z-schet-kak-pokazatel-jeffektivnosti-torgovojj.html.

Z.I. Maybe someone knows how to make a Z-account as an optimization criterion?

[Mykola Demko]

JohnyPipa:

There is a parameter in the test results called Z-Score. This is what will tell you whether you can or cannot use a martin. Here is a clear and straightforward example http://miotai.info/5070-z-schet-kak-pokazatel-jeffektivnosti-torgovojj.html.

Z.I. Maybe someone knows how to make a Z-account as an optimization criterion?

Send the result of Z-account to OnTester(), run the Expert Advisor on Custom optimization parameter (this is the last line in optimization parameter selection).

[Serj]

JohnyPipa:

There is a parameter in the test results called Z-Score. This is what will tell you whether you can or cannot use a martin. Here is a clear and straightforward example http://miotai.info/5070-z-schet-kak-pokazatel-jeffektivnosti-torgovojj.html.

Z.I. Maybe someone knows how to make a Z-account as an optimization criterion?

In MT5 there are more interesting parameters, e.g:

STAT_MAX_CONLOSS_TRADES

Number of trades in the longest losing seriesSTAT_MAX_CONLOSSES

[Alexey]

JohnyPipa:

There is a parameter in the test results called Z-Score. This is what will tell you whether you can or cannot use a martin. Here is a clear and straightforward example http://miotai.info/5070-z-schet-kak-pokazatel-jeffektivnosti-torgovojj.html.

Z.I. Maybe someone knows how to make a Z-account as an optimization criterion?

If you have a strategy where Z-account differs significantly in positive or negative side you don't need martin. Use any other MM it will be much safer and money plan much more efficient.