Arbitrage - page 22

 
DrawDown:
Paha wrote (a):

Or update beginPrice from time to time.
Which I did. I update this parameter at the beginning of each month. The drawdown has decreased, the quality has improved, but for one pair, because I cannot put it on the forward test yet.
Although I think it makes sense to bind beginprice not to time but to price trend. At the moment I'm searching for the tool most suitable for this purpose. There are a lot of levels, etc.

There are a lot of them. It's not difficult to find it, but the main question for me is how to physically bind it to my Expert Advisor. I want to know how to attach it physically to the Expert Advisor (I mean programmatically). I don't want or have time to sit and watch all movements.
I have one more thought: while working with the tester and online, I noticed that if the orders are successive, and moreover, five or six of them, then the order prices do not differ much and we can see a trend. You may try to close the first of a series of orders opened in a row with a negative sign, manually. The risk will decrease and you will lose less than if this order would have stayed there. I have recently put 3000 for 5 min. in online trading (since 23.04). As of now my depo is +406 and equity is +23 on 2 pairs. I got small lot 0.1 for each trade around 18-20 pips. My losses, even if they are small, my losses will not be big, I would rather earn 3000 or even 200 for the week. I will test it couple of months more, and then we will see. And about the periodic change in the initial price I have already said.


PS. And it would also be nice to be able to set two points bijin price. The first point is the beginning of trading, and the second one is manually set which would give the trend line, and then this EA could trade in the trend. But it would probably require a major revision of the Expert Advisor. Unfortunately, Mr. Reshetov refused to further improve it or introduce any variables. That is a pity. If anyone is able to do so, I will help you with testing. Thank you.
 
timbo:
And what would it change to run the EA on multiple currencies at the same time?
This is called risk diversification. See How to reduce financial risk
 
solandr:
corner_h:

For all of us who are not particularly gifted, here are a couple of pictures
explaining what beginPrice is

Apparently some people have not only an infinite amount of capital, but also an infinite amount of time, during which they are ready to wait for the market to return (1-2 years of waiting in the drawdown with a probability of getting a margin call, it seems to be a normal situation for this strategy).
Thanks for the pictures anyway! I am afraid they will not change anything in the opinions...

Who told you that this kind of explanation is needed? Leave us "strays" alone, let us discuss the subject in peace! If you don't like it, why waste your time on it? Apparently you have an endless supply of time?
I appeal not only to solandr, but to all the "defenders" of others' time and capital.

On the merits.
Pay attention, gentlemen, that Yuri noted that this advisor is a necessary minimum that he gave us to create his system. And also, remember that the first version of the EA (last May) was somewhat different...
 
Paha писал (а):

It's all there, it's all there. It's not difficult to find it, but the main question for me is how to physically bind it to the Expert Advisor? (I mean programmatically). I have no time or particular desire to sit and watch all the movements.

I'm sure you will. I tried to attach Murray levels to my Expert Advisor once. In my tester's log it registered all levels but traded on only one level. I did not have enough knowledge to analyze it and detect the error.

Paha wrote:

When I work in the tester, as well as online, I noticed that if the orders are placed one after another, and five or six of them, the order prices do not differ much and the trend is formed. You may try to close the first of a series of orders opened in a row with a negative sign, manually. The risk will be reduced and you will lose less than if this order would have remained there.
I have tried. I have added a Stop Loss parameter to the orders, but the result has not improved. Although, the stop-loss was static and did not depend on market dynamics. I believe this was the reason for the worsening of the result. That is why we should determine the criterion for increasing and decreasing the stop-loss but it is rather difficult, though we have some ideas concerning it.

Paha:

PS. And it would be nice to be able to set two points of bijin price. The 1st point is the beginning of trading and the second one is manually set, which would give the trend line, and then this EA could trade in the trend. But it would probably require a major revision of the Expert Advisor. Unfortunately, Mr. Reshetov refused to further improve it or introduce any variables. That is a pity. If anyone is able to do so, I will help you with testing. Thank you.
I have my doubts about two points (you may consider it a sure thing), because if we have two beginPrice points the orders will go both ways, which does not make sense. But the idea itself is realistic. I have added 3 indicators used in the "Patch" - trades become filtered. I.e. if beginPrice is lower than the current bid, then the sales take place only in moments of determining the downtrend by indicators, from short-term and to the very beginPrice. Of course errors may occur for every indicator, but they are rare and the overall result is better than without these indices. By the way, just at the moments of such errors the dynamic stop loss would be useful. I will decide on the latter.
 
DrawDown:
I have my doubts about the two points (you may think it is a sure thing). If there are two beginPrice points, the orders will go both ways, which does not make sense. But the idea itself is realistic. I have added 3 indicators used in the "Patch" - trades have become filtered. I.e. if beginPrice is lower than the current Bid, then the sales will take place only at the moments when the indicators detect a downtrend, from the short-term and up to the very beginPrice. Of course errors may occur for every indicator, but they are rare and the overall result is better than without these indices. By the way, just at the moments of such errors the dynamic stop loss would be useful. I will decide on the latter.

Maybe I have formulated the idea of two points incorrectly. One point set by beginPrice creates one horizontal line (page previous ) . Two points should set one line, but not a horizontal line, but a sloping line. If a sloping line is plotted on the bottom line, the result will be what I mean. Everything else should stay as it is. The lower one will not change. The only thing that will happen is that the start point beginPrice, for the current moment, will be kind of floating on the above sloping line with a slight lag (at the lag, it is just an assumption without grounds; it needs more elaboration).

By the way the second point can be automatically calculated as the arithmetic average of max. and min. prices, for example of a weekly chart (when playing at 1 o'clock). Please take a look, maybe there is something useful in it. And this point can be sent a little into the future if desired, but it is already
not for the seusecond moment.
Thanks!

PS. Unfortunately, I don't know a thing about codes. So I can't change anything myself, but I can think about it and test it. If you do not feel sorry and if you get anything out of it, will you let the adviser know?
 

I would like to ask Mr. Solandr not to use labels and epithets in his bazaar. I myself have long been not offended by many categories of people and the main thing for me is that no one is offended by me and I have not offended anyone, so less provocation.

To timbo: "And what will be the difference in running the Expert Advisor on several currencies at once?

If you assign experts=1 to each of the currency pairs and a unique magic number - then I think you can see an approximate picture of what will happen - by running an mql test for one time interval on all of these pairs in turn and then superimposing the resulting charts on each other. But if there is a group, everything is more complicated. The Expert Advisor sort of "mutates" under these conditions and its operation is much more difficult to calculate and evaluate. Personally, I cannot do this by reading the code, and therefore (and not only for this reason) I build my tester. Actually, there is a forextester, I don't have experience with it, but maybe something can be seen there. We can also "fill" a group in an EA with a slightly modified code and send it for testing - it will also be something similar. But this "something" will also most likely be not so good and similar to corner_h shown in the picture. Mr Reshetov gave recommendations on how to use it: switch it on and off for a while, etc. BeginPrice should also be changed, of course. Otherwise we may reach the absurdity and set this parameter to "0". Then everything would be abnormal. You cannot take everything so literally. This parameter is a starting point and it is important and should not be taken thoughtlessly from the "ceiling". I'm not a big expert of tests or Forex as well, but I think that the test for this EA should be run in discrete portions, rather than continuously. I think (reading the code) it would be better to insert a string into the body of the Expert Advisor (for each day): at a certain time (say, after 9:30 to 14:30, etc.) count the current price and assign it to BeginPrice. When the profit is reached, say 40 points or after a few hours, the Expert Advisor stops working on that day: a loss is a loss. I think this test is closer to the truth. Well, again, for those who are searching or trying to search for the truth. I personally have my own way.

Sincerely, Fed.

 
usdjpy:
timbo:
What would it change to run an EA on multiple currencies at the same time?
This is called risk diversification. See How to reduce financial risks
Thank you, of course... However, the main point in the article is the use of instruments with low correlation. How pairs based on one currency, for example USD, are low correlated? I once did an Expert Advisor for pairs USD, EUR, JPY - almost always the correlation coefficient was 100% on any timeframe and any period, i.e. there can be no diversification here.
 
timbo:
usdjpy:
timbo:
What would it change to run an EA on multiple currencies at the same time?
It's called risk diversification. See How to reduce financial risks
Thank you, of course... However, the main point in the article is the use of instruments with low correlation. How are pairs based on one currency, for example USD, low correlated? I once did an Expert Advisor for pairs USD, EUR, JPY - almost invariably the correlation coefficient was 100% on any timeframe and any period, i.e. there can be no diversification here.
Are you sure that usd eur jpy are pairs? Would you stop saying that the correlation coefficient is 100%.
The correlation coefficient has never been measured in percentages.

Timbo boy, go learn the basics. Read what pairs are and how correlation is calculated.
 
So there is nothing to answer on the merits of diversification with correlated instruments, let's get personal...

I was counting the correlations between all the pairs formed by the three currencies.
Reveal your intimate knowledge of how correlation is measured. Also think about why 0.1 equals 10%, for example.
 
I seriously doubt that it's 100%, timbo. What figure did you calculate?
Reason: