Ward 6 - page 51

 
jelizavettka:

... The doctor's actions are not quite adequate. He's digging a ditch with his hands when there's an excavator nearby.

Perhaps. Would anyone undertake to writea"universal advisor" that would:

1. every 10-15 minutes (after closing a pair or three M5 bars), it would dump quotes into a text file (in one file close columns for, say, 10 currency pairs, while filtering the possible missing bars, and if any, fill them with the latest available values - i.e. in the same line all values either refer to the same point in time, or in separate columns substituted by the latest available values, most importantly that there were no lines where quotes are asynchronous)

2. start executable file (the file name shall be selected in the Expert Advisor);

3. waited for 5 minutes until .exe would make calculations and save the text file results.txt;

4. opened (or not opened) trades, drawing instructions from results.txt file;

5. goto 1.

 
jelizavettka:
The market is quite consistently impulsive on small timeframes.

The market is the same on all tf's. I have already written about this. If I erase the numbers from the axes, you won't be able to indicate from which TF I took the pattern. Even for a sufficiently long sample (1000 bars, for example). On a bar or two, even less. This is called fractality.

P.S. I am not claiming that there is no difference between TFs. Moreover, I claim that there is. But it is not a determinant of the type of pattern. It is effects of lesser order of importance than those that define patterns. So in principle it is possible to "guess a TF" without numbers on the axes. With some (not too much) certainty. But it is: 1 - not for average minds, 2 - has no practical value.

 
Dr.Drain:

What is a z-account?


https://www.mql5.com/ru/articles/1492

 
and what does that mean? you've got some nonsense again. the horizontal equity line is nonsense.
 
Dr.Drain:
and what does that mean? you've got some nonsense again. the horizontal equity line is nonsense.

Did you read the link?
 
Rorschach:

did you read the link???
the horizontal equityline is nonsense.
 
Rorschach:

Have you read the link?

The method of calculation is stupid.

"This means that there is a 99.74% probability that trades in this trading account had a positive relationship with each other (Z-score is negative)" - of course if I open to sell or buy EURUSD, GBPUSD, EURJPY, GBPJPY, CADUSD - and then catch 5 TPs or 5 SLs they are dependent on each other. Who would doubt that. However, something makes me choose entry directions so that the TP is larger. The method described above is applicable for consideration of trading on one pair. Moreover, it is not even clear how to average values obtained for 10 pairs. And in general, the validity of the initial formula causes a smile.

 
Dr.Drain:

The method of calculation is stupid.

"This means that there is a 99.74% probability that trades in this trading account had a positive relationship with each other (Z-score is negative)" - of course if I open to sell or buy EURUSD, GBPUSD, EURJPY, GBPJPY, CADUSD - and then catch 5 TPs or 5 SLs they are dependent on each other. Who would doubt that. However, something is forcing me to choose entry directions so that the TP is larger. The described method is applicable for consideration of trading on one pair. Moreover, it is not even clear how to average values obtained for 10 pairs. And in general the validity of the initial formula makes me smile.


And in general, have you tried to take into account the direction of the filter, does it change the results?
 
Filter direction? Trade "in the direction of the filter"? Unfortunately it is not a smooth enough curve. The derivative is too jerky. The operation of numerical differentiation is noisy by itself, and the initial signal is lousy for this purpose and the derivative will constantly change its sign for no apparent reason. You can see it in the filter pictures. So I essentially don't know the "filter direction". But it doesn't matter either.
 
Rorschach:

did you read the link???

Now we can look at the Automated Trading Championship 2006 table with slightly different eyes.

Ha-ha-ha. The smaller z is, the larger is the profit. I may build a stochastic dependence on that table at home :-) My estimate is -3.59 (does not really make sense when dealing on different pairs and choosing them at my discretion) corresponds to almost the first ranking with maximal profit. Meow.

Reason: