Discussion of article "Exploring Seasonal Patterns of Financial Time Series with Boxplot" - page 4

 
Igor Makanu:

I can't get enough of this topic, as they say, "Hello again!"

Here I have sketched my verification code to test the hypothesis "Let's allow to open trades only in 0-1 hours, with the assumption that in the next few hours the trade will still be closed in profit,".

ran in optimiser 01.01.2017-01.0.12019, if you look for nice balance charts, the optimiser finds takek 100p and stoploss 4600p

if you look for reasonable take/stoploss, there is no pattern at all


or still the methodology of searching for a pattern does not work, well, as an option I lost practice and wrote a crooked code - I haven't written in MQL for more than a month.

it's some other TS that doesn't exploit the pattern.

You have to buy below average, and you buy anywhere.

A perfect representation of when you don't understand the pattern and optimise and optimise ... :)

 
Maxim Dmitrievsky:

it's some other TC that doesn't exploit the pattern.

You have to buy below average, and you buy anywhere.

then the phrase"We allow opening trades only in 0-1 hours, with the assumption that in the next few hours the trade will still be closed in profit" does not apply to the text of this article.

My code checks this very statement.

I pay attention to why I've been talking about the TS with MA(25) for several times now - it turns out that the analysis of the pattern search has been reduced to the evaluation of the TS efficiency by MA(25) or to the evaluation of MA(25) in general - what will happen to MA after 1 a.m., but in no way to the search for the pattern "the price will grow after 1 hour..... Well, at least it will ever grow relative to the price of 0-1 hour".


Maxim Dmitrievsky:

Excellent representation of when you don't understand the pattern and you are opting and opting... :)

no, the goal is not to optimise, but to get a stable profit on the investigated interval of the CD - so far I don't see where any regularity is.

 
Igor Makanu:

then the phrase"We will allow to open trades only in 0-1 hours, with the assumption that in the next few hours the trade will still be closed in profit" is not to the text of this article

my code checks this very statement

I pay attention to why I've been talking about the TS with MA(25) for several times now - it turns out that the analysis of the pattern search has been reduced to the evaluation of the TS efficiency by MA(25) or to the evaluation of MA(25) in general - what will happen to MA after 1 a.m., but not to the search of the pattern "the price will grow after 1 hour..... Well, at least it will ever grow relative to the price at 0-1 hour".


No, the goal is not to optimise, but to get a stable profit on the investigated interval of the MD - so far I don't see where any pattern is

read the article carefully, your code doesn't check anything.

you should buy below the average, then the return to it becomes highly probable on all bars. The distributions on the bars themselves overlap.

This confirms once again that you can't find anything at all through mindless optimisation, ever.

What I have shown works both with and without stops and on different TFs and is not very sensitive to the MA period even

 
Maxim Dmitrievsky:

read the article carefully, your code doesn't check anything.

you should buy below the average, then the return to it becomes highly probable on all bars. The distributions on the bars themselves overlap.

this confirms once again that you can't find anything at all through mindless optimisation, ever.

What I have shown works both with and without stops and on different TFs and is not very sensitive to the MA period even

I read the article both last night and an hour ago, my questions are to the last section of the article! - As I wrote last night, the last section of the article is a bit of a stretch.

your code checks the TS on MA, and the result of the test in the tester is only "TS on MA on the interval 2017-2019" has a positive result, but not " We allow to open trades only in 0-1 hours, with the assumption that in the next few hours the deal will still be closed in profit".

ZY: well, and about mindless optimisation, then the result of the article about scientific testing of TS by MA - I can't see any other way

 
Igor Makanu:

I read the article both last night and an hour ago, my questions are to the last section of the article! - as I wrote last night, the last section of the article is tacked on.

your code checks the MA TS, and the result of the test in the tester is only "MA TS on the interval 2017-2019" has a positive result, but not " We will allow to open trades only in 0-1 hours, with the assumption that in the next few hours the trade will still be closed in profit".

ZY: well, and about mindless optimisation, then the result of the article on scientific testing of TS on MA - I don't see any other way

hint: the EA result is obtained WITHOUT OPTIMISATION

at least the author should be thanked for this :-) And carefully air your head - "why it worked".

 
Maxim Kuznetsov:

hint: the EA result is obtained WITHOUT OPTIMISATION

everyone has his own hints in his head ))))

SZY: I remembered a colleague who boasted that at night he drove 200 km in 1.30 distance to the object, which during the day we drove 2.30, unfortunately I did not appreciate the hint, well he pressed the pedal as hard as he could, held on to the steering wheel, what's the point? - If he could run 200 km with his legs in 1.30, that would be it! - cool!!!

 
Igor Makanu:

everyone has their own hints in their head ))))

SZY: remembered a colleague who boasted that at night he ran 200 km in 1.30 distance to the object, which during the day we drove 2.30, unfortunately I did not appreciate the hint, well pressed on the pedal, as much as possible, held on to the steering wheel tighter, what's the fun? - If he could run 200 km with his legs in 1.30, that would be it! - cool!!!

When a grown man starts remembering army stories, it speaks volumes.

 
Igor Makanu:

I read the article both last night and an hour ago, my questions are to the last section of the article! - as I wrote last night, the last section of the article is tacked on.

your code checks the MA TS, and the result of the test in the tester is only "MA TS on the interval 2017-2019" has a positive result, but not " Allow to open trades only in 0-1 hours, with the assumption that in the next few hours the trade will still be closed in profit".

ZY: well, and about mindless optimisation, then the result of the article about scientific testing of TS by MA - I can't see any other way

I don't see any logical errors in my narration, Rashid pointed out lexical errors to me when publishing )

 
Didn't catch the discussion well. My vision is as follows
// https://www.mql5.com/en/articles/7038

#include <MT4Orders.mqh> // https://www.mql5.com/en/code/16006

#include <fxsaber\Virtual\Virtual.mqh> // https://www.mql5.com/en/code/22577

#define  BESTINTERVAL_ONTESTER // The optimisation criterion is the profit of the best interval.
#define  BESTINTERVAL_SLIPPAGE // Creating an artificial filter to calculate BestInterval.
#include <fxsaber\BestInterval\BestInterval.mqh> // https://www.mql5.com/en/code/22710

#define  Ask SymbolInfoDouble(_Symbol, SYMBOL_ASK)

input int inPeriod = 25; // Mashka period
input int inLow = -2;    // Lower boundary of the difference between the price and the MAP - entry
input int inHigh = 0;    // The upper limit of the difference between the price and the MAP is the output

const int hnd = iMA(NULL, 0, inPeriod, 0, MODE_SMA, PRICE_CLOSE); // https://www.mql5.com/ru/forum/170952/page152#comment_14131263

const double dLow = inLow * _Point;
const double dHigh = inHigh * _Point;

double maArr[], prArr[];

void OnTick()
{
  static bool Position = false;

  CopyBuffer(hnd, 0, 0, 1, maArr);
  CopyClose(NULL, 0, 0, 1, prArr);
  
  const double pr = prArr[0] - maArr[0];

  Position = Position ? !((pr >= dHigh) && OrderSelect(0, SELECT_BY_POS) && OrderClose(OrderTicket(), OrderLots(), OrderClosePrice(), 0))
                      : (pr < dLow) && (OrderSend(_Symbol, OP_BUY, 1, Ask, 0, 0, 0) > 0);
}
3-4 input parameters for Optimisation and see the result.
 
fxsaber:
Didn't catch the discussion well. My vision is 3-4 input parameters to Optimise and see the result.

Unfortunately, I have forgotten how bestinterval works.

maybe you can share the final result? best periods. I wonder how close it can "hit"