From theory to practice - page 1964

 
"
Дисперсионный канал стал практически постоянным, изменилось поведение скользящей средней и сделка была бы прибыльной.

"

And if you take increments of a constant value, the channel will be even flatter. Flatness of the channel, in my opinion, does not mean anything yet.

In the end there will be the same rake, once again getting into a "trend" there will be more losing trades than profitable ones, because that is the nature of price movement.

And the return of the sum of increments from the borders of the channel to zero - does not guarantee a price reversal.

The tail in the moving window makes noise (the new gain, the old lose).


Примем гипотезу, что истинное время рынка — это интервалы времени, при которых модуль приращения цены >= среднего значения приращения. 

How do you take the average increment? Is it calculated using the sliding window or statistics, or maybe it is a function of time?

I think I tried that on minutes, I took the increment that >= (sum of absolute/n). The result is the same.

 
Evgeniy Chumakov:

Hehe...

So not even a certain signalE convinces you that Toddler is right about something? Well, whatever.

 
Alexander_K:

Hehe...

So not even a certain signalE convinces you that Toddler is right about something? Well, whatever.


I'm looking at the signal, experimentation is good, but after experimenting on the m1 I have a gut feeling it won't help, we'll see.

 
Evgeniy Chumakov:


I'm looking at the signal, experimentation is good, but after experimenting on the m1 I have a gut feeling it won't help, we'll see.


No M1 test is going to help you see the Truth.
 
Alexander_K:
No amount of testing on the M1 will help to see the Truth.


Well yes, only a poker down on the balance sheet and on the real...

 

On the subject of thinning.

What if you quantise the time series in predetermined steps at different hours of the day?

Suppose the minimum step is 25 five digit points, then 50,75,100 etc. (or other gradation) .

Then, judging by the picture, if the step gets to 22:00, we should take its size = minimum, 23:00 increasing, etc.

Will this be of any use?

Who wants to do research, open a new thread. Maybe someone from the ticks can slice a row.

 
Evgeniy Chumakov:

On the subject of thinning.

What if you quantise the time series in predetermined steps at different hours of the day?

Suppose the minimum step is 25 five digit points, then 50,75,100 etc. (or other gradation) .

Then, judging by the picture, if the step gets to 22:00, we should take its size = minimum, 23:00 increasing, etc.

Will this be of any use?

Who wants to do research, open a new thread. Maybe someone from the ticks can slice a row.

This is exactly what A_K did with the ticks.

 
Алексей Тарабанов:
OK, we wait for tomorrow. Daughter in roubles, dollars sold at 75.47

Does the daughter place stop-losses? Or a martin?

It is very dangerous to counter a trend without stops.

 
Maxim Kuznetsov:

This is exactly what A_K was doing with ticks.


I think he was thinning out ticks by time (after n seconds, etc. ) or after n-ticks. And this is different.

 
apr73:

Does the daughter place stop-losses? Or a martin?

It's very dangerous to counter a trend without stops.

Scalping pennies. Sometimes doubles a month's deposit. I'm looking out.

Reason: