Thoughts on the random - page 11

 
tara:

Alexey, you're walking yourself into this trap, imho. A simple trend line will help more than a dozen pages of reasoning about the stationarity of a process, the characteristics of which we do not know:)

And we never will.

Constructing a "simple trend line" with real statistical advantage is not, I think, as simple a matter as you write about it.

How many trades have you had with such a trendline - and how many of them were profitable? If the average stoploss is different for the average and for the profitable one - then specify their values for the average and for the profitable one.

And I'll try to decide how valid your opinion about beating the market is.

The real meaningful bid for victory over the market is, imho, no more than three people out of a thousand.

 
Mathemat:

I am not sorry to admit defeat.

What defeat? Is it statistical methods and pair trading or what? I've forgotten what you were working on.

 

how can you draw trend lines on imaginary charts at all?

for example - the euro, as a currency in calculations, appeared in 1999, if memory serves me correctly....:-)))

and the posted history - as far back as 1971.... what's it mean ? :-)))

 
Mathemat:

Yeah, but those "overtones" won't be visible on the surface.

There is no primitive stationarity. There is no obvious pattern at all from which to pull a profit.

The laws (=steadiness) from which something can be extracted in terms of profit - exist. But to get to them you have to rack your brains.

----------------------------------------------------------------------------

But I have no doubts about their existence - until now, after almost 10 years since I first got to know Forex.

I'm not sorry to admit defeat. Many of the other "winners" shouting "I beat the market!" are deluded. They are engaging in self-delusion, but they don't know it.


In that case, maybe we should look at the designers of sea-surf power generators. These people, just like us, are trying to extract energy from Vibrations... and do NOT look for any law and order in the changes in the height of the water column...

And they do succeed...

 
Mathemat:

Yeah, but those "overtones" won't be visible on the surface.

There is no primitive stationarity. There is no obvious pattern at all from which to derive a profit.

Laws (=stationarity) from which something can be extracted in terms of profits do exist. But in order to get to them you have to break your brain.

My brain is already broken by searching for years (I've been searching for 5 years). But it is really difficult to extract the benefits. One could say that the regularities are non-parametric and therefore are hardly amenable to standardized analysis by statistical methods.

- One variant of further search I see in adaptive NS (retraining with high frequency), at present I am writing my NS in Excel environment;

- The second variant - using "primitive" heuristics, like "fast price movement will continue", "breakdown of maxima". Use of nearest stops;

- The third variant - I start to dig into the subject of range bars (Renko) and want to investigate them from the viewpoint of mutual informativeness;

- The fourth option - do not bother and give up. By the way, I invite Muscovites to the bath (a real one), where exactly to go can be decided. Need to steam the body before the New Year, to free it from toxins.

 
prikolnyjkent:


In that case maybe one should pay attention to designers of generators running on the energy of the sea surf? These people, just like us, are trying to extract energy from Vibrations... and do NOT look for any law and order in the changes in the height of the water column...

And they do...


Kent,

Keep one thing in mind: the height of the water column is a very stationary value. It is cyclical and strictly predictable. The charm of Forex is that no matter where you open, it is likely to be 50/50. It's an interesting chart.)

 
alexeymosc:

The brain is already fractured by years of searching (I've been looking for five years). But extracting the benefits is really difficult. One could say that the patterns are non-parametric and therefore almost impossible to standardised analysis by statistical methods.

- One variant of further search I see in adaptive NS (retraining with high frequency), at present I am writing my NS in Excel environment;

- The second variant - using "primitive" heuristics, like "fast price movement will continue", "breakdown of maxima". Use of nearest stops;

- The third variant - I start to dig into the subject of range bars (Renko) and want to investigate them from the viewpoint of mutual informativeness;

- The fourth option - do not bother and give up. By the way, I invite Muscovites to the bath (a real one), where exactly to go can be decided. Need to steam the body before the New Year, to free it from toxins.

I never cease to wonder: why absolutely anything, often a very complex tool, is being researched except the mainstream called econometrics.

PS. NS is nothing more than one method of classification.

 
faa1947:


PS. The NS is nothing more than one method of classification.


It is not just a classification. It is in a broad sense a non-linear filter whose values can be used as signals for trading.

 
alexeymosc:


Kent,

One thing to keep in mind: the height of the water column is a very stationary value. It is cyclical and strictly predictable. The charm of Forex is that no matter where you open, it is likely to be 50/50. That's an interesting chart to draw )


Thank you colleague. I have already taken everything into account. And I even have the ability to support the faith in success of those who are still looking, with totally free real-time trading signals, without fear of being ridiculed. I just do not know how to arrange it, if anyone needs it...
 
faa1947:

I never cease to wonder: why absolutely anything, often a very sophisticated tool, is investigated except the mainstream called econometrics.



I can answer that, too. I use statistical methods too, autocorrelation for example. But looking for autocorrelation in quotes makes no sense anymore. That was interesting 20-30 years ago. Now, in the conditions of Internet trading and even in Forex, there are no linear dependencies.