a trading strategy based on Elliott Wave Theory - page 207

New comment
 
<br / translate="no"> Sergey, those time series that we operate with (price series) are already integrated series of the first order (as a rule). By taking successive differences we will get a stationary series of residuals whose properties we will study. This is the right move. When opening a position, we actually operate not with the absolute value of the symbol rate, but with its expected increment for the time of holding the position, i.e. we work with a series of differences. As I said before, the entire variety of trading strategies comes down to a single action - predicting the price movement direction after opening a position...
It's too early to derive a criterion for detecting a deterministic trend. We need to build a coherent and, if possible, internally consistent picture of price formation, and only then will it become clear how to build an optimal predictive model. My hope.


Sergey, thank you for the clarification, I did not know that the price series already turns out to be integrated of the first order. It is kind of news to me. Is there any justification for this statement?

I disagree with the statement that it all comes down to predicting only one direction of price movement. Having guessed/predicted/calculated (as you wish) the direction, you still need to figure out how long to hold a position, or even a correct prediction may not help. The second point is especially important, especially now, as I understand you saying, that a long term forecast is impossible (by the way, if you please define the quantitative expressions, maybe I missed it somewhere).

So, we represent the series as a generally accepted superposition of four elements:
1. trend
2. seasonal fluctuations
3. cyclical fluctuations
4. random.

The above representation is a model or it is only a decomposition of the series in the mathematical sense of the term (quite a Taylor series for an example, it does not matter in general, "technology" is now enough). If we are talking about the primary pricing mechanism, I'm afraid it's very complicated and even fundamental analysis won't help here. The mechanism of galosha pricing is likely to be driven by season, country, cost of oil, etc. But apart from galoshes, there is a huge mass of everything (I think an FA supporter could make a decent turn here) that eventually translates into currency values: for example, many seasonal goods within the "planet" or rather forex lose this component and so on and so forth.

I suppose it will all eventually come down to "trend" and "noise", that will be the model. So maybe it's time to move on to criteria or do you have a better idea? :о)

PS: So are we moving exactly to a pricing model in the fundamental sense or to a decomposed series (and as you know very well, the same series can be decomposed in several, but in different ways and with different results) deepening in mathematics?
 
Neutron
It is too early to derive a criterion for detecting a deterministic trend. We need to build a holistic and, if possible, internally non-contradictory picture of price formation, only then will it become clear how to build an optimal predictive model.

I don't think this is possible. At least the micro picture.

And the macro picture, in general, is fairly well known. There are international markets for goods, raw materials, capital, investment, etc. There are national economies where all this is produced or used. The counter-flows of bilateral economic relations generate counter-financial flows. The relative exchange rate of the two currencies provides a balance of supply and demand. This is the first approximation. And in the second and beyond, there are crosses that take into account interactions with third countries. And so on. That's it - primitive and in a nutshell. But it follows that the long-term trend can be determined with FAs. And it really is.

However, for trading (not investing) this is not enough. The time horizon of trading is tens and hundreds of times shorter. Besides, for trading it is points, and for investing - cents. There is a 100-fold difference as well. And finally, for trading, the flow of events that influence the market in one way or another is incommensurably larger and much more heterogeneous. If the economic indicators of FA are more or less established and they are numbers, for all other events, for which there is no number, there is no meaningful, arbitrarily deprived estimations at all. How is it possible to construct "a holistic and, if possible, internally consistent picture of pricing"?

In my opinion this is only possible in one way. In the same way that the transition was made from a microscopic description of dynamic systems in Newton's theory, to their macroscopic description in thermodynamics. But to do this, we must see the market as a holistic system, one of the intrinsic properties of which is the price. And instead of trying to tie the price to individual events that take place on the market (the news release, for example), we must try to find other, also macroscopic, means of describing this system.

The role of experts in the field of mathematical statistics is probably the most important one. After all, the development of thermodynamics (statistical physics) and mathematical statistics were things interrelated and interdependent.
 
Sergey, I hasten to extend my comments as well. Following Yuri's approach, the definition of "model" should also be given. In this way, we have a good chance of getting bogged down in our endeavours.

Recall your calls to use proven, reliable approaches, in particular autocorrelation as opposed to CCS. Do you foresee a change in approaches to its calculation due to the construction of the model? If so, then autocorrelation will no longer be reliable. If not, why come up with a model, you can start researching it now.

PS: and by the way, don't consider it annoying, explain me where you get the sliding window in the autocorrelation calculation and what is the point and use of it. Or is it some kind of modification for some task?
 
Alex Niroba
Neutron, you're fucking full of shit!!! :))))))))))))
Believe me, it's a lot simpler than you think...

It's not! You're having an illusion about forex. It doesn't get any easier.
Proven.

[/quote]


Neutron , I may harbor some illusions but I'm not trying to "make my life harder". :))) WHY???? You're like Grandpa Susanin, you can get into such thickets :))))))))))))))))))))))))))))))))

Forex is first and foremost about people, or MTCs, which are also developed by people :))).
Do you really think that most traders deal
this kind of nonsense!!! :)))))))))))))))))))))))))))))))))))
Neutron , my dear friend, don't consider it rude !!! :)))))))))))))))))))
 
looking at where this discussion is heading, I thought I'd throw some more wood on the fire...

http://www.altertrader.com/publications03.html
http://www.altertrader.com/publications14.html
http://www.altertrader.com/publications02.html
http://www.altertrader.com/publications01.html

don't waste your time, take a look.

Also. since you have moved on to time series analysis anyway, it might be worth looking at the so-called "gooseberry" or SSA.
 
looking at where the discussion was heading, I decided to throw some more wood on the fire... <br / translate="no">
http://www.altertrader.com/publications03.html
http://www.altertrader.com/publications14.html
http://www.altertrader.com/publications02.html
http://www.altertrader.com/publications01.html

Please take the time to read it.




Yep, it's getting hotter from the campfire :)))
Quote: ( purple corresponds to the minimum value, red to the maximum).

I'm not color-blind, I've been looking at the picture http://www.altertrader.com/publications03.html for a long time , but I didn't see purple °/- :)))
it's a puzzle, though :)))))
 
<br / translate="no"> ...
Yeah, it's a challenge, though :)))))

You can see everything without colours, you just have to use your brain...
 

...
Мда-а-а, задачка, однако :)))))

It's easy to see without colours, you can tell from the isogypses.





I've been racking my brain, but I can't figure out which curves to look at,
North Wind , did you draw it?! :)))
 
I've been racking my brain but I can't figure out which curves to look at,

Alex, give it up. This is a forecast for the past year 2006.
Look at their pivot point calendar and the daily chart for 2006.
Even in January-February (and the forecast was made in December 2005) they are wrong.

This forecast may be of interest in theoretical terms, but in practice the reliability is very low. On the other hand, why make such an ambitious and thankless task - a forecast for one year?
Then it is possible and for the millennium. It has recently begun. :-)

Let them show acceptable accuracy for 30 bars ahead. At least on some TF.
They would not be of any price then.
 
... <br / translate="no"> Look at their pivot point calendar and the 2006 daily chart.
Even in January-February (and the forecast is from December 2005) they are wrong.
...

the authors claim that almost all of their predictions for this year have come true. haven't checked it myself. here's a quote:

"...And so. The result, which gave a trivial methodology exceeded my wildest expectations.
For now, 24 out of 30 predicted points have worked "correctly". Some simple mathematical operations showed that it was 80% of coincidence. Moreover, there are only 3 forecast points left until the end of the year, i.e. in the worst case scenario the accuracy of the calendar reading will be approximately 73% ...".

http://forum.viac.ru/viewtopic.php?p=81395#81395
1...200201202203204205206207208209210211212213214...309
New comment