That's interesting - page 20

 
Farnsworth:

I would suggest, theoretically of course, that proctologists also use arithmetic in their professional work. Is this also TA?

Well, if a proctologist uses arithmetic to predict the movement of the relevant "instrument", then there is no fundamental difference. :)
 
Andrei01:
Well, if a proctologist uses arithmetic to predict the movement of the relevant 'tool', then there is no fundamental difference. :)
Well, maybe, but he does fundamental analysis. By the way, he also uses arithmetic
 

I don't know shit, right off the bat. But I'll put in my own stupid thought.

A classic martingale is a random walkover. Price BPs are not really martingales. Because "mixing" several price BPs sometimes results in very stable readings.

It does not have to go far: highly correlated financial instruments. Or, more generally, a portfolio that has a constant MO of its equity. Therefore, it trades inside its equity channel from the levels of its RMS.

On Out of Sample patterns persist for quite a long time in some portfolios. Of course, there is no guarantee that it will break tomorrow, but neither is there a guarantee of anything.

 
HideYourRichess:


. I was wondering, you've been on the forum since 2006. You're practically an axakal, so I'll have to address you more respectfully.

Oh thank you, that's quite an honor and respect.

. Let's start from afar. The object you are trying to work with is a time series of prices. The only market characteristic you are given to see through the dc screen. I had to spend a lot of time (several years) and effort to make sure that the price shows the main characteristic of a martingale - namely, that the best forecast for the next value is the current value. That said, still, price is not a classical martingale, but rather a special class of martingale (a martingale made of "pieces", plus general non-stationarity, plus something else), but it's not that important. For that reason, I don't hesitate to write, price ~= martingale. Martingale has one property, any "transformation" of a martingale results in a new martingale. To put it simply, no matter how you convert the price using clever methods, you will still end up with a series-martingale. And that means "50/50 minus the spread" (c) someone else's.

. In fact, what is written is very simplistic, just a principle. I'm not going to convince anyone, especially since the price is not a strict martingale, but this non-strictness does not allow to be better (or worse) than the "market average", adjusted for commissions.

I understand very well that the only way to write a really stable trading system on price series is to use external data, i.e. additional information, external to the system. I also work in this direction, but it is too titanic.

My relation to the theory of an efficient market and to the theory of fractal market has already been stated on this forum. In short, both theories are not quite adequate, but both theories are a great way to earn money for authors with books, grants, articles, lectures, etc. Well, neural networks are just a mathematical apparatus, a tool.

Following the above, you should also understand that no TA in the classical sense of the term will give any advantage. All "characteristic" extremums and "movements" actually do not characterize the quoting process in any way.

Anyway, I haven't written anywhere that it is an ideal approach to "opening" the process. There are shortcomings, of course. But for now I am working with it. I will develop it further.

. there is no "beauty" because it is a martingale. :) Even if you insistently chisel away at it day after day, month after month, year after year with different methods, it won't work, because it is not realistic. Sorry!

The result is there, but there is no firm certainty. Including for the reason that you have named.

We need a completely different approach, a different view.

What is it? Although I suspect you will say trade secret.

. This is not important, especially for people who have not yet become "lucky" themselves. It's the methods and views that are interesting.

I have not yet created a sustainable automated (that's the key word) system. But there's my secret lab working on a superweapon. :о) Since I am not a gambler by nature, I do not need to model TCs one by one and tickle my nerves.

. you don't know. My understanding is that what you do is TA, price movement analysis, and it doesn't matter what "tricky" method you use, the data is the same (the same martingale).

"Analysis" is something many people do. Fundamental analysis is also engaged in analysis, and can you imagine, it also analyses prices.

 
joo:

Not everything that uses rhythmetics is TA.

TA has a definition, who is interested, he may Google his interest.

If an analysis uses different underlying assumptions than TA (and there are richmetic everywhere you look), then it is not a technical analysis. And basta. That's my, imho'naya basta.

Sergey, please continue, don't get distracted.

Yes, I'm kind of finished. I think the purpose of the topic has been fulfilled - I have read enough interesting materials, I have argued, I have briefly introduced my system, and I have received respect from HideYourRichess. And what else do you need here? :о) Next, everything is boring, you need to detail the concept, try to go to the theory, plan a few experiments, they will show what's what.
 
Farnsworth:
MathCAD doesn't really work with dates. I can't tell the first minute of Monday from something else. I'll read the documentation, maybe some functions have been introduced.


There's no need to distinguish the first minute of Monday, just like any other.

The file should be written from Monday, without any dates, only Close, one day in a row, every 5 rows - week, holidays should be filled with zeros, zeros should not be processed in the program, holes in the stream of minutes should be filled with the last preceding value, the file will end with the last Friday's quote. What else? Can't read a file like this correctly ?

 
Yurixx:


No need to distinguish the first minute of Monday as well as any other.

The file should be written from Monday, without any dates, only Close, one day in a row, every 5 rows - week, holidays should be filled with zeros, zeros should not be processed in the program, holes in the stream of minutes should be filled with the last preceding value, the file will end with the last Friday's quote. What else? Can't read a file like this correctly ?

Of course I can.
 
HideYourRichess:

On anecdotes, well just for you:

The difference between theory and practice in practice is much greater than in
theory.
 
Farnsworth:

I am well aware that it is only possible to write a really stable trading system based on price series using external data, i.e. additional information external to the system. I am also working in this direction, but the work is very titanic.

. And how has this understanding been reflected in the model presented?

Farnsworth:

Following the above, you should also understand that no TA in the classical sense of the term will not give any advantages. All the "characteristic" extremums and "moves" actually do not characterize the quoting process.

. This is so and not so, at the same time. The trap, as usual, is in the interpretation of the terms. In the "mass" mind of DT clients, TA is price analysis. Full stop. Simply because those clients have nothing else. And I agree, that shit doesn't work, in that form. In another definition, TA is price and volume analysis (this website says so). But this is already a market, not a decisive forex. And then it turns out that TA works in some places. And if we take the definition that TA is also analysis of interest (again, definition from this site) - the task starts to resemble the one described by traders in their memoirs. In general, depending on the context, I can say "TA sucks" and then "TA rules". And I won't lie once. That's it!

. One quote: "Technical market analysis is the study of the psychology of the stock market crowd. The purpose of technical analysis is to determine the balance of power between bulls and bears in order to bet on the stronger group." (c) A. Elder - and old Elder is very right, and mind you, nothing about prices, volumes or anything else.


Farnsworth:

which one? Although I suspect you will say trade secret.

. And that brings us back to where we started. To understanding the problem of "what is your process"! or more simply, "what processes are going on in the market?", "who is involved in these processes?" etc. and only then the question "what does price have to do with it and how does it behave".

. And here I am going to write a quote, I am tired of writing everything down: "To answer this question you have to understand what the market is. It is not a box with balls, not a "spherical horse in a vakuum".

On the market different groups of participants work, pursuing various purposes. Their methods of achieving those goals are different, and the frames/periods/scales at which they operate are also different. Also, different instruments will have different ratios between these groups.

Some biotech stock is highly likely to be driven by pump & dump style manipulators or insiders. SP futures are loved by "gender" and big players. ES is being spun by arbs and programmes.

But all this diversity does not negate the fact that the price has a "memory", which is what leads to the floating depth of process marginality. This "memory" may be due to an accumulated position, fanatical adherence to a particular trading strategy, fear/greed, etc.
.

In order to understand what is happening, you have to get acquainted with the basics of how this market is structured, which groups operate on it, what goals they have, which assets they trade. Then by reading forums and personal observations you will come to understand the processes that take place in this or that asset in this time window. And the application of some apparatus for the construction of estimates is the last step.

Note - be as specific as possible, financial instruments have their own characteristics
.

Farnsworth:

"analysis" is something many people do. Fundamental analysis also does analysis, and can you imagine, it also analyses price.

. It's not that simple with FA. I myself am very superficially familiar with it, so I only know that this is how various "global" data is analysed, which can affect "price".

Farnsworth:

oh thank you, just such an honour and respect.

.

 
Farnsworth:

On anecdotes, well just for you:

The difference between theory and practice in practice is much greater than in
theory
.
This is only for practitioners who do not know theory.