The most banal trading strategy - page 52

 
Unicornis:

What's the point? Brown line base from rsi stochastic etc, Aqua is three times easier and also 10 times or more faster. Same primus +/-, only in profile:


It's about channel boundaries there, not the centre line.

 
Yousufkhodja Sultonov:

Recipe for 1 kg rice (for 6-7 people):

1. Vegetable oil - 400g;

Onions - 1 kg;

Beef (lamb) - 1 kg;

4. carrots - 1 kg;

5. Rice - 1 kg.

Bring oil to a boil, add onion and fry until red, add meat and fry for 5-10 minutes, add half of carrot, fry for 10 minutes, add remaining carrot, fry for another 10 minutes, add water, until all mixture is covered with water, gently boil for 30 minutes, add rice, add water until rice is 1 cm over water. Boil on a high heat until the rice has softened and all the water has evaporated and the rice is saturated. Reduce heat to minimum, cover the cauldron with something for 20 minutes. All, pilaf is ready.

Yusuf Hodji, I understand that your problem is that it's difficult to grasp the main point; you are taking in everything at once.

You were asked what is the main ingredient of pilaw and you gave a recipe for cooking it. I agree that there is not only one main component of pilaw and I make it in a different way but I would not be worried by an answer like "zirvak" or "rice". Even an answer like paez, aez, lahm, olea... would have been fine too. But you don't need a concept - you need a result straight away.

You practically reached the model of Oleg (automaton); only he has linear differential equations, and you have algebraic equations of the same dimension.

 
Алексей Тарабанов:

Yusuf Hodji, I understand that your problem is that you find it difficult to grasp the main point; you take in everything at once.

You were asked what is the main ingredient in pilaf, and you gave me a recipe for cooking pilaf. I agree that there is no single main ingredient of pilaw and I make it in a different way but I would not be worried by an answer like "zirvak" or "rice". Even an answer like: paez, aez, lahm,

Ah, well done Alexei.

 
Yousufkhodja Sultonov:

You see, even in the case of a linear combination, things are considerably more complicated than one might imagine. Yuri, just because someone doesn't know or hasn't done it, doesn't mean you shouldn't try it. The first results will come soon and there will be food for thought. It's a promising start.

By the way. Bring the system to a canonical form, and any matrix library has everything for solving such systems. And work with any order. Within reasonable limits. )
As far as I remember, even Excel can do it.
 
Yousufkhodja Sultonov:

In that sense, you are right. We need 10 prices, not 8:

n x1 x2 x3 x4 y
1 1,1356 1,1367 1,137 1,1361 1,1361
2 1,1367 1,137 1,1361 1,1361 1,1356
3 1,137 1,1361 1,1361 1,1356 1,1359
4 1,1361 1,1361 1,1356 1,1359 1,1361
5 1,1361 1,1356 1,1359 1,1361 1,1364

Why guess when you can do a quick estimate with Excel? I was already sceptical, but it's quick to check, so why just throw words around?

You need 4 equations for 4 unknowns. Suppose for the first 4 lines x1i*k1 + x2i*k2 + x3i*k3 + x4i*k4 = y`, i.e. i is a line number.

Then we calculated the error in squares (to reduce to MOC) - (y` - y)^2 and in points, to be more clear how much we have made a mistake.

Then, all coefficients can be quickly calculated by the "Find Solution" function, minimizing the error (sum of squares).

Then I checked the same coefficients on the 5th line. So? The error is 9 points, taking into account that other values even fluctuate to smaller ones, i.e. a thumb in the sky...



The main question arises - why suddenly decide that this theory can be valid? I tried a lot of things once and it seems that randomly selecting all methods is not the best idea, at least that's what I decided for myself.

 
Galina Bobro:

Why guess when you can do a quick estimate with Excel? I was sceptical as it was, but it's quick to check, so there's no need to just throw words around.

You need 4 equations for 4 unknowns. Suppose for the first 4 lines x1i*k1 + x2i*k2 + x3i*k3 + x4i*k4 = y`, i.e. i is a line number.

Then we calculated the error in squares (to reduce to MOC) - (y` - y)^2 and in points, to be more clear how much we have made a mistake.

Then, all coefficients can be quickly calculated by the "Find Solution" function, minimizing the error (sum of squares).

Then I checked the same coefficients on the 5th line. So? The error is 9 points, taking into account that other values even fluctuate to smaller ones, i.e. a thumb in the sky...



The main question arises - why suddenly decide that this theory can be valid? I have tried many things in the past and it seems that randomly selecting all methods is not the best idea, at least that's what I've decided for myself.

This proves absolutely nothing. )
Wait, don't disturb the artist, Yusuf will invent linear regression in a little while).
 
Roman Shiredchenko:

WAAA!!!!!!!!! :-)

Uncles - cut the crap!

URM, not to be confused with BDSM - rules.

read, listen, use, do stuff!

2Jusuf - don't stoop to bullshit :-)

"The most trivial trading strategy is page 46"

*fuck... that was 8 years ago...

Speak of the devil... It won't be anything, but still, aye-aye, so to speak.
 
Yuriy Asaulenko:
This proves absolutely nothing. )

This and much else above proves only one thing - it's all about working with price derivatives, whatever they may seem to you.

 
Galina Bobro:

Why guess when you can do a quick estimate with Excel? I was sceptical as it was, but it's quick to check, so there's no need to just throw words around.

You need 4 equations for 4 unknowns. Suppose for the first 4 lines x1i*k1 + x2i*k2 + x3i*k3 + x4i*k4 = y`, i.e. i is a line number.

Then we calculated the error in squares (to reduce to MOC) - (y` - y)^2 and in points, to be more clear how much we got wrong.

Then, all coefficients can be quickly calculated by the "Find Solution" function, minimizing the error (sum of squares).

Then I checked the same coefficients on the 5th line. So? The error is 9 points, taking into account that other values even fluctuate to smaller ones, i.e. a finger in the sky...



The main question arises - why suddenly decide that this theory can be valid? I have tried many things and it seems that randomly selecting all methods is not the best idea, at least that's what I've decided for myself.

1. In my case, guessing and guesstimating is out of the question. I estimate 5 coefficients by ANC, using variation and covariance of variables for exact determination of C5 based on solving system of 5 equations.

2. In addition to 4 coefficients at 4 variables, in the final solution of the system appears a0 free coefficient, which you have not considered in your hasty and incompetent reasoning. Here is the exact solution of the above system, without any, residuals:

Ц1 Ц2 Ц3 Ц4 C5 actual. a4 a3 a2 a1 a0 C5 calculated
1,1356 1,1367 1,137 1,1361 1,1361 0,13082 0,10459 -0,48492 -0,23672 1,688584 1,1361
1,1367 1,137 1,1361 1,1361 1,1356 0,13082 0,10459 -0,48492 -0,23672 1,688584 1,1356
1,137 1,1361 1,1361 1,1356 1,1359 0,13082 0,10459 -0,48492 -0,23672 1,688584 1,1359
1,1361 1,1361 1,1356 1,1359 1,1361 0,13082 0,10459 -0,48492 -0,23672 1,688584 1,1361
1,1361 1,1356 1,1359 1,1361 1,1364 0,13082 0,10459 -0,48492 -0,23672 1,688584 1,1364

The calculation is based on the formula: Ts5 = a0 + a1C1 + a2C2 + a3C3 + a4C4 and a full match to fact is guaranteed.

3. Then calculated the errors and decided that "a finger in the sky...", exposing your illiteracy in this area.

4. Indeed, "The question arises in chief" - to what extent does the price of the 5th bar depend on the prices of the 4 previous bars? The answer is: it depends entirely! Moreover, this fact has now become a stone to those who believe that, price does not depend on history! How much the price depends on history is a subject for further research in a similar vein. Take 10, 20 or more bars deep into history.

 
Yuriy Asaulenko:
This proves absolutely nothing. )
Hang on, don't disturb the artist, it's only a matter of time before Yusuf invents linear regression).

Indeed, "Wait, don't interfere" and you yourself will be ashamed of such a mockery. By the way, I have to remind you that I was the one who extended Gaussian MNA to the field of non-linear regression, which absorbs Gaussian MNA for linear regression, in case you forgot about URM. But, this contribution will be appreciated by posterity, as no one ever noticed such "triviality".

Reason: