The Fairness Theorem - page 3

 
faa1947:

Dear Yusuf!

You are deeply mistaken, even though you are repeating the unchallenged pre-1987 theory called the "efficient market".

I once saw an article discussing the eligibility of a US antitrust committee in one state on the grounds that the antitrust rule was broken in that state: there must be more than 1000 sellers and more than thousand buyers of the same product.

This is a necessary but not a sufficient condition of your rules.

Because there are panicked price movements (stock market crashes) and deliberate price manipulation. On the latter question even a branch was opened on the forum. It was discussed in detail.

Everyone can be mistaken, but these "creators of market theories" you mention did not bother to understand and quantify the market price itself and are still saying that it cannot be quantified, attributing it to an unknowable virtual price. Meanwhile, it turns out that if the market price is the arithmetic mean of Ts1 and Ts2, the optimal price - is the geometric mean of these prices, and their difference - is the so-called "Cauchy difference", on which the maximum profit depends. You, too, have a habit of speaking freely about prices, so be careful.
 
yosuf:
Perhaps if Akio Morita refuses to sell his product at one of the prices listed above, he is no longer a market player, and his moves are accounted for by the market's changing elasticity and market capacity, determined by experimentally selling several batches of product at different prices. Sellers and buyers already know that he has a large batch of goods and this fact will indirectly lead to an adjustment of the market price. He is not going anywhere, sooner or later he will still sell his goods within the range of C7 - Cpr.

This is an oversimplified approach that will give too much error in forecasting. It only works in the potato market. But potatoes are only a small part of the economy. Modern economics works ON ORDER.

Akio Morita did not have a batch of goods. He had the knowledge, experience, equipment and employees to produce a large batch of transistor radios, which were not available in the USA AT ALL.

A US customer made Morita very happy with the order and said that he wanted a volume discount. Morita explained that he could not give a discount, because the contract would be for one year, and if it was not renewed, he would have to fire a lot of people and sell the machines he would hire and buy to fulfill the contract.

Based on my brief work with Samsung's top executives, I can tell you that this is how they operate now. The Koreans take into account a lot more circumstantial factors in their work that modern Americans don't pay attention to at all. The Germans also know how to count the little things (I judge from my work in a German company called World Top-100), but with them it is meticulousness, whereas Koreans have a lively attention to details that can delay (read - ruin) the whole project.

Yusuf, the American company Wallmart does not BUY a product, it .... orders it, and specifies all the desirable consumer properties, quantity, AND the PRICE for it. Any supplier who doesn't agree is screwed. And as a rule, Wallmart is not wrong. That's why Wallmart has the best IT department in the WORLD, their own special databases (not Oracle) and their own IT institute. They have the best marketing in the world. European retailers have not yet achieved this, but they also have a system of long term ORDERING through local distributors.

The Chinese are ready to produce ANYTHING, just name the properties of the product and the price you need to meet. And they will do it.

GlobalSources Chinese monthly electronics OEM directory weighs and looks like a small suitcase.

See? The ORDERED nature of the economy is in effect. It's not selling a bucket of potatoes.

 
Demi:

Remembered! Linear programming))))))

Finding an Optimal Solution - The Resource Optimization Problem in Production Planning

Target function, constraints, finding the optimum.

This is not a theorem

In terms of linear programming it is formulated as follows - if the LP problem is not degenerate, then the optimal plan can be found


You can find it through linear programming as well. How to make a system of linear inequalities I pointed out in the text.


PS. If the LP problem is degenerate it is not a problem. In linear programming such situations have already learned to handle successfully.

By the way, I implemented the arbitrator through linear programming in Java. I send it via a script the Asc and Bid values for 28 currency pairs (8 currencies) and put it all into a payment matrix. The arbitrator finds an arbitrage situation very quickly, i.e. the most favourable combination. But the problems are:

  1. The discreteness in the solution does not always fit the lot. I.e. in order to get more or less acceptable discreteness, it is necessary to open some deals with big volumes.
  2. It is one thing when arbitrage is done by a couple of trades. It is another thing when it is necessary to make 4-5 deals at a time for arbitrage. Requotes and slippages are unlikely to make a profit.

For this reason I have not started to port the code to mql. For an ordinary trader this system is most likely useless, i.e. it is hardly useful for home and family. The only way to make money with this method is on broker's server (if there are a lot of clients) or on trading platform's server. And most financial instruments are not suitable, because they are not tied to each other through barter, like currency pairs, so the arbitrage is negligible if all the goods are quoted only in money.

 
yosuf:
Everyone can be mistaken, but these "creators of market theories" you mention did not bother to understand and quantify the market price itself and are still saying that it cannot be quantified, attributing it to an unknowable virtual price. Meanwhile, it turns out that if the market price is the arithmetic mean of Ts1 and Ts2, the optimal price - is the geometric mean of these prices, and their difference - is the so-called "Cauchy difference", on which the maximum profit depends. You, too, have a habit of speaking freely about prices, so be careful.

Come on, there are plenty of noobiles out there. The theory flourished and still works for portfolio managers in monotonically growing markets. And then bang - and then the crisis. And the whole theory went down the drain. Together with the portfolioists.
 
faa1947:
Or you can give money to an official and sell the goods at non-market prices. The practice you describe does not exist, but mine does.

We are talking about purely market mechanisms of price formation, and there is a corruption component everywhere, as well as monopoly and oligopoly, which the above mechanism with the elasticity and volume of the market will take into account.
 
Reshetov:


By the way, I have implemented arbitrage through linear programming in Java. I send it via a script the Asc and Bid values for 28 currency pairs (8 currencies) and put all this stuff into a payment matrix. The arbitrator finds an arbitrage situation very quickly, i.e. the most favourable combination. But the problems are:

  1. The discreteness in the solution does not always fit the lot. I.e. in order to get more or less acceptable discreteness, it is necessary to open some deals with big volumes.
  2. It is one thing when arbitrage is done by a couple of trades. It is another thing when it is necessary to make 4-5 deals at a time for arbitrage. Requotes and slippages are unlikely to make a profit.

For this reason I will not port the code to mql. For the average trader such a system is most likely useless, i.e. it is hardly useful for home and family. The only way to make money with this method is on broker's server (if there are a lot of clients) or on trading platform's server. And most financial instruments are not suitable because they are not tied to each other through barter, like currency pairs, so the arbitrage is useless if all the goods are quoted only in money.


Yes, I had that idea too, but didn't get the hang of it. Variables - lots for each of the pairs. Restrictions - total amount of capital, non-negativity of lots and probably only integer values.

And the target function?

 
Demi:

Yes, I had that idea too, but didn't get the hang of it. The variables are the lots for each of the pairs. The constraint is the total amount of capital, non-negativity of lots and maybe only integer values.

And the target function?



The variables, i.e. unknowns, are lots not just for each pair, but for Ask and Bid of each pair a separate column should be started. After all, we need to know what to buy and what to sell in order to benefit from arbitrage.


The constraint, as stated in the text, must be less than or equal to 0 for each linear inequality.

For payment matrices, the target function is to maximise or minimise the sum of all the unknowns. In this case you have to look for the minimum.

It is also necessary to bring the payment matrix to a form so that it contains only positive numbers, i.e. all values should be increased by the minimum value in the matrix. See http://math.seme str.ru/games/linear-programming.php for details. There is no need to convert the solution back, because increasing or decreasing all matrix values by a constant has no effect on the final result.

 
AlexEro:

This is an oversimplified approach that will give too much forecasting error. It only works in the potato market. But potatoes are only a small part of the economy. Modern economics works FOR ORDER.

Akio Morita did not have a batch of goods. He had the knowledge, experience, equipment and employees to produce a large batch of transistor radios, which were not available in the USA AT ALL.

A US customer made Morita very happy with the order and said that he wanted a volume discount. Morita explained that he could not give a discount because the contract was for one year, and if it was not renewed, he would have to lay off a lot of people and sell the machines he would hire and buy to fulfill the contract.

Based on my brief work with Samsung's top executives, I can say that this is how they work even now. The Koreans take into account a lot more indirect factors in their work that modern Americans don't pay attention to at all. The Germans also know how to count the little things (I judge from my work in a German company called World Top-100), but with them it is meticulousness, whereas Koreans have a lively attention to details that can delay (read - ruin) the whole project.

Yusuf, the US company Wallmart doesn't BUY goods, it .... orders it, and specifies all the desirable consumer properties, quantities, AND the PRICE for it. Any supplier who doesn't agree is screwed. And as a rule, Wallmart is not wrong. That's why Wallmart has the best IT department in the WORLD, their own special databases (not Oracle) and their own IT institute. They have the best marketing in the world. European retailers have not yet achieved this, but they also have a system of long term ORDERING through local distributors.

The Chinese are ready to produce ANYTHING, just name the properties of the product and the price you need to meet. And they will do it.

GlobalSources' Chinese monthly catalogue of electronics OEMs weighs and looks like a small suitcase.

See? The ORDERED nature of the economy is in effect. It's not selling a bucket of potatoes.


I was a guest of the founder of Wallmart back in '95, though when the board was taken over by his sons. We were told how the not-so-young owner started with a small shop and created a retail empire according to the principle you mentioned, I was given his book.

In essence, any producer or intermediary in the form of sellers eventually has to face and surrender to market laws. And crises are an inevitable attribute of the market economy, its costs, but, as they say, humanity has not invented anything better than this evil.

 
Reshetov:
But the problems are:
  1. The discreteness in the solution does not always fit the lot size. I.e. in order to get more or less acceptable discreteness, you need to open some trades with large volumes.
  2. It is one thing when arbitrage is done by a couple of trades. It is another thing when it is necessary to make 4-5 deals at a time for arbitrage. Requotes and slippages are unlikely to make a profit.

For this reason I have not ported the code to mql. For an ordinary trader this system is useless. To make money with the method mentioned in the text, you can only on the server of the broker (if there are a lot of clients) or on the server of a trading platform. And most of the financial instruments are not suitable, because they are not related to each other through barter, as in the case of currency pairs, so the use of arbitrage if all the goods are quoted only in money comes to naught.

I have some vague doubts...... whether Metakvot ordered our dear Reshetov a good matching algorithm, which Metakvot needs to expand brokerage areas...... and therefore we are chosen as the cats to train .....
 
AlexEro:
I have some doubts...... whether Metakvot have commissioned our dear Reshetov with a good matching algorithm, which Metakvot needs to expand the brokerage sites...... and therefore we have been chosen as the cats to train .....

And even if they did, what good would it do for metaquotes to make the algorithm public? Don't they already have a working server that acts as an interbank for methaquot brokers? I read such information somewhere. Probably on the main website?
Reason: