Probabilistic Theory of Stock Exchanges. Anatoly Kondratenko

complex issues concerning the exchange. So, by definition and in its essence, the probabilistic function of demand D(p,q) (supply S(p,q)) is the probability of the buyer (seller) concluding a deal to buy and sell the traded goods in quantity q at price p. If this is so, then, according to the standard concepts of probability theory, it is natural to define the probability of the transaction under these conditions as the multiplication of these probabilities:

      We call this probability of making a deal a market deal function, and, for convenience, we also refer it to the market functions of supply and demand. Like the market functions D(p, q) and S(p, q), it is dimensionless. For the sake of certainty, let us explain that, generally speaking, purchase and sale transactions can occur in the market at any time, at any price and in any quantity, within reasonable limits, but with varying degrees of probability. But if the transaction function is a sufficiently high and narrow bell with a single maximum with the parameters p^M and q^M, then almost all transactions will occur in the proximity of these values, so it is reasonable to consider these very values to be market prices and quantities. If the function of transactions looks otherwise, of course, these definitions are somewhat meaningless, and one should consider the mechanism of probabilistic pricing in detail. Below we will always assume that the function of transactions is such as to allow market prices and quantities to be determined in a fairly simple way. This is exactly the case we have graphically presented in Fig. 1.10 for our two-agent model of the grain market.

      Fig. 1.8. Graphical representation in a rectangular three-dimensional coordinate system [P, Q, S&D] of the two-dimensional buyer demand function as a three-dimensional surface D (p, q) with a maximum at the point A (p^D, q^D) in the plane (P, Q).

      Fig. 1.9. Graphical representation in the rectangular three-dimensional coordinate system [P, Q, S&D] of the two-dimensional seller's supply function as a three-dimensional surface S(p, q) with a maximum at point B (p^S, q^S) in the plane (P, Q).

      As expected, the surface of the market transaction function F(p, q) has only one maximum. For multi-agent economies, the structure can be much more complex.

      Fig. 1.10. Three-dimensional graphical representation in a rectangular three-dimensional coordinate system [P, Q, F] of the three-dimensional deal surface F(p, q) in the form of a high and narrow bell with one maximum at the point C (p^M, q^M) in the plane (P, Q). The graphical method of calculation gives the following results for market prices and quantities: p^M = 281.4 $/ton, q^M = 51.9 ton/year.

      Let us now turn to the question of calculating market prices and quantities within the framework of probabilistic economics. It is well known from the standard course of mathematical analysis that extrema of a multidimensional function should be defined as points on the corresponding surface in which the total differential of this function is 0. In our situation this condition leads to the following equation:

      This equation is equivalent to the following two partial derivative equations:

      In terms of S&D functions, this system is transformed as follows:

      At this point it makes sense to introduce a new concept into theory, namely the concept of S&D market forces with such definitions:

      In terms of market forces we can write the system of equations (1.25) more compactly as follows:

      Obviously, this system of equations looks like a system of equality of S&D market forces at values of market prices and quantities. And it is similar to the system of forces equality at the static equilibrium point in classical mechanics. In other words, the system of economic equations (1.27) looks like a formulation of Newton's third law in classical mechanics. Substituting specific S&D functions from equations (1.16) and (1.20) into the system of equations (1.26), we obtain such simple and clear formulas for calculating market forces:

      As we can see, all forces have become one-dimensional functions in this model. Then, using these equations, we obtain a very elegant system of two independent linear equations to determine market prices and quantities:

      This system is so simple that you don't even have to solve it in the usual sense to get a very nice looking solution for market prices and quantities:

      Thus, probabilistic market prices and quantities in a two-agent economy, when using factorized agent functions in the form of Gaussians, are determined by averaging the corresponding agent parameters, with the frequency parameters of the agents serving as weights in this averaging. The fundamental point here is that these two simple, and independent, algebraic formulas, which include only four buyer parameters (p^D, q^D, w^DP and w^DQ) and four seller parameters (p^S, q^S, w^SP и w^SQ), uniquely determine the market price p^M and the market quantity q^M. Let us discuss some of the most prominent features of the resulting formulas.

      Feature 1. Market price and market quantity do not depend on each other in this model at all. Keep in mind that p^M and q^M in their meaning are the most probable price and quantity, the two most important conditions under which a deal is most likely to be concluded. This means that these two formulas unambiguously set the trend in the development of market dynamics at a given moment in the direction of these very values. There is practically no need to make any complex calculations. This fact is the advantage of this model.