Illustration of the binomial distributing an example serves as with throwing dice up. Thus an experimenter is interested by probabilities of “success” (falls a verge with a certain number, for example, with “six”) and “failure” (fall a verge with any other number).

Distributing of Puassona is used, when the followings terms are executed:

1.Every small time domain can be examined as experience the result of which is one of two: either “success” or his absence is a “failure”. Intervals are so small, that can be only one “success” in one interval probability of which is small and unchanging.

2.The number of “successes” in one large interval does not depend on their number in other, I.e. “successes” are helter-skelter sparse on temporal intervals.

3.The middle number of “successes” is permanent at all time.

Distributing of Puassona is ordinary illustrate the example of registration of amount of travelling incidents for a week on the certain area of road.

At certain terms, distributing of Puassona can be used as approximation of the binomial distributing, that especially comfortably when application of the binomial distributing requires difficult, labour intensive calculations, taking away much time. Approximation guarantees acceptable results at implementation of next:

1.Количество experiments great, preferably more than (n=3).

2.Вероятность “success” in every experience small, preferably less than 0.1.(p=0.1) If probability of “success” is great, then for replacement normal distribution can be used.

3.Предполагаемое amount of “successes” 5 (np=5).

In the cases when binomial distributing very labour intensive, him it is also possible to approximate normal distribution with a “amendment on continuity”, I.e. doing assumption, that, for example, value of discrete casual quantity 2 is the value of continuous casual size on an interval from 1.5 to 2.5.

Optimum approximation is arrived at at implementation of the followings terms: n=30; np=5, and probability of “success” of p=0.1 (optimum value of р=0.5)

It should be noted that in literature and practice besides statistical criteria other indexes of measuring of risk are used: size of loss of profit, received less profit et al, expected, as a rule, in monetary items. Sure, such indexes have a right on existence, moreover, they frequently simpler and more clear what statistical criteria, however for adequate description of risk they must take into account his probabilistic description.

On the basis of the conducted analysis an author offers the generalized complex criterion is a «cost of risk» which characterizes the size of conditional losses possible during realization of investment decision:

determined as a sum of possible direct losses from an investment decision.

For determination of cost of risk it is recommended to use such indexes, which take into account both co-ordinate of «vector», both possibility of offensive of unfavorable event and size of harm from him, only. As such indexes an author suggests to use dispersion, rejection and coefficient of variation. For possibility of economic interpretation and comparative analysis of these indexes it is recommended to translate them in a money format.