V. P. Yakovlev

The classic probable events as binary sequences test are considered, where the ones are positive outcomes. The probabilities were the Cesaro limit and were equals the altitudes of positive outcomes numbers to the whole numbers as its limit to the infinity. The sequence numbers for given probability corresponds the intervals ±ε nearby probability limit. The «correct» numbers in Shannon information theory was practically equals to whole sequence numbers for probability 0,5. The independent one-half probability sequences number from given sequence was practically equal to full its numbers. So the events in systems was independent, if it was absent the operative information exchange between its. There were absent the dependence between the events in physical systems without energy exchange. There was independence between systems example, if the information change takes place. The sequence aggregates with the same probability are the Rytov ensemble or subspace of Kolmogorov probability space. The probability means of the discrete numbers was equal to arithmetic means of test results with stringing its number to infinity. The test result for statistic process is given time function. The process value for fixed time is probable number realization. Its mean is ensemble mean and is compared with the time mean by the ergodic theory. The same results take place in mathematic statistics before replacing the realizations by «virtual» process and adding the independence proposition.