Hello Micha,
I calculated standard deviation for easch direction, according to horizontal distance.

I got these results:

standard deviation of direction in "

distance value from        distance value to standard deviation of direction in "
 0                        20                     80
20                        30                     60
30                        40                     42
40                        60                     28
60                        70                     21
70                        100                     15
100                        120                     10
120                        150                      7

I have small distances in network, so I got these results. For example, if distance is between 0m and 20m, standard deviation of direction is 80" etc.
How to define that in stochastic model - martix P, since I have various weights for almost each direction. For distance weights, I have no problem to define.
For this calculation I used formula
σ_α=arcsin⁡((σ_DV×√n)/(d[mm]))×ρ"

σ_D is standard deviation of points (needed to be)
n - number of directions from one point to destination points in network configuration
d - distance in mm

Is that possible to define in JAG, or I need to use some averaged value (in this case of various distance, averaged value wont respresent network model satisfactory)

I hope that you can help me.
Best regards,
Zaklina