The statistical distribution was developed starting from the values approximated
to the nearest metre, contained in the general table. The relative percentile frequencies
were calculated in respect to the total number of control points contained in the
different tests for each coordinate. This number is written in the same table, in
special columns. The percentile values of the frequencies are calculated with only one
decimal figure.
The last two groups of rows contain the number of points that have not been
plotted in the different tests, and the number of control points used for adjustment
of the blocks, respectively.
We consider the values of the mean and of the variance of these statistical variables
scarcely significant, therefore they have not been written in table.
The analysis of table 2 permits an immediate view of the asymmetry of some
blocks in respect to the 0. In this connection it is useful to consider the 5 central
classes of frequency, included within — 15 and -f 15 m. The remaining classes of
frequency are, in fact, much more influenced by anomalous points, namely by the
points with errors in the coordinates much higher than the average and due to excep
tional reasons.
First of all, one can note that the percentage of the errors equal to 0 is in correla
tion with the number of points used for the blocks adjustment. This number is ranging
from 3,1% in the X for test 11, to 24,7% in the Y for test 18. Test 11, in fact, was
adjusted with only 4 points in planimetry, while test 18 was adjusted with 38 points.
This statistical table of errors appears very significant as it permits more imme
diate comparisons among the results of the tests. That is, a comparison may be made
between those tests obtained by adjusting in different ways the same measures, or
those tests obtained from measures executed with different instruments and photo
graphic material, but adjusted on the same control points and by the use of the same
computational procedures.
145
10