[STATSGRASS] gstat: problem with prediction value at point location

[Jarek Jasiewicz]

Hi
when I tested prediction values kirging result I found that values at 
pixel exactly at point location has the same values as points. It gives 
that in points appear thin peaks with values strongly different from 
neigbours. It looks like gstat do not use prediction values but used 
exact values.
Variogram I used had the nugget

Jarek