The inverse distance squared, like all modeling algorithms, is a mathematical method for interpreting (within the data) or interpolating (outside the data) values for grid nodes or blocks in a computer mining model. 

Think of a scatter of data points on a planar surface that represent survey elevations. Over this is placed a grid with nodes every 25 meters apart. With the inverse distance squared method the value at each grid node is calculated as the average of the surrounding real values weighted by the inverse of their distance squared from the grid node. So if one value is 2 meters from the grid node and the other is 10 meters from the grid node the closest value will have 25 times as much weight in the grid node value determination as the furthest value. 

It can be seen that if there is a lot of data and a large grid then a computer is needed to calculate the grid values. In fact there are many computer programs which make these calculations and large grids can be calculated in a matter of seconds. The extrapolation of grid values beyond the data points requires special attention and care must be taken to ensure that the grid is calculated appropriately to the geological or other prevailing conditions. Because it was calculated by a computer does not make it correct.