A calculation is run but the result exceeds the total storage allocated for it
Each computation is rounded off to a certain number of digits, which can accumulate over time leading to loss of precision