Suppose we have two machines. Each machine measures things in a particular category. For example, one measures the weight of cars, and the other measures the weight of butterflies.
We have available many measurements by each machine in their category, and since we also have available the objective weights, we have available their errors.
Now we pose the question, which machine is most accurate? Obviously, the machine measuring the weight of butterflies has less error, but the comparison is unfair, because cars weigh much more on average than butterflies.
How can we bring the errors of each machine onto the same scale to compare how good they are at measuring?
