Principle of never throwing away information

Source: rlabbe

“Two sensors are better than one, even if one is less accurate than the other.”

Example 1

Given:

• A: a more accurate sensor
• B: a less accurate sensor

Should we therefore choose A as the estimate and discard B? No, because B can improve our knowledge when combined with A.

Using the measurements from B further narrows the range of estimates (overlap between the error bars).

Example 2

Assume:

• Scale A has an error of +- 1kg
• Scale B has an error of +- 9 kg

We measure: A = 160 kg B = 170 kg

Even though B is highly inaccurate, it still helps to narrow down the range of estimates.