Practice question · Multiple choice
A simulation reproduces the experimental data closely. Why is that weaker evidence of correctness than it appears?
Hints
- Ask how many parameters were adjusted to achieve the fit.
- What would a model have to do to earn more credit than fitting the data it was built on?
Show the answer
A. Because enough free parameters can be tuned to fit almost anything
Why
Fitting data you tuned against is nearly guaranteed, so the informative test is prediction, hold data back, or forecast a regime the model never saw. Von Neumann's line applies: with four parameters he could fit an elephant, and with five make it wiggle its trunk.
Practise Scientific Computing in Practice
The app has 5 more questions on this lesson, and keeps your place in the course. Mathematics I is free to start.
More questions on Scientific Computing in Practice
- Order the stages of the scientific computing workflow.
- A three-dimensional simulation halves its grid spacing in all three directions and also halves its time step.…
- A second-order scheme has an error of 0.08 on the current grid. Refining the grid by a factor of 2 changes…
- If halving the step size leaves the answer essentially unchanged, the answer is correct.
- Sort each claim by whether the computation described actually justifies it.
- A simulation can produce plausible-looking output while being completely wrong. What distinguishes…
- Select every legitimate way to validate a numerical result.