Interesting to think that a successful experiment is one that disproves a theory, or, in other words, one where the numbers aren't the ones one expected.
A successful experiment can also support a prediction. I.e., "using this model, we predict that a bagel with blueberry cream cheese will remain on the counter for 94+/-4 seconds before disappearing." The team will then release the data from their experiment, which either supports or disproves a model.
This kind of work is critical to refining theories and highlighting/fill gaps in knowledge.
