I always like change in allele frequency over time.
I think that's probably the shoddiest definition of evolution that gets kicked around; it is incomplete, it does not uniquely define evolution, it explains nothing and it completely fails to capture the grandeur of evolutionary theory.
It is incomplete because it does not make any mention of the mechanisms of evolution, or the historical reconstruction of evolution.
It does not uniquely define evolution because anything that changed allele frequency would be included - be it an utter randomly fluctuation or the subtle machinations of His Noodly Appendage.
It explains nothing because it does not include the reasons why allele frequencies change or the ways in which this produces change.
And, as a result of these failings, it completely fails to capture the grandeur of evolutionary theory; mentioning nothing of the overarching view of life it gives, the breadth of its explanatory power or the depth of evidence for this explanation.