Mathematics competitions involve nonroutine problems and most mathematics courses more heavily emphasize the routine (not to be confused with easy!) components of various subjects. Rather than addressing the original question in its fullness, here is just a fragment:
Can students who are doing well in the Calculus sequence (say, they have taken Calc I, Calc II, Calc III, and are currently taking Differential Equations) solve nonroutine Calculus problems?
The answer is mostly no. See, for example:
Selden, Annie, et al. "Do Calculus Students Eventually Learn to Solve Non-Routine Problems? Technical Report. No. 1999-5." Online Submission (1999). Link (no paywall!).
Even students with full/substantial requisite knowledge of the corresponding mathematical topic, who are presently studying in a relevant mathematics course, are unable to provide completely/substantially correct solutions to nonroutine problems – examples of which are included in the linked study above, and none of which is near the difficulty of a typical Olympiad problem.
Readers may find that this answer offers little else by way of insight as pertains to the original question, although I wish to advocate for first answering a weaker version of a question (can students doing well amid the Calculus sequence and who have the requisite knowledge successfully solve nonroutine Calc problems?) before zooming outward to a stronger question (can students who may be far removed from the material – e.g., years away from a geometry course – and/or have never seen the material – e.g., unfamiliar with stars and bars, proof by induction, generating functions – successfully solve Olympiad style problems?).
I hope the Selden et al piece is of interest for those who are curious about the manifestation of the routine/nonroutine phenomenon presented by the Putnam, IMO, etc.