> You don't need to think about graphs or runtimes every day- but that one time each year or so that you do need to know it, it's critical that you do.

That one time a year when it's critical, I would want someone who is likely to engage on the rest of the team to come up with a solid solution together and spend at least a day thinking about it.

This isn't how interviews work so if you think you're selecting for that one time a year when its critical at the expense of the rest of the year, I don't think you're getting what you think you are.

For quick workarounds to deal with a critical error while allowing time to solve it properly, I think an experienced person may have an edge.