I'm partial to the Goode Homolosine projection, next is whatever NatGeo standardized on. Of course, the best is an actual globe --but decent sized ones are too expensive and the beach ball ones are just not very well made.