Historically, single-player strategy games have primarily been focused around a campaign-based structure (see StarCraft, WarCraft, Age of Empires, etc). While it is possible to play these games single player against “bots”, there is very little or no variation in the starting conditions, so a strategy that you devise that works once will continue to work (not to mention that AI bots are typically not that fun to play against anyway, which is an article for another time).
The problems with this in my mind in terms of designing a strategy game are twofold. For one thing, as soon as the player has devised a solution that works every time (i.e after one match in the above examples, or after beating a mission once), the game has become a solved puzzle, which decreases the value of interacting with it considerably. Secondly, the sense of STAKES are greatly reduced in any given play session, because play can simply becomes a matter of trying solutions to the problem until one works.
Contrastingly, if a game takes its cues from multiplayer game designs (i.e sports, MOBAs, FPS games, fighting games, card games, etc) it can find answers to both of these problems. In these games, a play session is comprised of one or more matches in which some agent (usually the other team) makes moves or plays that a player must react to. Then, based on some combination of their execution and strategy as measured against their opponent, they are given clear, distinct feedback about how good that approach was (a win or a loss). This allows players to create and refine generalized inferences about the game system rather than specific solutions to static starting parameters.
Equally as important, a clear win/loss state provides players with a sense of accomplishment. If a game simply provides a “high score” system with which to judge your approach to the game, it does give feedback about the effectiveness of your approach, but each score always comes with the nagging sensation that it “could have been better”. A win tells the player “that was good enough!” rather than simply “that was X good”. Also of importance are strategies specifically tailored to have maximum payoff right at the end of the game, which is a facet of strategic thinking that many well-designed strategy games encourage.