If you were playing chess and could do whatever you wanted, you could take several goes in a row, take the opponents pieces off the board and move yours into a winning position. How does that help you play better under the constraints of the rules?

This metaphor doesn't really work because the entire point of a game—the thing that makes playing it playful—is that it has no effect on the world outside of the game itself. Thus, ignoring the rules of chess to reach a goal doesn't make sense. There are no chess goals that don't involve the game of chess.

This isn't true in programming or real-world tasks where you are trying to accomplish some external objective.

If you were playing chess, and you could do whatever you wanted, you might want to, e.g., set up a beautiful mating combination using the minor pieces, set a brilliant trap with forced mate in 10 moves and then trick the opponent into falling into it, keep control of the center all game and make the opponent play in a cramped and crippled style, promote a pawn to a knight for a win, skewer the queen and king, or turn a hopelessly lost position into a draw. The constraint is: you need to take turns and the opponent wants to win / doesn't want to let you win.

If you don't know you could take the king and win the game why would you bother with any of that?

It’s actually useful technique in chess to see what you would do if you could make multiple moves in a row.

If I could move my Rook there it’s a win, is there any way I can make that happen? How about if I sacrifice my knight etc.