Weaker notion of subgame perfect equilibrium

Question

It is well known that, for extensive-form games with perfect information, the notion of a Nash equilibrium is too weak, because a game may have Nash equilibria that involve "incredible threats". For this reason, the notion of a subgame perfect equilibrium (SPE) was defined, which is a pair of strategies that form a Nash equilibrium on every subgame.

However, it seems to me that this notion is actually too strong.

For example, suppose the following game:

• In the initial state, player 1 has two possible actions: A and B.
  • If player 1 chooses A, then player 2 has two possible actions: C and D
    • if player 2 chooses C, then the game ends with payoffs 100 and 90 for the two players respectively.
    • if player 2 chooses D, then the game ends with payoffs 90 and 100 respectively.
  • If player 1 chooses B, then the two players will play some really complicated game, say chess.
    • the winner of the chess game receives a payoff of 2,
    • the loser receives a payoff of 0,
    • and in case of a draw, both players receive a payoff of 1.

Finding the SPE of this game would be a hopeless task, because it would require solving the entire game of chess. However, it is immediately clear that the optimal strategy for player 1 is to choose A, and for player 2 to then choose D. For the states of the chess game, we can just select arbitrary dummy moves, since they will never be played anyway.

While the pair of strategies obtained this way will technically not be an SPE, it does capture its essence, because it's a Nash equilibrium that does not suffer from any incredible threats. So, I think there should be some intermediate solution concept, which one could probably describe as: "a pair of strategies that results in the same terminal state as the terminal state that would be reached under the subgame perfect equilibrium.". Such a solution concept would actually seem more useful to me than the SPE.

My question: is there a name for the solution concept that I just described? (or similar, that captures the essence of my problem) I can't imagine that this hasn't been described in the literature.