Today we consider the opening 1. e4 e5 2. Qf3.
Most common now is either 2. … Nf6 or 2. … Nc6.
The gambit 3. . … d5!?? seems suspicious. Black does not get compensation for the pawn. And keeping in mind that one might play against this opening in one out of ten thousands games suggests that black should take a more practical approach to the position. Apply the KISS principle here.
Since the f3 square is occupied by the queen, the move 2. … c6!? suggests itself. I’m suspicious of black playing to open the center when none of his pieces are developed. However, whether black is even able to get in the d5 advance is doubtful since white can put pressure on e5.
This is a classic example that comes up in chess openings all the time. Black can play to create an open and dynamic position in response to some inferior move by white. Or black can opt for a more positional option that should at least equalize without creating unneeded complexity.
From a general perspective, you need to think in terms of meta-strategic principles which are the principles by which you play not within a single game, but across many hundreds or thousands of games. You’re trying to optimize your average win rate and you’re trying to optimize your use of energy both for chess study and over-the-board play. Perhaps you accept a lower average win rate in positions such as this, but you maximize your average win rate across many games. That is preferable to spending so much effort trying to find the most critical move in a given game or spending too much time studying esoteric lines that are unlikely to occur in practice.
From that perspective, I prefer 3. … Nc6 and then we have 4. Bc4 Nf6 5. c6 Na5! or 4. Bc4 Nf6 5. Ne2 Na5! as what I would view as the main lines. Both give black very good play.