I tell folks that if they want to risk a few incorrect moves early, it’s necessary to be more alert to tactics than your opponents, and more proficient in endgames (“playing bad chess well”).
There are some gambit games in which the pawn sacrificed in the opening is never recovered, and the gambit player goes in the endgame a pawn down, but with some extra time. That’s the most critical time for gambit players, because that’s when you have to demonstrate to skeptics that sacrificing a pawn will not really, seriously, totally cost you in the ending.
In fact, the gambit player should thrive in games where material and time are in the same imbalance that arose in the opening. Endgames are a race to queen a pawn, and if the gambit player can cultivate that lead in time to be the first player to promote, then that the material value of that sacrificed pawn never came to mean anything.
When Fischer and Polgar played the King’s Gambit, they preferred the more adventurous 3. Bc4 to 3. Nf3. I share two games in which these great masters win endgames following a Bishop’s Gambit, which could fit together in a book chapter as two examples of one theme, but the conditions in which they are played could not be more different.
J. Polgar-Barua, Biel 1993, was played in an interzonal, the tuornaments that qualify players for the candidates’ matches, and the winner of those candidates’ matches earns a shot at the world championship. Interzonal games are very big deals, so winning a game at an interzonal with a Bishop’s Gambit to being, and a rook-for-knight sacrifice at the end, that’s extraordinary.
Fischer-Cunningham, Houston 1964 was a game from Fischer’s’ barnstorming tour of the United States, comprised of 40 simultaneous exhibitions and more than 2000 games. Really, the tour in itself was an outstanding feat, chronicled in IM Donaldson’s A Legend on the Road: Bobby Fischer’s 1964 Simul Tour.