Sometimes activity is FAR more important than material. Another similar opening in terms of one side sacrificing a lot of material for activity is the Traxler Counterattack.
Double/Triple Muzio Gambit

Stockfish overrates the moves. It changes the evaluation every move, even if they are the best moves.

8.Bxf7? Kxf7 9.d4 Qf5! was suggested as a refutation by WC Wilhelm Steinitz, some 140 years ago. Steinitz was right, and practical play/ engine analysis confirm that Black is rather easily winning in this line.
Taking the d4 pawn requires precise defence, and Black need not go for that in exchange of a measly pawn- he already has enough material in pocket.
9...Qf5 of Steinitz was strong indeed, as #8 points out, but 9...Qxd4+ was still winning for black.
11...Bc5+? spoils the win. 11...Ne7 was still winning: black has adequate defence and is ahead in material.
12...Bd6? even loses. A developing move like 12...Nc6 was still equal. Black will need to return all of his material to avoid checkmate, but can hold a draw.
"why is the win rate so high" ++ Because defending accurately is hard. Not everybody is a Steinitz, Capablanca, Karjakin, or Carlsen. In correspondence the Muzio is a sure loss for white, but in over the board play with a clock black players have a hard time defending it, especially if they have never met it before.

9...Qf5 of Steinitz was strong indeed, as #8 points out, but 9...Qxd4+ was still winning for black.
I guess that an engine with pawngrabbing addiction may well consider Qxd4+ as the best move. However, it does require precise defence for quite awhile, so it may even be decorated with a "dubious" mark.
9...Qf5 is quite safe, once you stick to the policy of returning material to neutralize white's attack. In the main line, Black even gives back a whole rook to get an easily winning ending.

[quote]Sometimes activity is FAR more important than material. Another similar opening in terms of one side sacrificing a lot of material for activity is the Traxler Counterattack.[/quote]
it's easier to attack than defend. i'm a FIRM believer in initiative as a tactician and have always disagreed with all of the anti-gambit GM babble when smith morra is a MAJOR spoiler (i took soooo many points with it!), there's only ONE variation of the (down a rook in at least one line) caro kann caveman that ISN'T losing by move #15 in stockfish vs everything i had analysis of dozens of games testing the theory, and the absolute STRONGEST chess player on the planet, leela chess zero LOVES throwing material away for the initiative... well, I still say gambit hate is BS!
as a double muzio specialist (took a +400 fisher player out with it because he missed Qxg5 after a Bxf6+ "krejcik" to unpin my knight), i've come to the conclusion that GM shaw is CORRECT to espouse the quade gambit! in looking at 1600-2000 games, the quade performs 54:42 at 11k games versus 49:48 in 66k games. 12 points difference is no laughing matter, and in digging deeper into it, white has the advantage in most lines by move #15!
i'm thinking that studying amateurs (people get violently upset over it! HAHAHA!) might just be the TRUE measure of a line's "inherent soundness"... people keep testing, and the line has better resources to defend. amateur games solve "unclarity" through trial and error testing EVERYTHING. it's a mind numbing process combing through a few hundred thousand games, one move at a time looking for best performing lines that stockfish agrees with more often than not because one sees all of the homework everyone's done, and how it pans out over the board.
i'm still looking for the hater thread i created to see if big mouth wants to test his mettle against crowdsourced theory that knows EVERY trap etc. once you START at 51:49, the numbers keep going up as long as you stick to best performing, or in extreme cases... evaluated moves when the crowd's out of ideas.
i've started gathering quade theory because THAT's where the wins are.
I'm very interested in why Double Muzio Gambit is such a good opening. It has a 70% win rate for white among masters, despite the fact that white is literally down six points of material! Furthermore, the engine also analyses this position as white being down "-5.5". So why is it named, why do masters play it, and why is the win rate so high despite the engine's protesting?
Below: Double Muzio Gambit
Below: One such game played between masters