In a Game of SuFi 19 , a Czech Benoni, Stockfish 11 as White could have forced , on move 28 , a line that forced an eight piece endgame on move 59. White would have had Queen , Knight and two advanced pawns vs Queen and Rook . The line was suggested by Stockfish own PV on move 27 , a plan that later abandoned for a small centipawn difference. In the game Stockfish was pressing hard in a pawn up endgame , but cannot overcame the stubborn defense by Leela zero ( at least in that game) and the game ended in a draw. Since I have extensively analyzed the eight piece endgame described above with Houdini 6.02 at d= 38/128 and concluded that it was a win, it should be concluded that Stockfish 11 took the wrong path on move 28. The game was a Czech Benoni defense, and , to be honest , is the only mistake by Stockfish 11 in the current SuFi when Stockfish was not in time trouble. Game 62 SuFi 19. At least even Stockfish 11 is not infallible!
Perhaps worth starting by explaining for the benefit of the majority of chess fans who have no interest in computer chess tournaments that a "game of SuFi 19" is not some new chess variant. "SuFi" stands for "Super Final" and "SuFi 19" is the 2019 computer chess super final tournament.
Is Stockfish 11 SuFi 19 still fallible?
It's also worth reminding ourselves of the meaning of the word "fallible" and its opposite, "infallible".
"Infallible" means "never makes a mistake" and "fallible" means "makes mistakes".
We are still a very long way away from the day that chess will be solved and so are still a long way away from the day that a chess engine is infallible. Until then all chess engines are fallible. In particular Stockfish 11 is fallible.