Problems at: <21...?>, <22...?>, big hash needed.
J Arnason vs Nunn, 1990
(C84) Ruy Lopez, Closed, 26 moves, 1-0
Rybka gets confused in the Rc2 line (see RV's post on page 1).
Yusupov vs J Nogueiras, 1985
(D43) Queen's Gambit Declined Semi-Slav, 20 moves, 1-0
Decision between Bd3 and Ba4 is not at all clear to engines.
Kuzmin vs A Choudhary, 1984
(C77) Ruy Lopez, 33 moves, 1-0
Best lines for black's 33rd move need very deep analysis.
Janowski vs Schlechter, 1899
(C78) Ruy Lopez, 36 moves, 1-0
See the zenpharaohs response to patzer2 re: 22th move for black
P Popovic vs M Lazic, 1993
(C92) Ruy Lopez, Closed, 29 moves, 1-0
Takes Hiarcs 17 plies and 50 CPU minutes to figure out 14.Rd8
Anand vs Robatsch, 1992
(C10) French, 20 moves, 1-0
It seems that Rybka 3 has some serious problems with this game.
Vescovi vs R Szuhanek, 1994
(B53) Sicilian, 27 moves, 1-0
Hiarcs does not find 19.Bxh6 even after 30m CPU, 17 plies.
Serper vs Barsov, 1988
(C05) French, Tarrasch, 25 moves, 1-0
Long time to decide between 19.Nd4/19.Rd6
Spassky vs Larsen, 1978
(B18) Caro-Kann, Classical, 21 moves, 1-0
17.Bf6 is very difficult to find for engines.
Spassky vs V S Korensky, 1973
(B65) Sicilian, Richter-Rauzer Attack, 7...Be7 Defense, 9...Nxd4, 28 moves, 1-0
All strong engines fail to find 44. Rxf6. But Crafty does !??!
F Olafsson vs Unzicker, 1970
(D41) Queen's Gambit Declined, Semi-Tarrasch, 47 moves, 1-0