Kasparov famously wrongly resigned the following position against Deep Blue in game 2 of their second match (which he went on to lose 2.5-3.5) :
PGN Viewer
Today I came across some comments in which fellow psephologist Nate Silver, writing about the match in a book about statistical predictions, had suggested Kasparov resigned because of anxiety about the computer's super-human capacities caused by a strange move in the previous game. Frankly, although that was a secondary comment based on some comments by others, I think Nate should stick to his day job on this one.Kasparov's resignation was something that could have easily happened between two humans in the same position.
But that made me curious about this game again and online coverage of it and I found some fresher analysis to that from the time of the game.
While Kasparov's resignation was obviously premature there has been dispute since as to whether it was even drawn at all. It does not seem, at least, that it was clearly drawn by perpetual check.
These are some lines from various sources mainly here.
The current position is that it is a draw with best play but I wonder if anyone can propose any wins for white that overturn this.
PGN Viewer
Further contributions welcome.