Bobby Fischer's
My 60 Memorable Games (MSMG) includes three losses, including a wild one to Efim Geller from Skopje 1967. Fischer achieved an objectively won position on the White side of a Sozin Sicilian, but in the complications lost his way and then the game. Here's the key position, after Geller's 19th move:
Fischer played 20.a3?, and after Qb7 21.Qf4(?) Ba4!! 22.Qg4 Bf6! 23.Rxf6 Bxb3! was forced to resign, as his intended 24.Rf4 gets mated by 24...Ba2+ and 25...Qxb2 mate, while 24.cxb3 Nxf6 leaves White down a rook without real compensation. In MSMG, Fischer claims that 20.Qf4 ("!!"), with the threat of Rh5, is a winner. He analyzes 20...d5, 20...Nd2+ and 20...cxb2, and some brilliant lines culminate in White wins.
In Edmar Mednis' first chess book, the 1974 work
How to Beat Bobby Fischer, this conclusion is disputed. After 20.Qf4 cxb2 21.Rh5 (threatening Bxg7+), Mednis suggests 21...Nf6 ("!!") instead of 21...Nc3+ and 21...Bf6 as given in MSMG. Mednis writes: "I see no winning opportunities for White", suggesting only a possible draw with 22.Rxh7+, but without supporting analysis.
Going forward a few years from Mednis' book, but back 27 from the present day, to mid-1981, my friend Jeff Gallegos and I tried to defend Fischer's honor by busting the 21...Nf6 line. Jeff and I were just shy of 1900 at the time; not world-beaters, but not bad for 17 and 14 years old, respectively. Of course we couldn't analyze at the level of a Fischer or even a Mednis, but with persistence and the cockiness of youth we didn't see any reason why we couldn't get to the bottom of things.
It was generally Jeff who would initiate the attacking tries for White, and then I'd do my best to bust them. Mednis' claim stood for a while, and I was starting to think he was right, but Jeff finally found the right move. We were going to send our finished analysis to Chess Life (to Larry Evans' Q&A column), but in a sad coincidence the issue that came out even as I was writing the letter saw then-IM (now GM) Nick De Firmian write the column with the same solution and main idea. Alas!
Yet as sad as it was to be scooped, I now see it as a mixed blessing. Looking back on our analysis, it turns out that it wasn't very good. We had the right move and the right idea, but on the details we failed to find the best defenses and had plenty of errors in the lines we did cover. It's also nice from a nostalgic perspective to have the actual letter we were going to submit.
I've deliberately avoided saying what that move and idea happens to be, so that all of you can try to find it for yourself. Before digging up the old letter earlier today, I decided to have a look for myself to see if I could figure it out, and it turned out to be pretty easy to find. (Of course, it's possible that at some level I remembered the solution, so maybe I'm not as much stronger than my 14-year-old self as I'd like to think.) Taking a look at what the contemporary computers think was very interesting, though. The three of us were right, but Black can put up a tremendous amount of resistance. (I just noticed that Kasparov also presents the line in
My Great Predecessors, vol. 2, game 100, in his section on Geller. Yet his analysis too leaves out a line that took me quite a while to work through using Rybka 3 and Deep Fritz 11 on a computer at least five years newer than the one he used at the time.)
So see what you can come up with, and feel free to post analysis to your heart's content. (Readers be forewarned: don't read the comments if you want to look at the position without any preconceptions!) I'll post my analysis sometime later this week, and will incorporate readers' suggestions if they add to or overturn any of my own conclusions.
