Exploration and Discovery

The other day Yury Bukayev sent me a game played by World Champion Anatoly Karpov in a simultaneous display which sent me off on a great quest of exploration and discovery:

Karpov - Delgado, Simul, Terrassa, Spain, 1976: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 b5 5.Bb3 Na5 6.Bxf7+ Kxf7 7.Nxe5+ Ke7 8.d4 Nf6 9.O-O Qe8 10.Nc3 Bb7 11.Re1 Kd8 12.d5 Qh5 

I will stop the game here, for now, except to point out that Karpov lost in 58 moves, the result of an atypical endgame error - perhaps due to tiring in the later stages of the exhibition.

Backing up, I would like to look at the opening moves: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 b5 5.Bb3 Na5

As Andrew Geet writes in Play the Ruy Lopez (2006)

You are looking at the Norwegian Variation,  which acquired its name having been used extensively by a number of masters hailing from the Scandinavian country.

Jonathan Tisdall explains further, in his article in the New In Chess Yearbook, #37, 1995

Since the earliest explorers of this unusual line were Furman and Taimanov, the variation has appeared under various names, but has eventually grown attached to Norway after the persistent an imaginative efforts of the leading Norwegian players of the 1960's-70's, IMs Johannessen, Arne Zwaig and Terje Wibe, as well as other members of the national team such as Ragnar Hoen and Ole Christian Moen. The current generation of Simen Agdestein and Einar Gausel have carried on this tradition...

I would like to back up even further, however, over 100 years.

In the February 1901 issue of the Deutsche Schachzeitung, Carl Schlechter wrote of the line in "A New Defense to the Spanish Game [Ruy Lopez]". He looked primarily at 6.Nxe5, assessing equal chances, and 6.d4!, where Black looks very good.

This was noted in The British Chess Magazine in its May 1901 issue, in a brief article by C.E. Rankin, and in an adjoining article where W. Timbrell Pierce proposed 6.Bxf7+

Mr. Rankin's interesting article on a New Defence to the Ruy Lopez reminds me of some games played about ten years back at Harrogate, and again in 1896-7 with Mr. Sutcliffe, who used to play that defence. I remember I thought at the time and still think White may venture to sacrifice his KB for two Pawns and a strong attack...

Pierce's work was quickly noticed in Deutsche Schachzeitung, and this was followed by another article by Schlechter on 6.Bxf7+. Not surprisingly, the attack was found to be incorrect, and two lines of analysis were given, one leading to a win for Black and one leading to Black's advantage.

So, we have an obscure attack - 6.Bxf7+ - against an infrequently played defense, whose origins, according to Pierce, go back to the 1880s or 1890s. I have not been able to track down any game examples, yet. 

Interestingly enough, the line has picked up the name the Nightingale Variation, possibly due to a correspondence game covered in the BCCA Magazine: 

Nightingale, Cyril Alfred - Ellinger, Maurice, correspondence, 1950: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 b5 5.Bb3 Na5 6.Bxf7+ Kxf7 7.Nxe5+ Ke7 8.d4 d6 9.Bg5+ Nf6 10.Nc3 dxe5 11.Nd5+ Kf7 12.dxe5 h6 13.Bh4 g5 14.exf6 gxh4 15.Qh5+ Ke6 16.O-O-O Bd6 17.Qg4+ Kf7 18.Qg7+ Ke6 19.f4 Bb7 20.f5+ Ke5 21.f7+ Kxe4 22.Rd4+ Kxf5 23.Qg4+Black resigned

As Tisdall also notes, the line is also

Known as the Swedish Variation due to some serious testing in Scandanavia, but I believe that the Latvian Vitolinsh was in fact the move's most devoted adherent.

It is characterized on the blog "Sverre's Chess Corner" as the "Acid Test" of the Norwegian Variation

The positional ramifications of Johannessen's 7...Nxb3 and Zwaig's 7...f6 - or for that matter the Stein/Wibe variation with 7...exd4 - obviously are only relevant if Black can survive this more direct attacking attempt. 

All of this brings us - almost - to the play of this line by a world champion in a simultaneous exhibition of 19 boards. I say "almost" because it took place a half-dozen years before the player of the White pieces became world champion.

Jose Raul Capablanca - Percy H Moise, New Orleans Chess & Checkers Club, New Orleans, simul, 1915: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 b5 5.Bb3 Na5 6.Bxf7+ Kxf7 7.Nxe5+ Ke7 8.d4 d6 9.Bg5+ Nf6 10.Nc3 Bb7 11.Nd5+ Bxd5 12.exd5 Ke8 13.O-O dxe5 14.dxe5 Qxd5 15.exf6 Qxd1 16.Raxd1 gxf6 17.Bxf6 Rg8 18.Rfe1+ Kf7 19.Bc3 b4 20.Rd7+ Kg6 21.Re6+ Kf5 22.Rf6+ Kg5 23.Rd5+ Kh4 24.Rf4+ Rg4 25.Bf6 checkmate

The New York Tribune, of  April 18, 1915 covered the game, and pointed out

Here is where a master's knowledge and experience counts. Capablanca, when handing the score to the Tribune, said that he never before played or saw such a variation, but he figured pretty promptly that two pawns with the position in hand and the promising attack would be ample compensation for his bishop, especially so against and inferior opponent.

Rashid Nezhmetdinov would go on to play the sacrifice against Seman Furman in 1954 (1/2-1/2, 34), the same year Boris Spassky played it against Mark Taimanov (1/2-1/2, 49).

Eventually, Anatoly Karpov would give up his bishop against Delgado in 1976.

I am still searching for answers, starting with the question "Didn't Karpov's 12th move simply drop a piece?" and following with "Why didn't Delgado play 12...Qxe5 ?" My suspicion is that Delgado didn't believe that his opponent would make such a mistake, and so declined the sacrifice, only to take advantage of another error, about 30 moves later...