Marshall - The Endgame Master

If one was to think of an American Chess Grandmaster who dominated during his time, then Bobby Fischer would naturally come to mind for many. There is no denial in what he brought to the game of chess; this is especially true for players from the United States of America. However, Fischer was not the first dominant American chess player. Frank Marshall was born, in New York City, in August of 1877; although from age 8 to 19, he lived in Canada.

Marshall learned how to play chess at age 10 and by 1890 (age 13) he was already one of the best chess players in Montreal, Canada. Marshall is usually known for his attacking style and tactical play. In fact, the Marshall Attack (1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 O-O 8. c3 d5) in the Ruy Lopez Opening is highly respected even today. However, lesser known today (yet well-known by his contemporaries) is that Marshall was also a master of the endgame. When I think of great endgame players, Capablanca is typically the first player to come to mind; yet, Marshall was a great endgame player as well.

In 1909, Marshall agreed to play a match against the young Cuban chess player Capablanca. Of the 23 chess games played: Marshall drew 14 games, lost 8 and only won just a single game. Rather than be resentful, Marshall was just the opposite. He realized Capablanca’s chess talent and thought he deserved recognition. The American player Marshall used a lot of his influence to make sure Capablanca had a chance to play at the highest levels; he even insisted that Capablanca should be allowed to play in the exclusive San Sebastian tournament of 1911. Capablanca did enter and ended up winning the entire event! Just a few years later, in 1915, Marshall opened a chess club in New York City. The Marshall Chess Club is now one of the oldest chess clubs in the United States. It currently operates as a non-profit organization and gold affiliate of the United States Chess Federation (USCF).

Franks Marshall’s chess skill was unwavering though. He dominated the American chess world during his peak years - much like Fischer. The U.S. Chess Championship was won by Marshall every year from 1909 to 1936 and his winning streak only ended because he declined to play in the invitational tournament that began in 1936. Samuel Reshevsky won in 1936 and then went on to win through 1942. There wasn’t a player this dominant until a young Bobby Fischer won the 1957/58 tournament at age 14, becoming the youngest United States Chess Champion ever. Fischer then won the event every single year through 1966/67 except for the 1961/1962 tournament won by Larry Evans and that was only because Fischer did not play that year; this gives Fischer a record of eight championship wins out of eight attempts!

In 1913, just four years after Marshall lost his match to Capablanca, they had another encounter: this time Marshall was able to outplay Capablanca in the endgame and score a win as the Black pieces. Marshall was one of the few chess players to ever beat Capablanca when Capablanca had the White pieces and also worth noting is that Capablanca rarely lost an endgame. This is the game we will investigate here.

This chess game was played in Havana, 1913. Capablanca had the White pieces and Marshall had the Black pieces.

1. e4 e5 2. Nf3 Nf6 Marshall has the Black pieces, but no Marshall Attack today: unless play transposes into the Ruy Lopez Opening. Here, we instead have the Petrov’s Defense.

3. Nxe5 d6 4. Nf3 Nxe4 5. d4 d5 6. Bd3 This is all opening theory and is the Classical Attack of the Petrov’s Defense.

Bg4?! This move is questionable; the modern interpretation is probably 6…Nc6, but another possible improvement may be the book line 6…Be7 7. O-O Nc6 8. c4 which is known as the Jaenisch Variation of the Petrov’s Defense: Classical Attack. The Bishop move to g4 doesn’t score too well for Black since White has many choices to play for an advantage.

7. O-O Nc6 Marshall is still allowed to get …Nc6 in anyhow it appears.

8. c3 Be7 9. Nbd2 Nxd2 Capturing on d2 is a fine move, but an aggressive alternative may have been the interesting idea of 9…f5!? A sample line may continue: 9…f5!? 10. Qb3 Rb8 11. Ne1 Nxd2 12. Bxd2 O-O and the game is still in the air, but the position is energetic for sure. The f-pawn advancing naturally comes with drawbacks such as the more open King (currently x-rayed from the b3 Queen), but also comes with some interesting positive points as well. One point may be that …f4 may come into play soon. White will almost always kick the g4 Bishop back with f3 and then …f4 at some time later - if allowed, …f3 might even come into play! Perhaps something like 13. a3 Qd7 14. Nc2 f4 15. Nb4 Nxb4 16. axb4 f3! when keeping the Kingside “closed off” with 17. g3? backfires to the calm 17…Bf5 and the f3 pawn is menacing with light-square weaknesses all around the White King. Naturally, this is just one long sample line continuation, but the concept to note is that the move 9…f5!? is interestingly playable despite the drawbacks it comes with. Even if this isn’t all meticulously calculated, it is fairly clear to see the potential 9…f5!? brings once this move is considered.

10. Bxd2 O-O Back to the game text, we see both sides continue developing pieces smoothly and Black castles to safety.

11. h3 Bh5 It is worth noting that 11…Be6?! is the kind of dubious move players at the lower to intermediate levels of chess routinely make. It looks tempting to retreat this way because it looks to support the center d5 pawn, but it is much stronger to retreat to the h5 square as Marshall has done. It is about keeping the tension in the position. By maintaining the pin on the f3 Knight, Black’s Bishop is more active; after 11…Bh5 12. g4 Bg6 we actually still have a fairly even game in this exact position, but this is the exception rather than the rule. Usually advancing the pawns in front of your castled King like this (12. g4) is far too ambitious. From a positional standpoint, moves like 12. g4 are usually ugly and Black should provoke these moves from the opponent whenever possible. In this unique case, 12. g4 is surprisingly all right because of the combination of Bishop trades on g6 giving Black doubled pawns and the fact that Black’s pieces are not ready to take advantage of White’s opened Kingside. White’s King is paradoxically safer than it appears even though pushing the pawn shelter is something that can’t be un-done (as pawns can’t move back, unlike pieces can).

12. Re1 Qd7 Both sides instructively continue development. White activates their Rook by placing it on the potentially valuable e-file and Black moves the Queen off the backrank to “connect” the Rooks. Maybe …Qd6 was slightly better because it takes away the f4 square from White’s Bishop (which is currently doing little on d2). Of course, one point with the Queen on d7 might be indirectly keeping the d1 White Queen on the defensive. If the White Queen were to leave defending the f3 Knight then Black can chop and get to the castled King. For instance, 13. Qb3?? Bxf3 14. gxf3 Qxh3.

13. Bb5 Bd6 Both sides overlook something that may have been game changing, so the moves are kind of a wash. 13. Ne5! was an excellent move that is tricky to find over-the-board (OTB).

Obviously, if 13. Ne5! Bxd1? 14. Nxd7 keeps White ahead with play maybe continuing 14…Rfd8 15. Ne5 Nxe5 16. Rxe5 Bd6 17. Rxd1 but the critical line to calculate comes after 13. Ne5! is met with 13…Nxe5 where White’s winning move is 14. Bxh7+! and 14…Kh8 or 14…Kxh7 are both met with White’s Queen capturing on the h5 square.

14. Ne5! Now Capablanca finds this move. It is not quite as powerful as it would have been a move earlier, but it is still strong enough to grab the immediate advantage and initiative. Back on move 12, the idea of a White Knight hopping into e5 may have been another small reason to consider …Qd6 over …Qd7, but the difference is only a small improvement.

Bxe5 15. Qxh5 Bf6 White has the slight advantage and the Bishop pair, but nothing seems concrete yet - only positional considerations. One creative way for White to push for a greater advantage may be to utilize their Bishop pair by relocating the Bishop on b5. It isn’t doing a great deal on this side of the board and chopping on c6 would give up the Bishop pair. Ideally, the Bishop should be helping coordinate something with the Queen against Black’s King; after all, White’s Queenside pawns are “pointing” at Black’s Kingside indicating the White should generally be playing on this side of the board.

Creatively putting this plan into action may take the form of Ba4-Bc2 such as this sample line: 16. Ba4 Rae8 17. Qf3 Qd6 18. Bf4 Qd8 19. Bc2! with a “mission accomplished.” Black would then probably have to address the Kingside pressure by bringing more pieces over to defend with a follow-up like 19…Ne7, but White’s advantage is still present and more substantial than the theoretical ideas from the drawing board after move 15…Bf6. Continuing the alternative line from here we left off, after 19…Ne7 20. Qg3 looks instantly attractive to keep the Kingside pressure coming for White and Black is losing.

16. Bf4 Rae8 White decides to improve the passive Bishop on d2 first; this is a solid plan as well. Black uses this chance to contest the e-file, but if any side is winning: it is still White.

17. Re3?! Rxe3 Placing the Rook on e3 threatens doubling Rooks on the e-file, but this is too ambitious of White. After the obvious Rook trades on e3 (Black can’t afford to allow White to double Rooks on an important file), White has lost the initiative. It reminds me of a quotation from the first official Chess World Champion Wilhelm Steinitz, “Only the player with the initiative has the right to attack.”

As the text game illustrates, White lost the initiative of their potential Kingside play after 17. Re3?! while an interesting alternative may be found in 17. Qg4!? Offering a Queen exchange on g4 looks funny because material is equal and getting Queens off the board usually favors the defender; plus, White would get doubled pawns. However, White’s initiative doesn’t fade and White still has the play when Queens come off the chess board in this exact position. Perhaps 17. Qg4!? Qxg4 18. hxg4 Rxe1+ 19. Rxe1 Bd8 and Black’s Bishop is forced into a passive defensive role. Meanwhile, White has the Bishop pair that may become really powerful in the endgame. I have little doubt that Capablanca wouldn’t go for this position if he evaluated the power of trading Queens; we must remember that Capablanca has several endgame masterpiece games and seldom ever lost an endgame. Even an elite chess player like Marshall would have trouble holding the endgame against Capablanca after 19…Bd8. I know I wouldn’t want the Black pieces against Capablanca with this position.

18. fxe3 a6 I can easily see Marshall’s perspective with attacking the b5 Bishop. Black wants to free up the Queen from the pin, since a Queen move now might give White the chance to give Black doubled c-pawns. The problem is I think White would never trade off the Bishop when it is attacked here. White will just maintain the tension of the pin by playing Ba4 and I think Black is just pushing the Bishop to a better square.

19. Ba4 b5 Noted and virtually forced if Black elects this line of play.

20. Bc2 g6 21. Qf3 Bg7 With Bc2 aiming at Black’s Kingside, 20…g6 was practically forced, so Black uses this to fianchetto the Bishop from the unconventional route of f6-g7. When the pawn was pushed to g6, Black gained dark-square weaknesses around the King and placing the Bishop on g7 is an attempt to plug those holes and keep the White forces from raining into Black’s King shelter.

22. Bb3 Ne7 Bb3 is almost an exclaim if it wasn’t so obvious: we can see why Capablanca is such a strong endgame player. White has a slight advantage and Capablanca is constantly provoking weaknesses and attacking them. Here, Bb3 adds more pressure to the Black d5 pawn and …Ne7 is the logical way to defend this. This may sound like a small detail, but White is playing instructively because it isn’t about attacking a little pawn: it is about piece activity and keeping the opponent pieces passively on the defensive. Threatening the d5 pawn was just the means to provoke a Black piece (Knight now on e7) into a defensive role. It is a kind of endgame motif that is like zugzwang really far in advance; when Black reaches the point where they have everything forced into passive defensive duties, then the position is usually close to collapsing. It is as strategist and Dutch Grandmaster John van der Wiel once said: “When you absolutely don’t know what to do anymore, it is time to panic.”

23. e4 dxe4 24. Qxe4 c6 25. Re1 Nd5 The endgame is playing itself and White is slightly better.

26. Bxd5 cxd5 Marshall probably plays the pawn recapture instead of the Queen because he too realizes that White is slightly better in this endgame; by keeping Queens on the board, Black can hope to gain counterplay out of the complexity. Any great simplification will just give White a simple theoretical endgame to convert.

27. Qe7 Qc8 Again, we see both players acutely aware of the chess position being played out. Capablanca wants to simplify by trading Queens off the chess board and Marshall wants to avoid this to give Black more chances for counterplay and complexity.

28. Bd6 h6?! Perhaps Black could have tried 28…Rd8 instead, but Marshall is understandably trying to create a luft for the King by anticipating an incoming attack soon.

29. Rf1 f6 30. Re1 Rd8 We again see Capablanca’s endgame mastery at work reminiscent to the concept on move 22. Bb3 to provoke …Ne7 into the defense. In this case, White has play on Black’s f-pawn and after it is forced into advancing, then the White Rook returns to the e-file. With Black’s f-pawn advanced slightly, Black’s King is now more exposed and White can infiltrate on the 7th rank if given the chance.

31. Bc5 Kh7 Black utilizes the luft created on move 28. I don’t think playing the h-pawn up to the h6 square was best, but it is hard to defend a losing position. As Marshall himself once said, “A bad plan is better than none at all.”

32. Qf7 Qf5 33. Be7? This looks like a clever continuation because the Bishop attacks the f6 square and gains time attacking the d8 Rook. A Rook move like 33…Rc8 fails to 34. Rf1 and White will win material since the f6 pawn would be attacked 3 times with only 2 defenders. What makes Bishop to e7 a mistake is that it overlooks the possibility that Black may not need to move the Rook to defend the f6 square.

Qd7! Now Black has endgame chances! Pinning the e7 Bishop to the Queen like this is irksome for White. Better than 33. Be7? would have been for White to play 33. Re7 Rg8 (to defend checkmate with Qxg7#) 34. Re6! and only now does White gain significant pressure on the f6 pawn without any Black counterplay.

34. Kf1 Rf8 35. Qe6 Qxe6 36. Rxe6 Re8 It is amazing how Black now has a fairly even endgame to play out.

37. Re2 Kf8 38. b3 Kf7 39. Bc5 Rxe2 40. Kxe2 f5 The endgame seems to play out naturally with both sides getting their remaining pieces active - including Kings. The recent move 40…f5 has the intension of getting more pawns on light-squares to help their dark-squared Bishop remain active; the f5 square may also become a nice location for the Bishop to help keep White’s Bishop off of the d8-h4 diagonal.

41. Kd3 Ke6 42. c4 bxc4+ 43. bxc4 g5 The endgame is still up for grabs, but the best move for White isn’t extremely intuitive at first glance. 44. g4 is actually White’s strongest move here, since White needs to stop the Black pawns from advancing too far. If the Black pawns can’t advance too deeply into enemy territory, then the White monarch can always chase down the straggler pawns to defensively prevent pawn promotion.

An example of a mistake in this endgame would be the normal-looking move 44. cxd5+? because 44…Kxd5 gets the King opposition and Black’s Kingside pawns may prove troublesome theoretically. In practical play, the Bishops still keep the game roughly even (as each side can avoid zugzwang by shuffling Bishops to pass tempi), but Black has a 3 versus 2 pawn majority on the Kingside. This is a stronger advantage than White’s Queenside potential because of the better pawn structure and space advantage.

Pragmatically, a sample line may go 44. cxd5+? Kxd5 45. Ke3 Be5! and the Black Bishop is immune to capture because 46. dxe5? Kxc5 gives Black a winning pawn ending.

44. g4 f4 Of course, Capablanca sees this and finds the best move 44. g4 to give Black stubborn resistance.

45. Bb4 Bf6 46. Bf8 dxc4+ 47. Kxc4 f3! Of course, as mentioned earlier, Marshall was known by his peers as an exceptional endgame player as well. Here 47…f3! is the only move that gives Black a slight advantage! The passed pawn is threatening to make a break at promotion on f1 and this forces the decoyed White King to give up controlling the d5 square that Black may use for their own King with …Kxd4 being only a matter of time.

48. d5+ Ke5 The idea behind the d5 check is simply to take away the option of Black putting their King on d5, but then Black has …Ke5 instead.

49. Kd3 Kf4 The Black King has infiltrated and now is going after the weak h3 pawn or threatening to “build a bridge” with the Black King on g2 to escort the f-pawn to the f1 square.

It is worth pointing out that deflecting the White King too far with the f-pawn is not in time. 49…f2?? would be a blunder that gives up the game. 50. Ke2 f1=Q+ 51. Kxf1 Kf4 or 51. Kxd5 and White is the one playing for a win.

50. Bd6+ Be5 51. Bc5 Kg3 52. Ke4 Bf4 Black is surely winning!

53. d6 f2 0-1 White resigns because the pawn race to promotion is being won by Black and sacrificing the White Bishop for the dangerous pawn only loses slower due to Black having time to catch the White passed pawn on d6. For instance: 54. Bxf2+ Kxf2 55. d7 Bc7 and the Black Bishop is just in time to halt White from a safe pawn promotion on d8.

Capablanca’s endgame mastery is well-recognized today, but we should not forget that Marshall was an endgame master too. This is often overshadowed by many games illustrating Marshall’s attacking talent, but he was more than capable of converting an ending against the best of them and Capablanca is certainly one of the best. What makes this a good game for demonstrating Marshall’s endgame prowess is that after the 33. Be7? mistake, Black only has an equal ending versus Capablanca. As the game text reveals, Marshall was also a fantastic endgame player and we should not forget this as many chess players have forgotten over the years.