The unfinished symphony from Cambridge Springs

I felt sorry for my opponent. The guy played an excellent correspondence game, had me running out of ideas where to find an advantage and then just blundered it away, probably by a misclick.

Yes, a misclick 😉 These things happen even in correspondence chess.

It is an incomplete game because of the blunder, but since it is strategically instructive, quite rich and didn’t evolve into a purely tactical mess too early I annotated it nevertheless.


I’ve put all the moves below into a Lichess study for you to replay and analyze more conveniently. The game also represents the newest addition to our database of annotated games.

Advertisement for the “Waterbury chess clock” that was used in the 1904 Cambridge Springs tournament (won by Frank Marshall) after which the opening below was named.

Hero (2.205) – Villain (1.933)
Lechenicher SchachServer 2017
Queen’s Gambit, Cambridge Springs Defense

1. d4 Nf6 2. c4 e6 3. Nf3 d5 4. Nc3 c6 5. Bg5 Nbd7 6. e3

(6. e4?! is premature, it only accelerates Black’s play. After 6… dxe4 7. Nxe4 Bb4+ White needs to misplace or move back one of his knights, and Black is fine.)

6… Qa5


The Cambridge Springs Defense, one of the most mature systems in the d4 d5 complex. It was first played in 1892 by Emanuel Lasker and then discussed intensely between Capablanca and Alekhine in their 1927 WCh match. As a child I was proud to know that 7.Nd2 needs to be played in order to unpin the Nc3 and prevent Black from going …Ne4. These days 7.Nd2 remains the main move, but lately especially the top level guys like to offer a pawn sacrifice via 7. cxd5 and 8. Rc1.

7. cxd5

(7. Bd3? Ne4 8. Qc2?? Nxg5 9. Nxg5 dxc4 10. Bxc4 Qxg5 is a common mistake White must avoid.)

7… Nxd5 8. Rc1


8… Bb4

Theory wizard Tomashevsky’s new concept


(8… Nxc3 9. bxc3 Ba3 (Black cannot afford 9… Qxa2 10. Bd3 The black queen needs way to
long to get back into the action, and White can easily organize a crucial kingside attack, based on his lead in development.) 10. Rc2 b6 11. Be2 Even top guys like Carlsen or Ding used to go 11.Bd3 here, but had a hard time proving anything. The innocent looking 11.Be2 in combination with 13.Rxe2 was an important new concept, introduced by Russian theory wizard Evgeny Tomashevsky in 2015. 11… Ba6 12. O-O Bxe2 13. Rxe2 The main point of 13.Rxe2 instead of the more obvious 13.Qxe2 is to keep guys on the board and have better central control. White anticipates Black going …Rc8 and …c5 and prevents exchanges on the c file. 13… O-O (13… Qxc3? 14. Rc2 White seizes the c file himself and won’t give it back anytime soon.) 14. e4 (diagram) with comfortable play for White in  Tomashevsky,E (2743)-Ipatov,A (2624) Reykjavik 2015 (1-0, 34).)

9. a3 


White insists on giving a pawn in order to shake off the pressure immediately. His central control and perspective to build up a kingside attack should compensate.

9… Bxc3+

(9… Nxc3?! 10. axb4 Qxg5 11. Rxc3 Qe7 12. b5 is pleasant for White, who has a much easier game than his opponent.)

10. bxc3 Qxa3 11. Qd2 h6 12. Bh4 O-O 13. e4 Ne7


The knight heads towards g6 in order to support the central counterstrike …e5. Black can’t just sit and do nothing, he will be bulldozed then.

14. Bd3 Ng6 15. Bg3 e5 16. O-O


Now that development is complete, the question is how to approach this. White must choose between a bunch of possible setups, all of them reasonable:

  • The most obvious attempt would be to go h2-h4-h5, chase the Ng6 away and then organize some kingside play.
  • It may also be an idea to transfer the Bd3 on the a2-g8 diagonal (blocks the a passer, pressures f7) and harass the black king from that side.
  • Instead of going for the kingside White can even try to press Black’s queenside on the open a and b file.

16… Qe7

Prepares the following march of Anton, the a passer, towards a1.

(16… Re8 In order to understand possible White plans that don’t involve h2-h4 I looked, among others, at a game played by Kramnik more than 10 years ago – and decided that this is way above my paygrade. I don’t understand it. Kramnik seems to aimlessly shuffle his guys around, suddenly he is winning. After spending some time with this game I decided I better go for the mere mortal’s approach, push my h pawn and see what happens. 17. Rfe1 Qa5 18. Qb2 Qd8 19. Bb1 a5 20. Rcd1 a4 21. Ba2 Qe7 22. Qc1 Ra5 23. Qd2 exd4 24. Nxd4 Qc5 25. Bc7 Ra8 26. Bxf7+ Kxf7 27. Qa2+ Kf8 28. Ne6+ and White was winning in Kramnik,V (2729)-Bruzon Batista,L (2652), Turin 2006 (1-0, 32).)

17. h4 a5!N


A novelty, and I like it a lot. 17…exd4 had been played before in correspondence and OTB chess, but 17…a5 is more consistent with 16…Qe7. Black is ready to give back a pawn (on f4) in order to win time for his a passer to march down the board and create counterplay.

18. h5 Nf4 19. Bxf4 exf4 20. e5!


Important inbetween move to interfere with Black’s development. Before capturing on f4, White prevents the Nd7 from reaching f6. Also threats against the black king on the b1-h7 diagonal become an option. Thanks to the h5 advance Black will never have …g6 in order to close b1-h7.


Now Black needs the f8 square to get the Nd7 into the game and free the Bc8. Note that Black must not try to transfer his knight to d5 via b6 since he must keep h7 covered. Once he goes …Nb6 White sets up a Bc2/Qd3 battery, and immediately the black king is under heavy fire.

21. Qxf4 a4

White needs to play f2-f4-f5 in order to make progress. But there’s a traffic jam on the f file, caused by Black’s pawn sacrifice on f4.

22. Qe3

The only way for White to make progress is the pawn advance f2-f4-f5. This illustrates another upside of Black’s pawn sacrifice on f4: it created a traffic jam on the f file. White will need a bunch of tempi to free his f pawn and get it rolling, time that Black can use to find a good setup that supports the further advance of his a pawn.

(22. Nh4? Qg5 and Black takes over.)

22… a3 23. Ra1


Another tempo not used to make progress on the kingside, but White needs at least one guy to make head against Black’s queenside advance.

23… Nf8 24. Ne1


The decision was between 24.Nd2 and 24.Ne1. The former looked much more desirable to me, especially because of the option Ne4 and possible jumps to d6 or f6. But during lots and lots of analysis I couldnt find a way towards a White advantage after 24.Nd2 c5 25.Ne4 cxd4 26.cxd4 (note that Black has two connected passers now). So in the end I settled for 24.Ne1, a move that supports the kingside and central play less, but the option Nc2 stabilizes the centre and may enable White to switch plans and go for the intrusive guy on a3.

24.Ne1 is more a defensive than an aggressive move, but while it doesn’t help Whte’s attack much it also limits Black’s options. However, at this point I was quite worried. Black’s play so far had been impeccable, and if he would be able to keep this up, I would have to concede a draw to an opponent rated almost 300 points lower than me.

24… Ne6?!

An inaccuracy, finally. Black should keep the knight on f8 and install his bishop on the e6-a2 diagonal instead. The moment White advances f4-f5 the knight will go …Nf8-h7 find a beautiful eternal square on g5, one downside of White’s h2-h4-h5 advance.

(Of course after for instance 24… Be6 25. f4 f6 26. f5 Bb3 the game just starts, things are doubleedged and complicated, but it seems to me that the objective assessment of this position is “unclear”.)

If only White could go 25.f4 with his queen being on e2. But wait, he actually can!

25. Qe4!

(After 25. f4 Nc7 …Nd5 comes with tempo, and Black has sufficient counterplay.)

25… Ng5 26. Qe2

With f2-f4 coming Black has nothing better that to return his knight.

26… Ne6


Now the position at move 25 is reached, only with the white queen on the better square e2. I would love to take credit for this sophisticated maneuver, but 25.Qe4/26.Qe2 is simply the engine’s first choice. As much as I like to overrule my silicon friend, when it suggests something that makes a lot of sense, it also makes sense to follow its suggestion.

27. f4 Nc7 28. Nc2

(28. f5?! TIL what “ungestüm” is in English: Impetuous! 28… Nd5 29. Qe4 (29. f6 gxf6 30. Qd2 fxe5 31. Qxh6 f5 =) 29… Qg5 and the option …Qe3+ keeps Black in the game: 30. f6 Qe3+ 31. Qxe3 Nxe3 32. Rf3 Ng4 33. fxg7 Be6 with a doubleedged position that should be equal objectively.)

28… Nd5 29. c4


29… Nc3??

A pity, this just blunders the game away. A misclick I assume, given how well Black had played so far.


(29… Nb4 was neccessary. After 30. Nxb4 Qxb4 31. Qe4 Kf8 32. Kh2 Qc3 33. d5 f5 (diagram) it is important not to fall for the engine’s suggestion 34. Qe3?! that promises a +1 advantage. More or less by force the game will fizzle out in an endgame, in which White is an exchange up, but it is unwinnable nonetheless. (After 34. Qf3! instead White keeps pressing.) 34… cxd5 35. Rfc1 Qb4 36. cxd5 Rxd5 37. Bc4 Qd2 38. Rxa3 Qxe3 39. Rxe3 Rc5 40. Rcc3 b5 41. Bb3 Rxc3 42. Rxc3 Bd7 43. e6 Be8 44. Rc7 Bxh5 45. e7+ Ke8 46. Bd5 Ra6 47. Bc6+ Rxc6 48. Rxc6 Kxe7 =)

30. Qe1 Na4 31. Rxa3


Black has lost “only” one pawn, but it was the most important one, the major source of his counterplay. Now White can just bulldoze him, and Black has nothing else than sitting there and waiting for his execution. White is winning.

31… Bd7 32. Qe4 f5 33. Qe3 b5 34. Rfa1


Black resigned.


5 1 vote
Article Rating
Benachrichtige mich bei

Most Voted
Newest Oldest
Inline Feedbacks
View all comments
1 Jahr zuvor

Gibt’s das auch als daunladtbare pgn, bitte? Der Autor wird’s hier ja nicht handschr. eingefügt haben? Es würde das allg. Spülniwo heben helfen.