Thursday, January 1, 2015

Danish Gambit (Also evaluation bug in Stockfish 5 and Komodo 8)

     In the Danish Gambit, aka the Nordic Gambit, (1. e4 e5 2. d4 exd4 3. c3) white sacrifice one or two pawns for the sake of rapid development and the attack. However, with care, Black can either accept one or both pawns safely, or simply decline the gambit with good chances. From the very beginning the nomenclature of the Danish Gambit was very confusing. The idea stems from a famous correspondence game London–Edinburgh, 1824: 1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Bc4 Bc5 5.c3 Qe7 6.0-0 dxc3 7.Nxc3.
     The Swede Hans Lindehn played 1.e4 e5 2.d4 exd4 3.c3 on a regular basis from 1857 at the latest. He defeated the later World Champion Wilhelm Steinitz with his gambit in London, 1864. It is possible, that Severin From met Lindehn in Paris in this period and learned about the gambit there. According to Graham Burgess, in Denmark itself, the opening is called the Nordic Gambit.
     The opening was popular with masters of the attack including Alekhine, Marshall, Blackburne, and Mieses, but as more defensive lines for Black were discovered and improved, it lost favor in the 1920s. Today it is rarely played in top-level chess. Frank Marshal wrote that he used to use it all the time at simuls at his club, but eventually everybody got booked up with the best lines so he had to give it up.
     The idea is for white to threaten to establish a strong pawn centre with pawns on d4 and e4, controlling the key squares c5, d5, e5 and f5. White also offers a sacrifice of the d-pawn by challenging the d4-pawn with c2-c3.
     If Black declines the gambit, White will often end up with an isolated pawn on d4 but will aim to compensate by activating the pieces into the early middlegame.
     If Black accepts the gambit, White will generally aim to put pressure on f7 with Bc4, Qb3 and/or Ng5, and Nc3-d5 is also a common motif. If Black plays ...Nf6 too early then the knight is a target for an e4-e5 push from White. In lines with ...Nc6, White has to be careful of the ...Na5 fork, which can force the exchange of White's c4-bishop for the black knight on c6.
     However, if Black refrains from playing an early ...Nc6, then there is greater scope for Black to get in ...d7-d5.

     While analyzing this game with Komodo 8 and Stockfish 5 I noticed a problem I have run into several times in the past...incorrect evaluations
     I posted a warning a couple of years ago HERE about evaluation bugs in Komodo 3. I also discovered that a few players have posted regarding evaluation bugs in Stockfish 5.
     I have run into this problem several times in tactical situations and/or positions with material imbalances. Both K8 and SF5 give totally incorrect evaluations. Is that reason to cause suspicion about the programming? I have no idea. Anyway, for me, it's enough of a problem that I have started running Houdini 2 as kibitzer just to alert me to these situations! 

But this post is about the Danish Gambit and this snazzy little Mieses vs. Marshall game.  If you ever get tired of the Ruy Lopez and Sicilians then it's fun to play loer games like this once in awhile!



1 comment:

  1. A great example is the party blindness Lafuente - Shredder from 2005:   (move black) 5rk1 / pB2qppp / 1p1rpn2 / 1Pp5 / 2P5 / P1Q1P3 / 3N1PPP / 3R1RK1 b - - 0 19     White flocked black bishop on b7. Each engine immediately sees that Black must beat the white bishop, and the difference in assessment between this and the next on the list is the movement from the first second huge: two - three points. And what did Shredder, famous tactics? He "forgot" to beat runner ..

    ReplyDelete