Steinitz Defense

In the Main Line of the 4.Ng5 Variation of the Two Knights' Defense, White can play the Steinitz Defense, kind of a weird opening where White plays his Knight to the h3 square where Black can exchange it inflicting White with doubled h-pawns.

Here we see the basic problem for White---the weakness of the h2 square. That is just enough for Houdini to equalize with Black. Houdini earns its nickname here steering for a piece down ending with Black it can't lose.

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Na5 6.Bb5ch c6 7.dxc6 bxc6 8.Be2 h6 9.Nh3 Bc5

10.d3 O-O

11.O-O It seems that Black should play Bxh3 at some point, but looking ahead Houdini has the idea of Bf3/Bg2 where the Bishop plays the role of the “tall”pawn) Rb8

12.Nd2 (This is dead equal) Bxh3

13.gxh3 (Steinitz and Fischer reckon that White gets compensation for his busted pawn structure; the only winning chances for White long term is to mobilize his Queenside pawn majority, but it is clear that there is a lot of chess to be played before that happens) Nd5 (So far there have been no surprises; Black has made a series of “obvious” moves.)

14.Bf3 Nf4 (Black, so far, has excellent compensation. What is clear is that the Black King Knight is worth at least as much as White's King Bishop, so the minor exchange confers no advantage to White.)

15.Bg2 f5 (This is the thematic move in the 8.Qf3 variation and the 8.Bd3 variation so it is not surprising to see it here. One problem facing Black is that he cannot afford just to exchange pieces. If we strip off all the pieces the pawn on h3 becomes hugely important. In this variation, Black has to create threats on the Kingside to Queen. What is not generally recognized in endgames is that when deciding to permit your pawns to be doubled ask yourself one simple question, “What do I intend to do with the pawn? Am I going to try to Queen the pawn or just prevent my opponent from Queening?” If the latter is the case then the pawn on h3 is just fine. You will observe that in a pure King and pawn ending that Black has a 4-3 majority on the Kingside; White has a 4-2 pawn majority on the Queenside. White is better.)

16.Qf3 Ng6 (This is a sign that Houdini may not like its position; avoiding Nxg2 suggests that Houdini thinks the Knight is better retreating rather than exchanging)

17.Kh1 Rf6

18.Qe2 (I like White here +/=; Houdini is not quite to the point where it says this is +/=) Qd6

19.Qe1 Bb6!

20.b4! (Houdini seeks simplification; Bc7 is a very real threat) Qxb4

21.Ne4 Qxe1 22.Nxf6ch gxf6 23.Rxe1 Bxf2 24.Bd2 Bxe1

25.Bxe1 Rb2!

26.Bxa5 (It will be interesting to see how Houdini salvages this endgame for Black) Rxc2 27.Rg1 Kf7 28.Bf3 Rxa2 29.Bd8 Ne7 30.Bxe7 Kxe7 31.Rc7 Rc2

32.Rxa7 = Since the Bishop is the wrong color of the Rook pawn White has no realistic winning chances.

Dude, i have to ask: so what?

The only reason for this post is that GM Nigel Short led a renaissance in this opening in the 1990's feeling it gave White no worse chances than the more common 9.Nf3. What is clear is that h2 is a huge weakness that cannot be rectified. 9.Nf3 may be good for a slight plus---the Steinitz Defense---equality.

Yes, White's sub-40% score in the 9.Nf3 variation is a strong argument for it giving a slight plus.

At a practical level I have long preferred 8.Qf3. At a practical level I'd rather play Black against 9.Nf3. In theory White might have a slight plus; here is one of the most complicated positions in chess I have ever seen. The better player will win:

4.Ng5 d5 5.exd5 Na5 6.Bb5ch c6 7.dxc6 bxc6 8.Be2 h6

9.Nf3 e4 10.Ne5 Bc5! (For Bd6, see below) 11.c3 Bd6 12.d4 exd3 13.Nxd3 O-O 14.O-O Re8 15.g3 Bc7 16.Na3 Bb6 17.b4 Nb7 18.Nc4 Ne4

19.Bf3 Bf5 (Nxc3 20.Qc2 Qf6 21.Bb2 Bd4 22.Rfe1 Bf5 23.Ne3 Bh7 24.Kg2! h5! (I doubt most players would find this so the game more or less leaves the path where a strong player would find the best moves. Until now the move sequence, while sharp, is not that difficult to find; with h5! Houdini earns its rating) 25.Rac1 h4 26.gxh4 Nxa2 (Although Houdini gives White a slight edge here (+.17 at a depth of 32), this is an incredibly complicated position where in a human-human contest anything could happen)

20.Nde5 (Houdini is beginning to like White's chances) Qxd121.Rxd1 Bxf2ch 22.Kg2 Nxd3 23.Kxf2 Nxd1ch

24.Bxd1 (We leave the game here with the observation that it has unbalanced elements and that White, perhaps, has a slight edge)

In the next game we will look at the other primary choice for Black, 10...Bd6

10.Ne5 Bd6

11.d4 (Houdini has f4 as equal) exd3

12.Nxd3 O-O 13.O-O c5 14.Nc3 Qc7 15.Bf3 Bb7 16.h3 Rab8 17.Bxb7 Qxb7 18.Bd2 c4 19.Ne1 Rfd8 20.Nf3 Bf4 21.Bxf4 Rxd1 22.Raxd1 Re8

23.Bc1 (Houdini clearly saw this position many moves earlier; the question many players would ask is why would White want this position? In terms of material, White has Rook, Bishop and pawn for the Queen, but often in middlegames like this the Queen might be preferred. It is instructive that Houdini has had advantage White for the entire opening and middlegame; now at a Depth of 30 Houdini finally begins to realize that White can claim only a tiny plus (+.21, =). At a practical level many players would prefer to have the Queen rather than Rook, Bishop and pawn. The problem for White is that he could play flawless chess for thirty moves only to wind up drawing by perpetual check. At a practical level 10...Bd6 seems equivalent to 10...Bc5).

Based on these two centaur games the main line as theory has already suggested is playable for both sides. The only caveat we would offer the reader is that 8.Qf3 is vastly more complicated than either 8.Bd3 or 8.Be2, and, as a practical matter, offers greater scope for original play.

Troll thread.