Counterplay / Readers Respond
Practically Speaking
QUERYING NARODITSKY Congratulations to Daniel Naroditsky for
three excellent articles to begin his col- umn “The Practical Endgame.” The writing style is interesting because of what is not said, leaving the readers to work out many details for themselves. Also very interesting is that he poses actual game positions as problems, rather than posing compositions. Actually, in the August article working
out details of the GM Roman Dzindzi - chashvili-GM Lubomir Kavelek ending leads me to some questions. The position after 48 moves is shown below as an endgame problem.
rather than a “computer move.” What I consider the “main line” because
it is most analogous to the actual game, despite a different move order, goes:
50. ... Ne4 51. a5 h4 52. a6 Kh2 53. a7 Kxh1 54. a8=Q h3 55. Qd5
With the black king already on g2, this only draws. If 55. ... Kg2 56. Qxf5, and the lines
finally appear in the six-piece endgame tablebase offered on the Internet. Also 55. ... h2 56. Qxf5 Nf2 succumbs, even though the knight cannot be captured. More resistance is offered by ...
54. ... Kh2 55. Qd5 Ng3
... avoiding immediate material loss (and staying out of the tablebase). But ... 56. Qf3
... ties Black in knots.
WHITE TO PLAY AND WIN. White’s two previous moves were
awarded a total of three question marks, yet White has a forced win that begins with logical moves, as far as I can see. In the game White played 49. Kb6 and only drew. My proposed winning line begins
49. Kb4 Kxg3 50. Bh1! I am only a class player and I have no
chess engine, but I was mindful of Daniel’s comment that the bishop cannot go past e4, or the h-pawn becomes unstoppable due to Ne4. I was also mindful of GM Irina Krush’s emphasis on making minor improvements in one’s position before embarking on aggressive maneuvers, in her discussion in the same issue of one of her games (p. 29). I envisioned the Zwis- chenzug 50. Bh1 Kh2 51. a5 Kxh1 as being a minor improvement over 50. a5 Kxg2, as occurred in the game. Obviously, the black king is worse off on h1 than on g2. This is why I consider Bh1 to be logical,
8 October 2014 | Chess Life
The loss of a tempo with 50. Bh1 gives Black another main try: 50. Bh1 f4 51. a5 f3 52. Bxf3 Nxf3 (The lines now appear in the tablebase.) 53. a6 Nd4 54. Kc5! Nxb3+ 55. Kd5 ... and later White uses the knight’s
presence to trap the black king on h1 without it being stalemate.
With this long introduction, I come to
my questions for GM Naroditsky: Is it true that the most powerful chess engines do not find this win for White? Possibly there is a flaw in my hand analysis, but if it stands up, it shows that machines are not as infallible as many chess writers seem to assume. More generally, are most of the general rules in Reuben Fine’s Basic Chess Endings (BCE; see last chapter) now considered to be unreliable due to computer analysis? The first three articles, as well as recent “Endgame Labs,” seem to convey the message that the only safe method is exact calculation from scratch. At first I thought the above analysis
might be worthy as a composition (such as a Benko Baffler), but it turns out that Bh1 is too strong. Although the first two moves are unique, there are alternative
wins after that, in which White is not required to sacrifice the bishop outright. (A) If 50. ... Kh2, both 51. a5 and 51. Bc6 suffice.
(B) If 50. ... Ne4 51. a5 Kh2 tries to provoke Bxe4, but both 52. a6 and 52. Bf3 h4 53. a6 suffice.
(C) In the latter case, after 53. ... Kg3, both 54. Bh1 (intending to sacrifice) or 54. Bxe4 suffice.
—Allen Van Gelder
professor of computer science University of California, Santa Cruz via e-mail
GM Naroditsky responds: Your analysis and conclusions are spot-
on from top to bottom. In my analysis, I took 49. Kb6 for granted, and did not thor- oughly check the consequences of 49. Kb4. Given about 20 seconds, my engine (Hou- dini 4) was indeed able to discover the win. Computer infallibility is an asymptotic notion, but given enough time, a strong engine running on good software should be able to accurately deconstruct even the most complex endgame positions. Conse- quently, the two question marks should be assigned to 49. Kb6 rather than to 48. Bg2 —I stand corrected! To answer your second question (about
Fine’s rules in BCE), I do believe that accu- rate calculation, rather than the application of general rules, lies at the heart of endgame mastery. Of course, a line must be drawn between over-the-board play and analysis, but the universality and reliability of endgame generalizations have certainly been put in question.
Send your letters to
letters@uschess.org or post on the Chess Life Facebook group or the
uschess.org Issues Forum. Letters are subject to editing for style, length, and content.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76