26 May 2019

Textbook Ending

While continuing to work my way through drawn endgames from my personal database (see "King Position"), I found that I missed a textbook win that I have taught to my chess students. However, my failure dates from 2006 and I've been teaching it since about 2014.

Reuben Fine, Basic Chess Endings (1941) offers this position (No. 40).

White/Black to move

White to move draws, Black to move wins for White.

Mark Dvoretsky, Dvoretsky's Endgame Manual (2003) offers a similar position.

White/Black to move

Again, it is only a win for White if Black is on move. The key, as Dvoretsky points out, is that White's king must occupy h6 while a pawn remains on the second rank, preserving the option of moving one or two spaces. That way, Fine's position can be forced with Black to move.

In a game on Free Internet Chess Server, I had this position.

White to move

1.Kg5 wins.

I played 1.h4 and the game was drawn.

25 May 2019

King Position

I am working my way through pawn endings that I have played over the past twenty years that ended drawn (see "Four on Four"). Some of these are completely lost for one player who was saved by the clock. Many of them are blown wins by one side or the other. Several games reveal that both players threw away a win, as yesterday's "More Endgame Errors".

Of course, these games nearly all come from blitz games played online. By the time a pawn ending is reached, the players may have only a few seconds remaining. Even so, better instincts can be trained, leading to better endings when one must play instantly.

This morning's review turned up an interesting position with both sides possessing pawn majorities, but White's king is better placed.

Stripes,J -- Internet Opponent [D32]
rated blitz match  freechess.org, 17.10.2005

White to move


Seems like the wrong pawn, but ...


33...a5 34.Kf3 Kg7 35.b4 axb4 36.Ke2 Kf7 37.a5 Black's king is outside the square.

34.b4 Kf7

White to move


35.Kf3!+- Ke7 36.Ke4 Kd6 37.Kd4 a6 38.b5

Black to move
Analysis diagram


(38...axb5 39.axb5 Kc7 40.Kd5)

39.f4 Kc7

(39...g5 40.fxg5 fxg5 41.Ke4)

40.Kc5 Kd7 41.Kd5 Kc7 42.Ke6 g5 43.fxg5 fxg5 44.Kf5+-


Black, nonethless, managed to fail, too. The game was drawn at move 61.

24 May 2019

More Endgame Errors

Working my way through more than 200 drawn pawn endgames is revealing how poor my endgame skills were fifteen to twenty years ago. I bought Dvoretsky's Endgame Manual in 2003 or 2004, and that was after training against Hiarcs or Fritz with 100 pawn endgames downloaded from a website with terrific tactics and ending PDFs.

I plan to develop training material for my students from these games.

White to move

Internet Opponent -- Stripes,J. [A46]
ICC 5 0 Internet Chess Club, 24.11.2003


37.Ke3?! a5

     (37...g5 38.g4?? fxg4 39.hxg4 h4 40.Kxe4 a5 and Black has the upper hand)

38.g3 a4 39.Kd2 g5 and Black has the upper hand



37...g5 and Black has the upper hand

38.b3+ Kd5

White to move


39.h4 is the only move that holds the draw.



Did I know that the game had suddenly turned in my favor?

40.Ke2 a5 41.Kd2 Kc5 42.Kc2

Black to move


42...g4 shows a grasp of the right idea.

43.axb4+ axb4 44.cxb4+??


44...Kxb4–+ 45.Kb2 g4 46.hxg4 hxg4 47.Kc2

Black to move


There is a breaktrough idea based on the square of the pawn. Either I was extremely harried because of the clock, or simply lacked understanding of this elementary endgame concept.

47...e3 48.fxe3 f3 49.gxf3 g3 and the pawn will promote.

47...Ka3 48.Kc3 e3 49.f3 gxf3 50.gxf3 e2 51.Kd2 Kxb3 52.Ke1 Kc2 53.Kxe2 Kc3 54.Ke1 Kd3-+

48.gxf3 exf3 49.Kd2= Kxb3 50.Ke3 Kc4 51.Kf4 Kd4 52.Kxg4 Ke4 53.Kg3

Black to move


The only move, but at least at this point in my life, I had an understanding of the opposition.

54.Kxf3 Kf5= 

Black was stalemated on move 73

23 May 2019

Four on Four

While reading Van Perlo's Endgame Tactics, I was inspired to review some of my online games over the past two decades. I searched for games that resulted in a four vs. four pawn endgame, thinking that maybe I would find a breakthrough position similar to those in that section of the book. With a database of nearly 100,000 games, there are many pawn endings. I opted to focus on games that were drawn, reducing the number to 202.

There are plenty of examples where my or my opponent's lack of endgame skill surrendered half a point. These are from the years 1999-2002.

Black to move

I had Black and played 40...f4. The game was drawn at move 55.

White to move

With White, I played 44.Kf4, offering victory to my opponent whose own errors let me back into the game.

White to move

For reasons I can no longer explain, I played 45.Kd7, which my opponent graciously failed to exploit.

White to move

My opponent played the correct 48.Kxb6 and the game was drawn at move 74. Would I have proceeded with competence had he tried 48.h3??

In some games, other pieces were on the board that were exchanged into a drawn or even losing ending.

White to move

I had White. The game continued 49.Rf6?? Rxf6 50.gxf6 Kxf6 51.Kf3 c3?? and was drawn at move 61. White has a win from the diagram, and Black does after 51.Kf3.

22 May 2019

Playing against the Isolani

In the Inland Empire Open last weekend, I was on board two in the last round, fighting for second place. My opponent and I stood at 3.0 points, while those on board one stood at 3.5 and 4.0. There were other players with 3.0 as well. I faced Pat Herbers, who is at his rating floor of 1900. Several prior games against Pat have been draws, and I have lost a few. I have never beat him.

Stripes,James (1889) -- Herbers,Pat (1900) [D30]
Inland Empire Open Spokane (5), 19.05.2019

1.d4 d5 2.c4 e6 3.Nf3 c5 4.cxd5! exd5 5.Bf4?!


5...Nc6 6.e3

6.Nc3 Nf6 7.e3 cxd4 8.Nxd4 Be7 9.Be2 0–0 10.0–0 Be6 11.Rc1 Rc8 12.Qd2 Nxd4 13.exd4 Bb4 14.Bd3 Ne4 15.Qe3 Bxc3 16.bxc3 Nd6 Schleifer,M (2240)--Polacek,J (2250), Toronto 1992 was drawn in 47 moves.

6...Nf6 7.Be2 Bd6 8.Bxd6 Qxd6 9.dxc5 Qxc5 10.0–0 0–0

White to move


An instructive game appears in the database.

11.Nc3 Rd8 12.Rc1 Qb4 13.Qc2 d4 14.a3 Qe7 15.Nxd4 Nxd4 16.exd4 Rxd4 17.Nb5 Rd8 18.Nxa7

Black to move
Analysis Diagram


(18...Rxa7 19.Qxc8 Rxc8 20.Rxc8+ Ne8 (20...Qe8 21.Rxe8+ Nxe8) 21.Bb5)

19.Qxf5 Qxe2 20.Qb5 Re8 21.Qxb7 Rab8 22.Qc6 Rxb2 23.Nc8 Re6 24.Qc7 h6 25.Nd6 Kh7 26.h3 Rd2 27.Nf5 Rd7 28.Qb8 Qd3 29.Ng3 Qxa3 30.Rc8 g6 31.Kh2 Qd6 32.Qa8 h5 33.Rd1 Qe7 34.Rxd7 Nxd7 35.Rh8+ Kg7 36.Qg8+ Kf6 37.f4 Rb6 38.Qc8 Kg7 39.Nf5+ gxf5

(39...Kf6 40.Nxe7)

40.Qg8+ 1–0 Morozevich,A (2678)--Grischuk,A (2701), Togliatti 2003.

11...Be6 12.Nb3 Qe7 13.Nbd4 Rac8

White to move

I have a clear plan to play against the isolani.

14.a3 Bg4 15.Qa4 Nxd4 16.Nxd4 Bxe2 17.Nxe2 a6 18.Rfd1 Rfd8 19.Rd3 Rc7 20.Rad1 Rcd7

White to move


I had a vague sense that I would want my queen on the kingside after Nd4. Perhaps I could have found some other way to probe for weaknesses.

21...Qe4 22.f3

22.Qxe4 dxe4 23.Rxd7 Rxd7

(23...Nxd7 24.Ng3 Kf8 25.Nxe4)

24.Rxd7 Nxd7 25.Ng3 Nf6 should be equal

22...Qxf4 23.Nxf4 Kf8 24.Kf2 Ke7 25.Rd4

25.Ke2 g5 26.Nh3 Rc8 27.R3d2

25...h6 26.R1d3 Rd6 27.Rc3 Rc6

White to move


Abandoning my plan against the isolani in exchange for temporary control of an open file.

28.Rb4 maintains my opening edge.


I have converted my opponent's isolani into hanging pawns.

29.Rb4 c5 30.Rb6 Rd6

White to move

My opponent offered a draw here. We both thought my position was slightly better. Conversing with several players in the room while we were waiting for the round's pairings, I pontificated that if any of us were there for money or for chess rating, we were delusional. All of us are relatively weak players compared to professionals, and none of us will earn substantial money for chess. We are there because we enjoy the game.

Naturally, I needed to play on. I paid an entry fee to play chess.

31.Rb8 Rd8 32.Rb3

32.Rxd8 Kxd8 33.Nd3 Nd7 34.b4 Kc7 35.bxc5 a5 36.e4 dxe4 37.fxe4= The computer says equal, but this ending can be very difficult to play.

32...g5 33.Nd3 Nd7

33...c4 34.Rb7+ Rd7 35.Rxd7+ Kxd7 36.Nc5+ Kc6 37.Nxa6 Kb5 38.Nb4=

34.Rb7 c4 35.Ne5 Ke6

White to move


I considered exchanging into a pawn ending, but could not see clearly that I was better. When I arrived home from the tournament, I entered the game this far into the database and played the pawn ending against Stockfish. I lost.

36...Rc8 37.Nd4+ Kd6 38.Ra7

38.g4 seems unclear
38.Nf5+ Kc6 39.Ra7 Rb8 40.Nd4+ Kd6 41.Rxa6+ Kc7 42.Rc6+ Kd8 43.g4 perhaps White is slightly better.

38...Rb8 39.Rxa6+ Kc7 40.Rxh6

40.Rc6+ gains a critical tempo in the play of kings. 40...Kd8 see at 38.Nf5+. Hastily grabbing the pawn on h6 without first driving back the king let the initiative slip away. It was easy to spot this error after the game.

40...Rxb2+ 41.Kg3 Ne5

White to move


42.f4 gxf4+ 43.exf4 Rb3+ 44.Kh4 Ng6+ 45.Kg5 Rd3 with an edge for Black
42.Ra6 seems best.

42...Rd2 43.Rf5

43.Nb5+ Kc8 44.Rf5 Nd7 45.Rxg5 Rd3 46.Rg4 Nb6 47.h4±


White to move


44.f4 gxf4+ 45.Rxf4=


44...Kb6 45.a4 Ka5 46.Rxf6 Kxa4 47.Nd4 c3 Black is better.



45...Nd7 46.f4??

Black to move

This move would have been useful a couple of moves back, but now simply gives up a pawn for nothing.


Pat could have ended my hopes sooner with 46...d4! 47.Nd5 gxf4+ 48.exf4 Rb2 49.Nxf6 Nb6–+ White is ahead two pawns and irredeemably lost.


47.Na2 Rxe3+ 48.Kg4 Re4 49.Nb4+ Kd6 50.Rxd5+ Ke6 51.Kf3 Rxf4+ 52.Ke3

47...Rxe3+ 48.Kf2 gxf4

White to move



49...Re5 50.g4

I looked at 50.Rxe5 fxe5 51.Nxd5 Kxd5 and thought that Black should win


50...Rxf5 51.gxf5 d4 52.h4= I was optimistic about this line, but the engine says it is equal.

51.Ne2 Re4 52.Rf4


52...Rxf4+ 53.Nxf4 Ne5

White to move


54.h4! may salvage the draw 54...Nxg4+ (54...d3 55.Ke3 Nxg4+ 56.Kd2) 55.Ke2.

54...c3–+ 55.Ke2 Kc5 56.g5


56.Ke1 Kc4 57.h4 Nd3+ 58.Nxd3 Kxd3–+

Black to move

56...fxg5 57.Ne6+ Kc4 58.Nxg5 Kb3 59.Ne6 d3+ 60.Ke3 d2 61.Ke2 Kb2 62.Nc5 d1Q+ 63.Kxd1 c2+ 0–1

Herbers finished in second place and won $100. I still have not beat him, although I had some chances. We went over the game quickly together and were the last to leave the playing venue.

14 May 2019

Exchanging into Pawn Ending?

This position arose in Lasker -- Tarrasch, St. Petersburg 1914, the source for Van Perlo 7 (see "Endgame Tactics").

Black to move

The game continued 37...Bxg7 38.Bxf5 Kxf5 39.Kxg7 a5, reaching the position where Van Perlo celebrates Lasker's "amazing escape" (Van Perlo's Endgame Tactics, 21).

White to move
White has a drawing combination
Did Black have something better?

13 May 2019

Endgame Tactics

Van Perlo's Endgame Tactics, New, Improved and Expanded Edition (2014) arrived on Friday. It is a large book of more than 600 pages and more than 1300 positions. The diagrams are small and set next to detailed analysis of the position that includes the source, what happened, what might have happened.

These two positions arose in the course of my play against Stockfish, and analysis of that play, from Van Perlo 6, a position that arose in Taimanov -- Cuellar, Leningrad 1973.

White to move

White to move

01 May 2019

Stubborn Defense

How can Black put up the strongest resistance to White's ambitions?

Black to move