Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

5/29/18 - Bereolos-Zapata, 2018 Ronald Finegold Memorial

I've added my recent game against Alonso Zapata to the GM games section. As I mentioned in the post for my game against Ben Finegold, I played at the Chess Club & Scholastic Center of Atlanta for the first time in April. The occassion of that tournament was in memory of Ben's father Ron, who passed away in 2014. I never had the honor of playing the patriarch of the Finegold family, but was certainly aware of his reputation as one of Michigan's strongest players. His chess legacy was carried on by Ben and his brother Mark, who also reached the master title. Ben's son Spencer has also crossed 2200 to make it three generations of masters from a single family.

Against GM Zapata, I move ordered myself out of my normal repertoire. I still kept a decent opening edge when he played a line I knew from the Black side. I played a bit tenatively in the middlegame and eventually he reached an equal ending. It looked like he was slowly grinding out an advantage, and may have even missed a winning chance in the time scramble. However, I managed to hang on for the half point.

5/23/18 - Bereolos-Spalding, 1993 Purdue Hoosier Tour Bonanza #6

Another major piece battle occured in my game against Jim Spalding in the 6th Purdue Hoosier Tour Bonanza in 1993. Since the Hoosier Tour allocated points based on the number of rounds, our club president, Brad Watson, came up with the idea of holding 6-round single-day tournaments with fast time controls to attract players to our tournemants. These got reasonable turnouts, this one had 25 players.

This ending is not as complex as some of the others I've looked at recently, but still has many interesting points. This was a pretty ragged game, like many rapid games are, but had settled into an equal ending after 43. Ke4-f4

43...Kf6 43...Rb5 44.Rxd3 h2 45.Rd1 Rxb2 46.Kg3 Kf5 47.Rh1 Ra2= 44.Ke4? this allows the Black king to penetrate. It was time to head for a draw with 44.Rh2 Rb5 45.Rxh3 d2 46.Rh6+ Ke7 47.Rh1 Rxb2 48.Rd1= 44...Kg5 45.Ke5 Kh4 46.Kf4 b5? This is a poor move locking in the Black rook and turning a winning position into one that is close to lost. After 46...Rb5 47.b4 Rd5 48.Ke4 Rd7 49.Kf4 Rd8 White is in zugzwang 47.Rf2 better is 47.Rh2 but it looks like Black holds the draw by the skin of his teeth after 47...d2! (47...Kh5? 48.Rxh3+ Kg6 49.Rh2+-) 48.Rxd2 Kh5 49.Kg3 Kg5 50.Rf2 (50.Kh2 Rxf3 51.Rd5+ Kg4 52.Rxb5 (52.Rd4+ Rf4=) 52...Rf2+=) 50...Re3 51.Kxh3 Kf4 52.Kg2 Rb3

and it doesn't look like White can make progress despite his two pawn advantage since the pawn ending after 53.Rd2 Rxf3 54.Rd4+ Ke5 55.Kxf3 Kxd4 is drawn 47...b4? allowing White what should be a decisive passed pawn. Black should mark time with 47...Kh5 when it looks like White doesn't have anything better than transposing to the previous note with 48.Rd2 Kh4 49.Rh2 48.a4 Kh5 49.Rd2? showing a complete lack of awareness of Black's resources. There is really no reason for this move as pushing the d-pawn does not free the Black rook. White should just get on with pushing the a-pawn 49.a5 Rc3 50.bxc3 bxc3 51.Ke3 +-

49...Kg6? missing the saving resource 49...Rc3! 50.a5 (White loses after 50.bxc3? bxc3 51.Rxd3 c2! 52.Rc3 h2!) 50...Rc2 51.a6 (51.Rxd3 h2 is dangerous for White 52.Rd8 Rc6) 51...h2 (51...Rxd2 52.a7 h2 53.a8Q) 52.Rxh2+ Rxh2 53.a7 d2 54.a8Q d1Q 55.Qh8+ Kg6 56.Qg8+ (56.Qxh2? Qd6+!) 56...Kf6 57.Qf8+ with perpetual check since the Black queen is lost to a skewer if the Black king goes to the d-file 50.a5 Kf6 50...Rc3 doesn't work now because after 51.bxc3 bxc3 52.Rd1 the threat of Rg1+ gives White the tempo needed to bring his king back to the defense. 52...h2 53.Ke3 51.a6?! while White is still winning after this move, it makes the job much tougher. 51.Kg3 controls the h-pawn so none of the ideas with Rc3 work and the Black king is still too far away to catch the a-pawn. 51...Ke5 52.a6 51...Rc3!?

Black finally recognizes his last chance 52.bxc3? 52.a7! Rc8 53.Rxd3 h2 54.Rd6+ Kg7 55.Rd1 Ra8 56.Rh1 Rxa7 57.Rxh2+- is winning for White, but Black is not without practical chances to hold 52...bxc3 53.Rxd3 c2 54.Rc3 h2 55.a7 h1Q 56.a8Q Qh2+ This doesn't spoil anything, but with the king in front of the pawn and the White queen stuck in the corner, there is little to fear from the Q+P vs. Q ending after 56...c1Q+ 57.Rxc1 Qxc1+ 57.Ke3 Qg1+ 58.Kd2 Qd1+ 59.Ke3 Qe1+ 60.Kd4 Qd2+ 61.Kc4 Qe2+? this was the last chance for Q+P vs. Q after 61...Qxc3+ 62.Kxc3 c1Q+= 62.Kb3+- Qd1 63.Rxc2 Not a mistake, but White should probably improve the position of his queen before taking the pawn. The tablebase indicates 63.Qf8+ as the fastest mate. 63...Qxf3+ 64.Qxf3+ 1-0

The lessons from this game: 1. Recognize your opponent's resources 49. Rd2? and 51. a6?! showed that I was completely oblivious to ...Rc3 by Black. 2. Activity is always a key in the ending. 44. Ke4? let Black's king become too active and 46...b5? made the Black rook passive instead of activating it with 46...Rb5. 3. Don't be lazy in calculation. It was relatively straightforward to calculate the variation from 52. bxc3? all the way to 56...Qxc1+ and I should have recognized that as a draw. By process of elimination that should have led to 52. a7! Although, that line can't be calculated to a win, the moves up to 57. Rxh2 are easy to find when White is a pawn up and has all the chances.

5/22/18 - BCE-599b

Position 599b in BCE provides a good follow up to the discussion of Q vs. R+P. Here, Black has a far advanced d-pawn, but it is not enough to hold. Fine declares it lost for Black to move, but a draw if it is White to move. Benko corrected the analysis in the revised edition, showing a win for White. I will quibble about him keeping Fine's language still calling it "a critical position". Since there is no mutual zugzwang, the position isn't critical, it is just won for White.

This position illustrates the difference between a central (or bishop's) pawn and a knight's pawn. If you push position 599b two files to the left so that Black has a b-pawn, we have the following drawn position.

Trying to follow the same method as in the 599b analysis we have 1. Qa5 Rc2 and there is no check on the third rank, nor any other way to utilize the queen on the left hand side of the pawn, the king is able to cover the key squares and the rook has enough safe squares along the c-file.

5/12/18 - Colias-Bereolos, 1981 Dunes Country Spring Open

My game against Billy Colias at the 1981 Dunes Country Spring Open provides more complex material on queen versus rook and pawn endings. Here, we pick it up a few moves before Q vs. R appears on the board after 58. d7

It is readily apparent that the rook will not be able to stop the White pawn, but Black still has some defensive resources. My choice 58...Kg2 is not the most testing. It does stop White from immediately queening because 59. d8Q? is met by 59...Re1+ 60. Kd5 Rd1+ and Rxd8 winning. However, after 59. e6 Black's pieces are too uncoordinated and Billy finished me off without much difficulty 59...Rd1 60.e7 Rxd7 61.e8Q Rd6 62.Qe7 Ra6 63.Qb4 Rc6 64.Qd2+ Kg3 65.Qe1+ Kg2 66.Ke5 Rc5+ 67.Kd4 Rf5 68.Ke4 Rf8 69.Qe2+ Kg3 70.Qf3+ Kh2 71.Ke3 Rb8 72.Qe4 Rb3+ 73.Kf2 Rg3 74.f5 g5 75.f6 g4 76.f7 Rg1 77.f8Q g3+ 78.Ke2 1-0

Instead of 58...Kg2, 58...Kg4 is a more challenging move. It carries the same idea to stop White from immediately queening and positions Black to be able to capture the remaining White pawns. If White continues as in the game, it appears that Black has a narrow path to a draw, although it would be very difficult to find in practice 59.e6? Re1+ 60.Kd5 Kxf4= 61.e7 (61.Kd6 Kf5! 62.e7 Re6+! 63.Kd5 Re5+! 63...Rxe7? loses a sample line is 63...Rxe7? 64.d8Q! Re5+ 65.Kd6 Re6+ 66.Kd7 Re4 67.Qa5+! Kf4 68.Qd2+ Kg4 69.Qd5 Kf4 70.Qf7+ Kg5 71.Qf3 Re5 72.Kd6 Rf5 73.Qe3+! Kg4 74.Qg1+ we'll see something similar in the note to Black's 63rd move 64.Kd4 Rxe7 65.d8Q! Re4+ Black doesn't have this check in the previous line with the king on d5 66.Kd5 Rf4! and Black will be in time to set up a barrier on the f-file to prevent White's king from entering) 61...Rxe7! 62.d8Q! Re5+! 63.Kd6

65...g5! The only way to draw. I'm not sure how a human would find this move. It is fairly simple to see that Qh8 will fork the rook and h-pawn. Even if you knew for sure that the resulting R+g vs. Q position was a draw, do you really give up a pawn to get there? The seemingly plausible 63...Rf5? loses 64.Qh4+ Kf3 65.Qh3+ Kf4 66.Ke6 Kg5 66...Rg5 67.Qh4+ Rg4 68.Qf2+! Kg5 69.Qf6+ Kh6 White can use his king to create mating threats.70.Kf7 h4 71.Qh8+ Kg5 72.Qe5+ Kh6 73.Qe3+ Kh5 74.Kf6 Rg3 75.Qe2+ Kh6 76.Qd1 h3 77.Qe2 g5 78.Qe4 Kh5 79.Kf5 Rg2 80.Qd4 64.Qh8 Re4 65.Qxh5 g4! 66.Qf7+ Kg3! 67.Kd5 Rf4 68.Qc7 Kf3! 69.Ke5 Kg3! and Black holds.

So does this mean I missed a study-like draw? No. The position after 58...Kg4 is a good training position. What is White's best move?

Instead of entering the labyrith above with 59. e6? White has the killing 59. Ke3! threatening both d8Q and e6. 58...Rd1 trying to steer back into similar waters with 58...Kg3 fails most simply to 59. Ke2 and the Black rook is denied the d1 and e1 squares 59.e6 Kg4 60.e7 Re1+ 61.Kd3 Rxe7 62.d8Q and Black doesn't get to pick up the f-pawn.

As further illustration how difficult this type of position can be for even the best players in the world, I want to present the game between Boris Gelfand and Vassily Ivanchuk from the 1995 Belgrade Investbank tournament. Both of these super-GMs were in the world top 10 at the time. Gelfand provided notes to the game in Informant 65. From his analysis, we will see that he did not even have access to 5-piece tablebases, and we can see the difficulty presented in analyzing these positions without a computer let alone in playing them over the board.

Some background knowledge on Q vs. R+P that is relavent to this ending. If the pawn is advanced beyond the second rank the defense becomes more difficult. Most of the positions with a bishops pawn or central pawn are lost because there is enough room on either side of the pawn attacking king and queen to squeeze the defense from both sides.

A knight's pawn offers more possibilities for the defense. If the defending king is behind the pawn, the positions are almost always drawn. There are two exceptions: 1. If the attacking king can get all the way behind the defending king crossing both rank and file. 2. The rook gets pinned as we saw in Carlsen-Topalov.

If the king is in front of the pawn, the situation is less clear cut. The primary goal for the defense is to cut the king off along a file as we saw in the analysis of the Colias game, Black can hold with a rook on the f-file cutting off the White king. However, with a pawn on the third rank attacker can work his way in with the king. On the fourth rank or beyond, the rook has extra space to stay safe from forks while maintaining the barrier along the file. The following two positions illustrate the differences.

Here, Black loses. 1. Qh3 Rf4 2. Qg3+ and the king crosses over after either 2...Kf5 3. Kf7 or 2...Rg4 3. Qh5+ and 4. Kf7

With everything advanced one rank, Black holds because his rook has the safe square f8

1. Qh2 Rf8! and White can't make progress even if it was Black to move here, the rook can return to f4

Other good rules of thumb for the defense are to keep the king in close contact with the pawn and avoid getting the rook pinned. All of these are good guideposts to know, but the defense is very fragile and we will see that even a great player like Ivanchuk was unable to cope with the problems that Gelfand set to him.

The queen versus rook and two pawns ending was reached after 62. Qxh8

In the notes below, I've used the Nunn convention for the symbols on the moves rather than those provided in Gelfand's notes. 62...Kg5! Black would prefer to advance his pawn, but that would allow the king to cross over the f-file. [Gelfand gives 62...g5? 63.Kf2 (63.Ke2 Kg3=) 63...Ke4+ (63...g4) 64.Ke2 (64.Kg2 Re5 given as equal by Gelfand is also winning for White.) 64...Re5 He ended the line here with no evaluation, but I presume he felt it was a draw like the other evaluations he gave in this variation. The fastest win for Whtie is 65.Qc8 stopping both pawns from advancing.; 62...Kg4? is also instructive since Black doesn't block the advance of his g-pawn. 63.Qh6! g5 64.Qe6! the pinning theme 64...c4 (64...Kf4 65.Kf2 crosses the king over) 65.Ke2 (65.Qxc4+? Rf4 allows Black to set up the barrier) 65...c3 (65...Kf4 66.Qxc4+ Kg3 67.Qd3+ Kg4 68.Qd7! Kf4 69.Kf2+-) 66.Kd3 c2 (66...Kf4 67.Qe4+) 67.Kxc2 Kf4 68.Kd3 Re5 (68...g4 69.Qe3#) 69.Qf6+ Rf5 70.Qd6+ Kf3 71.Qe6 Rf4 it looks like Black has reached his desired setup, but he needs the king on g3 to prevent White's next. 72.Qh3+! and the king gets separated from the pawn. 72...Kf2 73.Qe3+ Kg2 74.Ke2! Kh2 75.Qe5 winning the pawn] 63.Ke2 Rf7 64.Kd3 Rf3+ 65.Kc4

65...Rf5? This move went uncommented by Gelfand, but that doesn't surprise me. What could be more natural than to guard the c-pawn and put the rook on an anchor square? [Informant doesn't have a symbol for mind blowing, but giving up the pawn with 65...Rf4+ 66.Kxc5 Kg4! secures a draw as the pawn will be able to advance.] 66.Qg7 Kh5 67.Qh7+! Kg5 68.Qh8? That was a nice bit of triangulation, but it lets Black off the hook [68.Kb5 Kf6 69.Kc6 c4 70.Kd6 Rf4 71.Qh6 Kf5 72.Ke7 c3 73.Qh3+ Kg5 74.Qg3+ (74.Qxc3? Kg4!) 74...Kf5 (74...Rg4) 75.Kf7] 68...Kg4 69.Qh6 g5 normally it is undesirable to allow the following pin, but here the c5 pawn helps the defense by robbing White of the d4 square. 70.Qe6 Kf4! 71.Kd3 Re5! [71...g4? 72.Qe3#] 72.Qf6+ Rf5! 73.Qd6+ Re5 [Gelfand gave 73...Kf3 74.Qe6 Rf4? (the narrow path to the draw is 74...c4+! 75.Kd2 taking allows 75...Rf4 and 76...Kg3 reaching a draw. 75...Rf4 76.Qh3+ Ke4! 77.Qg3 g4 78.Qh4 c3+ 79.Kxc3 g3 80.Qh3 Rf3+ 81.Kd2 Rf2+ 82.Ke1 Kf3 after move) 75.Qh3+! Kf2 76.Qh5 c4+ 77.Kc3? (77.Kd2! g4 (77...c3+ 78.Kxc3 g4 79.Qh2+ Kf3 80.Kd2 is also winning) 78.Qh2+! Kf3 79.Qe2+! Kg3 80.Qe5+- is the familiar pin) 77...g4!= I can understand Ivanchuk's reluctance to go for this as the timing of the push of the c-pawn needs to be very precise.] 74.Qf6+ Rf5! 75.Qh6 Re5 [Gelfand gives 75...Rd5+ 76.Ke2 as interesting without further evaluation. Then, the only move to hold is 76...Rd4! which leads to an entirely different class of position than we have been considering. 77.Qh3 Preventing Black from playing Kf4 or Kg4 followed by Rf4 (The white king can cross the f-file, but it doesn't help. 77.Kf2 Rd2+ 78.Kg1 Rd1+ 79.Kg2 Rd2+ 80.Kh3 Rd3+) 77...Ke5!; Gelfand showed 75...Kf3? losing to the pin after 76.Qh3+ Kf4 77.Qe3+ Kg4 78.Qe6+-; and showed 75...Kg3 leads to a draw, which was probably Black's simplest method.] 76.Qh2+ Kf5! 77.Qf2+ Ke6? Gelfand correctly identified this as the losing move. [giving 77...Kg4 78.Qf6 Rd5+ 79.Ke3 Rd4!= which is similar to the lines after 75...Rd5+] 78.Qf3 Kd6 79.Kc4 Ke6 80.Kb5 Rf5 81.Qh3 Kf6 [81...Ke5 82.Kxc5 Kf4+ 83.Kd4+- Gelfand (83.Kd6 is a longer, but perhaps more thematic way to win, see the note to Black's 82nd move.) ] 82.Qh6+ Kf7 Gelfand devotes the largest part of his analysis to the lines following 82...Ke5 83.Kxc5

a) 83...Kf4+ 84.Kd6 (84.Kd4? Gelfand ends his line here, perhaps evaluating it as a win based on the similar position in the note to the previous move. But here the White queen is not on h3, so Black has the saving 84...Kg3! 85.Ke3 Rf3+ 86.Ke2 g4=) 84...Kg4 85.Qe6! pinning! Here it is a little trickier with the king already crossed over the 5th rank, but here is a sample line 85...Kf4 86.Qd7 Kg4 87.Ke7 Kf4 88.Ke6 Kg4 89.Qd1+ Rf3 90.Ke5 Kg3 91.Qg1+ Kh4 92.Qh2+ Kg4 93.Qg2+ Rg3 94.Qe4+ Kh3 95.Kf5 crossing the f-file and winning b) Somewhat strangely, Gelfand does analyze the king going around after 83...Ke4+ 84.Kd6 (Here 84.Kc4 still wins for White since he can prevent the Black king from reaching g3 after 84...Kf4 85.Qg6 Kg4 86.Qe6! the pinning theme again.) 84...g4 85.Qg6 Kf4 86.Ke6 Re5+ 87.Kf6! Ra5 (87...Re3 88.Qf5+ Kg3 89.Kg5+-) 88.Qd3+-; c) 83...g4 84.Qg6 (Gelfand's 84.Qe3+? lets the Black king get back behind the pawn. 84...Kf6+ 85.Kd4 (85.Kd6 is also met by 85...Kg6!=) 85...Rf3? (85...Kg6!=) 86.Qe5+? (86.Qh6+! Kf5 87.Qh5+! Kf4 else the pawn is lost 88.Qe5#) 86...Kg6 87.Ke4 Kh6! holding the draw) ; 83.Kc4 White can't advance the king 83.Kc6?? Rf6+! but with the Black king cut off, the rook cannot maintain the defense of the pawns on its own. 83...Re5 84.Kd3 Ke7 84...Rf5 85.Ke4 Rf4+ 86.Ke5 85.Qg6 Rd5+ 86.Ke4 Rd4+ 87.Ke5 Kd7 88.Qb6 Rf4 89.Qe6+ Kc7 90.Qd6+ Kc8 91.Qxc5+ Kd7 92.Qd6+ Kc8 93.Qg6 1-0

So what lessons do we draw from this ocean of subtle variations? 1. I think first off, the attacking side in these endings needs to be like Magnus and not believe in fortresses. While fortresses certainly exist, they can't always be played mindlessly. Especially, deep in the endgame, fatigue can start to weigh on the defender. 2. The age old endgame theme of "Don't Rush!" In many of the variations of the Colias game, we saw that there was a narrow line to a draw after 58...Kg4, but if White looks to first kill Black's counterplay with 59. Ke3! he avoids a possible draw with a much clearer variation. 3. In a queen versus rook and knight's pawn ending, the mobility of the rook is key. We saw numerous variations above where the queen pinned the rook, which stopped Black from reaching an ideal defensive setup.

5/6/18 - Bereolos-Kudrin, 1986 Midwest Masters Invitational

I've added my first game against Sergey Kudrin to the GM games section. This was the start of a so far unsuccessful duel with Sergey in the exchange Grunfeld (+0 -4 =0 none of which were due to the opening). I had a very nice position, but overplayed it and let him equalize. I then sacrificed a piece to open his king, which should have led to a draw, but instead I blinked at the wrong moment and lost.

This was the 4th edition of the Midwest Masters, which eventually evolved into the US Masters. Sergey went on to tie for first with Joel Benjamin and Yehuda Grunfeld at 4.5/5. I ended up with 1 win, 2 draws, and 2 losses.

5/1/18 - Bereolos-Schmidt, 1992 Indiana State Team Championship

I never gave the queen versus rook ending of my game against Ulrich Schmidt in the 1992 Indiana State Team Championship a very deep look. I knew he could have established a simple fortress, but instead stumbled into a pin and lost his rook. It turns out that there was more subtlety to it than that and I only capitalized on his second blunder. We entereded the queen versus rook ending after 45.Qxc1

Black has a fortress here by moving his rook between g6 and e6. If White manages to put his queen on the long diagonal, the king can go to h7 or g8. 45...Rf4 46.Qc7 Rf5 47.g3 Kg6? There was still time to set up the fortress with47 47...Rg5 and 48....Rg6 Instead, the king comes forward like a goalie leaving the net 48.Kg2? Letting him off the hook, 48.Qe7 is decisive 48...Kg7 (48...Rf6 49.Qxe5) 49.g4 Rf6 50.Qxe5 and because of the pin, Black can't reestablish the fortress. A sample line to show how White worms his way in is 50...Kg6 51.Kg2 Re6 52.Qh8! It is vital to stop the Black king from getting back via h7 or g7. 52...Re4 53.Kg3 Re6 54.Kh4 Re4 55.Qg8+ Kf6 56.Kh5 Re5+ 57.g5+ Ke7 58.Kh6 Re6+ 59.Kh7 Rd6 (59...Rg6 60.Qxg6) 60.Qg7 Re6 61.Kg8 Rg6 62.Qxg6 fxg6 63.Kg7 with a winning pawn ending. 48...f6? He could have corrected things with 48...Kg7! 49.Qc2 Kg5 50.Qe4 1-0

Keres analyzed a similar position from the game Samisch-Prins, with the Black king on e7 so the Black rook can cut the king off from crossing over.

White still wins here by transposing to a winning pawn ending at the appropriate moment. 1. Qh4+ Kf8 2. Qh8+ ke7 3. Kf2 Rg6 4. Kf3 Re6 5. Kg4 Rg6+ 6. Kh5 Re6 7. g4 Rg6 8. g5 Re6 9. Qb8 Rg6 10. Qb4+ Ke8 11. Qe4+ Re6 12. Qxe6+ fxe6 13. Kh6 and wins. I wanted to look at how the game actually continued to pinpoint where where Samisch went wrong. However, the game was not in my database nor in several other online databases that I checked. I thought this was a bit strange since both players are of some reknown having opening variations named after them. This led me on a bit of a research study.

I found two references to this game in books Keres authored ( Practical Chess Endings and The Art of the Middle Game) and attributed to Keres in three others ( Dvoretsky's Endgame Manual, Encyclopedia of Chess Endings and the 3rd volume of Averbakh's Comprehensive Chess Endings). The location was referred to as Hastings 1938 or 1938/39 (the famous Hastings tournament spans the new year so it often gets referenced in different ways). The Keres books give a story that Samisch played inaccurately and allowed the Black King to come to g7 then the game ended in a draw, where Averbakh and Dvoretsky have the draw agreed from the diagram. Further research seemed to indicate that neither player was in the 1938/39 edition of Hastings. Keres did play in the 1937/38 event, so I thought he could have got the date wrong, but I didn't find Prins playing at Hastings until 1945/46 and 1946/47 while I didn't find any time where Samisch played Hastings.

Finally, a found a snippet from the 1939 volume of the British Chess Magazine on Google Books saying that both Samisch and Prins would be playing in the Premier Reserves. So at least I partially solved the mystery. It looks like the games from this event didn't make it into the databases, which is again surprising since Mieses is another notable name in that field. That is where the trail finally went cold for me, although I also turned up an interesting article on the Google Books scanning project.

After that bit of diversion, back to the chess. There are two keys to this fortress are the position of the defending king and rook. In the Samisch-Prins game, Black was unable to get his king in the proper position in front of the White pawn. In my game, Black was unable to establish his rook on an anchor point. These two features are well illustrated by moving the White pawn from g3 to e3 in the Samisch-Prins game.

White to move wins instructively with 1.Qh4+! Ke8 1...Kf8 2.Qd8+ or 1...Rf6 2. Kg2 Ke6 3. Qb4 both force the Black king to abandon e7/e8. 2. Qa4 and the pin decisively compromises the defense.

Instead, Black to move from the diagram holds with 1...Rg6+ establishing the rook on an anchor point. However, Black cannot play completely on autopilot from here. For example, 2.Kf2 Re6 3.Qh4+ Ke8 4.e4 Rg6 5.Ke3 Re6 6.Qh8+ Ke7 7.e5 Rg6 8.Ke4 Re6 9.Kf5 Rg6 10.Qc8 Black must avoid 10...Re6? 11.Qxe6+ fxe6+ 12.Kg6 instead 10...Rh6! and White cannot take advantage of the rook leaving an anchor point.

Finally, in positions with a defending g-pawn versus and attacking f-pawn are always won for the side with the queen because of zugzwang. A very timely example of this is the game a few days ago between World Champion Magnus Carlsen against Veselin Topalov in the Gashimov Memorial tournament.

51...Rf6 [White even wins without his pawn after 51...Rxf5 52.Qe6+! Rf7 again the pin is key to breaking down the fortress. 53.Kxg4! Kf8 54.Qc8+ Ke7 55.Kg5 Rf6 56.Qg8! and the Black pawn falls.] 52.Qe8+ [White can even allow the Black king to the f-file in this situation with 52.Qc8+ but with everything shifted one file to the right, the equivalent of Carlsen's move Qb6-d8+! would be the only way to win. 52...Kf7 53.Kxg4 Rh6 54.Kg5 Rf6 (Black is in zugzwang without the equivalent of the saving 10...Rh6! from the previous example) 55.Qd7+ Kf8 56.Qd8+ Kf7 57.Qxf6+ +-] 52...Kh7 53.Kxg4 Rh6 54.Kf4 Rf6 55.Ke5 Rh6 56.Qe6 1-0

Some lessons from these endings. 1. Don't relax in a drawn ending, you're opponent may give you an opportunity (47...Kg6? in my game or 10...Re6? in the altered Samisch-Prins analysis). 2. You have to be able to recognize endgame themes in different situations. The presence of the black e-pawn and white g-pawn in my game didn't alter a textbook fortress. 3. In trying to win with queen (+pawn) against rook+pawn, two key methods are pinning the rook to the king to break the fortress (50. Qxe5 in the notes to my game, 2. Qa4 in the altered Samisch-Prins position, and 52. Qe6+ in the Carlsen-Topalov game) and transposition to a winning pawn ending at an appropriate moment (62. Qxg6 in the notes to my game, 12. Qxe6+ in Keres analysis, 11. Qxe6+ in the modified Samisch-Prins position, and 57. Qxf6+ in the notes to Carlsen-Topalov)