Where to move the rook? My first inclination was to move it far down the board, to e1 or e2. But at second blush those squares seem inferior because after Black plays ...f3 the rook may hang to the bishop on one of its checking squares, e.g. 1...Re1 2.Bxg5 Rd1+ 3.Kc6 Rc1+ 4.Kb7 f3? 5.Bxc1 or 1...Re2 2.Bxg5 Rd2+ 3.Kc6 f3? 4.Bxd2. Maybe the latter line is still winning for Black after 3...Rc2+ 4.Kb7 f3 5.Be3. However, 1...Re3! looks more accurate since now 2.Bxg5 is out of the question: 2...Rd3+ 3.Kc6 f3 0-1, since 4.Be3 is met by simply 4...Rxe3.

Even in the final position(-1) Black comes up with the nice check where any move by the White King allows the capture of the Bishop, but even that was not needed. Simply advance the a-pawn. The Bishop cannot stop both pawns. The King is too far away to help. I saw people talking about Rxc7...why? Make White waste a tempo pushing the pawn to c8 before taking it.

The pawns on a4 and f3 are guarantee a Queen is coming Black's way. Finding the cute Rook check is nice and makes Re3 the best of the best but other lines are plenty good enough as well.

In the Re2 line was there any discussion of fxg3 in the position with the pawn on c7 and the King on c8? It appears to win for Black as well...at least Bxd2 gxh2 has Black Queening first with both Rook pawns left on the board.

So after gxh2 which way does White's King go? In this position all Black needs to do is force the Queen trade and with the position of White's King once again one of the two Rook pawns will Queen.

Kb7 loses immediately to h1Q+.

38. rxe6 Kxe6
39. Kxc8

Hi, <David2009>. Your end-game trainer has been very useful to me (and others), so thanks for your welcome.

< <agb2002> wrote: <johnlspouge> End of sabbatical? >

Hola, <Antonio>. I am on vacation and accordingly, able to post earlier (as you must have noticed). I am under scientific review next year, with one student's project lagging a little, so my return is only temporary, about another week.

If you want to simulate real game situations then it can be a useful exercise to set yourself a time limit which mirrors the amount of time you might have in a game.

I play most of my chess at 90 minutes for the entire game, so I can rarely afford too many thinks of more than 5 to 10 minutes. So I mostly look to solve POTDs in under five minutes.

If you play that way, you are not looking for perfection in your analysis, just a pragmatic decision about which move to play. Judgement can replace analysis in order to save time.

On the other hand, if you want to focus on accuracy and board vision, it can be useful not to have a time limit. Then you strive for a thorough and detailed analysis, where judgement plays a lot less of a part.

In Think like a Grandmaster, Kotov describes how he trained himself to analyse by looking at the most complicated positions for between half an hour and an hour.

What I will sometimes do is to spend five to ten minutes in initial human mode. Then look up the solution, read the kibitzing, write a story. Then go back over any bits I didn't understand without a time limit.

And sometimes, like yesterday's POTD, I use <David2009>'s crafty trainer to see how easy or hard it is to win a supposedly won position.

click for larger view

plus a Crafty End Game Trainer link to play the ending out interactively: http://www.chessvideos.tv/endgame-t.... You are white, drag and drop the move you want to make.

48 Rxh5? drops the Bishop after which Black should win. However, 48.Rh8+ Kf7 49.Rxe8 Kxe8 50 Kb5 appears to draw with the Bishops of opposite colours, although Black can give White a hard time: 50...Kd7 51 Bf6 f4 52 Bxg4 f3 54.Be3 Kd5 55.Kb6 Ke4 56.Bg1 Kd3 to reach

click for larger view

I wasnt sure between e4 and e3 but felt both should work. Re4 might not prevent a later Bd3 in some lines, but I dont see how that matters. Unless I missed a clever finesse by W the WB cannot stop both the B passed pawns.

While some have said that they don't like the links, I think they are great. It does not matter if you are a tyro or an accomplished player, my belief is that everyone should practice their technique. (If they want to remain sharp - and I do.)

http://www.chessvideos.tv/endgame-t...

click for larger view

1.Re6, gxf5;
2.gxf5, Kd3;
3.Rc6, Bh5;
4.Kg3, Bf7;

Black wants to play ...Bc4; building a bridge and winning the game. (This explains the next move.)

5.RxP/c2, KxR/c2;
6.Kf4, Kd3;
7.a5, Kd4;
8.a6, h5!?; (Hmm. Horizon effect?)
9.a7, Bd5;
10.f6, Kc5;
11.Ke5!, (zugzwang) 11...Bb7;
12.f7, " " and White wins.

Sunday - 30 minutes is an absolute minimum.

I DO think that Kotov was right. I.e., that essentially the more strenuous your training regimen is; and the more it SIMULATES the actual conditions of a real tournament game, the more likely you are to improve. (Or at least not lose a step.)

<Result: Succeeded with lucky guess.>

White goes downhill fast,

52. Kd6, -9.66. Best, c8N, -8.41.

53.Bxg5, -10.73. Best, gxf4, -9.66.

54.Be3, -#14. Best, Bd2, -10.79.

and resigned after 54...Rd3+, -#14.

Both sides made errors earlier but the rot set in for White with,

Black has two potential passers on a5 and f5, but must first find a way to take care to defend against the eminent promotion of White's c-pawn to c8.

Obviously, a Rook retreat is called for to deflect the White King and capture the white c-file passer after it promotes.

Three Rook moves appear to win for Black, with 51...Re3! being the strongest and 51...Re4! being a close second. Also winning is 51...Re1, but the win is a bit trickier for Black and offers him a few chances to go wrong and let White back in the game.

Losing for Black is 51...Re5?? The move 51...Re2?! is interesting but dubious as it is give Black practical drawing chances.

Here's some analysis, starting with the losing move for Black:

<50. Bf6 ?> This is the losing move that sets up Black's winning combination.

Instead, 50. Bd6! holds when play might continue 50...a4 51. c8=Q Rxc8 52. Kxc8 Ke6 53. Kc7 f4 54. gxf4 gxf4 55. Bxf4 a3 56. Bd2 a2 57. Bc3 Kf5 58. Kd6 h4 59. Kd5 Kg4 60. Ke4 Kh3 61. Be5 a1=Q 62. Bxa1 Kxh2 63. Be5+ =.

<50...f4!> Now Black has a win with this somewhat obvious push of the passed f pawn.

<51. Bd8>

Black has nothing better. If 51. gxf4 gxf4 52. Bd8 Re3 53. c8=Q Rd3+ 54. Kc6 Rc3+ 55. Kd5 Rxc8 56. Bxa5 f3 57. Be1 Re8 58. Bh4 Re2 59. h3 f2 60. Bxf2 Rxf2 wins easily.

<51... Re3!> This move solves our Friday puzzle, and is the strongest of at least three winning Rook retreats.

A second winning option is my attempted solution 51... Re1 52. Bxg5 Rd1+! (not 52... f3 ?? 53. c8=Q ), when play might continue 53. Kc6 (53. Kc8 f3 54. Kb8 f2 55. c8Q f1Q ) 52...Rc1+! 54. Kb5 (54.Kb7 Rxc7+ 55. Kxc7 f3 56. Be3 a4 57. Bc5 a3 58. Bxa3 f2 ) 54... Rxc7 55. Bxf4 Rc2 56. h4 Ra2 57. Be3 Ke6 58. Bd4 Kf5 59. Bg7 a4 60. Bh6 Kg4 61. Bf4 a3 62. Ka4 Rg2 63. Bg5 Rxg3 64. Bc1 a2 65. Bb2 Kxh4 66. Be5 Kg4

A third winning option is 51...Re4 when play might continue 52. Kd6 Rc4 53. Kd5 Rc2 54. Ke4 a4 55. Bxg5 fxg3 56. hxg3 Rxc7 57. Kd3 a3 58. Bf4 Rc5 59. Kd4 a2 .

Giving White practical drawing chances is 51...Re2?! when 52. Bxg5 Rd2+ 53. Kc8 is far from a clear win for White.

Turning the tables and losing for Black is 51...Re5?? 52. c8Q .

<52. Kd6>

Also losing quickly for White is 52.
Bxg5 Rd3+ 53. Kc6 f3 ,

52. gxf4 gxf4 53. c8=Q Rd3+ 54. Kc6 Rc3+ 55. Kb7 Rxc8 56. Kxc8 a4 ,

and 52. c8Q Rd3+ 53. Kc6 Rc3+ 54. Kb5 Rxc8 .

<52... Rc3 53. Bxg5 f3 54. Be3>

If 54. Bd2, then simply Rxc7 55. Kxc7 f2 56. Bxa5 f1=Q is sufficient for an easy Black win.

<54... Rd3+ 0-1> This little in between move check forces immediate resignation, but is not essential as there are several winning alternatives.

51... Re4!! appears to be the most accurate move, ensuring that black can neutralize the c-pawn (or short-lived new queen) *and* maintain two winning passed pawns that beat the bishop:

A) 52.c8=Q Rd4+ 53.Kc7/c6 Rc4+ 54.Kb7 Rxc8 55.Kxc8 (Bxg5 fg wins) a4 56.Bxg6 f3 57.Be3 a3 wins, the bishop being unable to stop both the a-pawn and f-pawn.

A.1) 55.Bxa5 fg 56.hg Rc2 is easy.

B) 52.Bxg5 f3! 53.c8=Q Rd4+ 54.Kc6/c7 Rc3+ 55.Kb7 Rxc8 56.Kxc8 f2 wins.

C) 53.Bf6! Rc4! (Kxf6?? 54.c8=Q Rd3+ 55.Ke8) 54.fg g5 and again the black passed pawns win.