Latest Updates:
Hot Topic (More than 10 Replies) Tattersall A Thousand Chess Endings (Read 5642 times)
an ordinary chessplayer
God Member
*****
Offline


I used to be not bad.

Posts: 1667
Location: Columbus, OH (USA)
Joined: 01/02/15
Re: Tattersall A Thousand Chess Endings
Reply #11 - 05/21/19 at 04:10:06
Post Tools
When I posted the Lasker and Reichhelm position yesterday, I should have mentioned that it is also No. 65 in Tattersall; but I didn't remember (I created the pgn too many years ago), Tattersall does not have an index, and I didn't find it by searching for the fen string (Tattersall's position reverses the K- and Q-sides). I guess it's time to learn CQL, which would have found it using the -flipvertical option.
http://www.gadycosteff.com/cql/

So far out of five Tattersall compositions, we have two correct studies (36, 61), two duals (13, 62), and one outright cook (58). Those statistics are not too impressive, but it might just be bad luck, due to the small sample size. The remaining four are sound, and the last of them (a Queen ending) is pretty nice in my estimation. This next one might seem like a "basic" ending, in fact I wouldn't call it a "study" by modern standards, but Tattersall put his name on it, and maybe for the day it qualified. You can certainly find it in the online tablebase. But have a go anyway.
http://www.k4it.de/index.php?lang=en&topic=egtb

No. 70 C. E. C. Tattersall
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
*
White to play and win.
  

Tattersall-0070_001.pgn ( 0 KB | 153 Downloads )
Back to top
 
IP Logged
 
an ordinary chessplayer
God Member
*****
Offline


I used to be not bad.

Posts: 1667
Location: Columbus, OH (USA)
Joined: 01/02/15
Re: Tattersall A Thousand Chess Endings
Reply #10 - 05/19/19 at 21:53:22
Post Tools
One more corresponding squares digression before I get back to Tattersall.

Fine (1941) Basic Chess Endings, pages 53-54. (Lasker and Reichhelm, 1901)

No. 70
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
*
White to play wins; Black to move draws.

Fine gives the correct solution, but he makes a mistake in a sideline he gives. The error is discussed in Whitaker / Hartleb (1960) 365 Ausgewahlte Endspiele / 365 Selected endings, diagram "C". Notably, it was not covered in Chess Life and Review, so does not appear in the corrections by Louie, or those by Crane / Chew.

This endgame is a pretty good illustration of my geographical method.
  • North pole h4/g6, South pole c4/b6. These are corresponding squares.
  • Longitudes h4-g3-f3(f2)-e3(e2)-d3-c4 and g6-f7-e7(e8)-d7(d8)-c7-b6. Also corresponding squares.
  • Equator d3|e3 vs d7|e7.
  • Maneuvering space - black has enough. Unlike in Tattersall No.62, black has two ranks near the equator, so triangulation won't work here. Note that there is not an absolute 1:1 correspondence in the middle. Black has a choice. So if white is on e3, black can be on d7 or d8. The correspondence is more on files than on squares. Reuben Fine makes this point in his analysis as well.
Edited:
Changed one instance of Not to Note.

Unlike in Tattersall No.66 (Locock), the play near the equator and towards the North pole is mostly rectangular. There will still be some triangular correspondence near the South pole, but the absolute number of correspondences is pretty small, so can be kept in the head.

Now we begin enumerating the correspondences.
  • c4/b6 + d3/c7 implies c3/b7.
  • c4/b6 + c3/b7 implies b3/a7.
    -- or maybe c7! This latter square illustrates what I call "folding", of which distant opposition is a well-known case. Let's ignore c7.
  • d3/c7 + c3/b7 +b3/a7 implies c2/b8.
    By now we can see the pattern. Black needs to be one file to the left, and on the opposite color. This makes sense because of both h4/g6 and c4/b6. The pattern will help us remember, but I still want to calculate the next square from known squares instead of from the pattern. It's the only correct way, because the pattern will break down when we reach the edge of the board.
  • c3/b7 + b3/a7 + c2/b8 implies b2/a8.
  • d3/c7 + c3/b7 + c2/b8 implies d2/c8.
  • d2/c8 + c2/b8 + b2/a8 implies c1/b7.
    "Folding" in action again. Black can't use the 9th rank, so switches to the 7th rank. Note the color pattern still holds.
  • b2/a8 + c2/b8 + c1/b7 implies b1/a7.
    Bad news for black. The white problem is solved already (Kb1!), but let's continue, so we can solve the "Black to move draws" part.
  • b3/a7 + b2/a8 + b1/a7 doesn't imply anything yet for a2.
    Black has to use folding on the b-file, but it could be b7 or b8. So it's no use trying to get a corresponding square for a1, it would also be either b7 or b8. Fine gives for black 1...Kb7!!, but based on this last point I don't see why 1...Kb8 would not draw as well. 2.Ka2 (2.Kb1 Ka7) (2.Kb2 Ka8) 2...Kb7 3.Ka3 Kb8 etc.

Note the method for building the correspondences. We generally need three squares to build a fourth, although in the neighborhood of the poles we might need only two because of the nearby pawns. This makes it difficult to get started. That's why I like the geographic method, which gives a nice set of corresponding squares right away (poles and longitudes), as well as the possibility to skip the whole thing if we notice a chance for triangulation. We still have to do some work with true correspondences, but we only need to keep in our head 7-9 of them, which is quite do-able for an ordinary short term memory.

One final point concerns how to make progress. Normally when the kings are on corresponding squares and black moves, the white king will step onto a corresponding square again. But if white does this perpetually, it might be a draw. The correct method is at some point to step in the opposite direction of the black king, then when black reacts, step onto a corresponding square that is closer to the target. This cut-back move by white happens at the equator. A trivial example is this one.

* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
*
1...Kb5 2.Kb3 Kc5 3.Kc3 Kd5 4.Kd3 Ke5 5.Ke3 Kf5 6.Kf3 Kg5 7.Kg3 (equator) 7...Kh5 8.Kf4 (not 8.Kh3) 8...Kg6 9.Kg4 etc.
  

FineBCE-070.pgn ( 0 KB | 130 Downloads )
Back to top
 
IP Logged
 
an ordinary chessplayer
God Member
*****
Offline


I used to be not bad.

Posts: 1667
Location: Columbus, OH (USA)
Joined: 01/02/15
Re: Tattersall A Thousand Chess Endings
Reply #9 - 05/19/19 at 16:58:56
Post Tools
Jupp53 wrote on 05/19/19 at 12:53:37:
The normal method is so hard for most because it is mathematical. Mathematics are no language for most humans. Chess like everything demanding conscious cognitive work is best stored and processed with language.
For a long time I have strongly believed that chess is best stored with patterns and processed with moves. But "most people" either can't do that or don't want to do that. Adults in particular try to substitute natural language thinking for the required pure chess thinking, which leads to blunders.

Although pretty strong in mathematics myself, I also find the corresponding squares problems difficult.
  1. The types of patterns usually associated with chess (pawn structure, tactical operations, etc.) are no use here.
  2. The actual mathematics required is not something we use much in everyday life -- just a long chain of transitive relationships.
  3. The number of relationships is far greater than the limits of short term memory.

I think the last one is the biggest reason I can't do it. Someone like Pillsbury or Koltanowsky with their phenomenal short term memory could do it, I bet. So I came up with my geographical method to help out. I find that I can focus on one area of the chessboard, solve it, then "forget" that part while I work on another part. Later when I return to the first part, it is easier to reconstruct what I did before because of the geographic imagery.

Usually. But sometimes my method is no use whatsoever. Here is the examplar from Tattersall, the best corresponding squares puzzle I have seen.

No. 66 C. D. Locock
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
* * * * * * * *
*
White to play and win.

The north pole (e3/g5) is too close to the south pole (d4/f6), the longitudes are too short (e3-d3-g5 and f6-g6-g5), and my geography doesn't help at all, therefore I can't afford to forget any of the relationships. So it all comes down to memory, but this composition has a particularly evil character: All the middle relationships are triangular (c3-c2-d2 vs g7-h7-f6) and the edge relationships are rectangular (b4-b3-b2 vs f8-g8-h8), which is a difficult transition in itself. But black's shortage of squares means that some of black's "triangular" squares map to some of white's "rectangular" squares. Finally, it's pure correspondence. In some puzzles, once you reach a win, you can just execute a plan. In this one, the only plan is to keep moving to the next corresponding square. Even though I have seen the solution, I still can't reproduce it in memory. It's hard enough using paper and pencil.
  
Back to top
 
IP Logged
 
Jupp53
YaBB Moderator
*****
Offline


be

Posts: 988
Location: Frankfurt/Main
Joined: 01/04/09
Gender: Male
Re: Tattersall A Thousand Chess Endings
Reply #8 - 05/19/19 at 12:53:37
Post Tools
an ordinary chessplayer wrote on 05/19/19 at 10:35:16:
Leon_Trotsky was asking about corresponding squares in another thread
https://www.chesspub.com/cgi-bin/chess/YaBB.pl?num=1554857629. So I will use this example to explain my personal thinking method around these endgames. My whole idea is to be able to solve these over the board (i.e. without marking the corresponding squares on a piece of paper). To help with this, I think of the chessboard as a globe. (If you find my method idiosyncratic, remember that it is personal. Feel free to come up with your own method. Anything that helps you get to the correct solution is fine.)

Let's start from this analysis diagram:

First I label the North and South poles: f5/f7 is the North pole, b5/b7 is the South pole. These are actually the first corresponding squares. If black cannot reach these corresponding squares in time, the white king gets in.
  • Next I find the lines of longitude for each player. b5-c4-d3-e4-f5 for white, b7-c7(c8)-d8-e7(e8)-f7 for black. Obviously if black's route is longer than white's, black will just lose to a king-march by white (assuming no counterplay). Otherwise, these should all be corresponding squares as well.
  • Then I identify the equator, which is the mid-point of the longitudes. In the diagram, the kings stand on the equator. I find it helps to mentally place the kings here before doing any calculations. Your mileage may vary.
  • The last thing I look for is defensive maneuvering space. These are different ranks (usually ranks) that black has available on the longitude. Here c7(c8) and e7(e8) are examples. Often if black has a 3x2 zone of ranks and files to maneuver in, then once the king reaches this zone, white cannot win. But in this example, black's d7 pawn prevents this zone.

  • Thank you! If I will use this for my blog  I will give the source. The main point: The normal method is so hard for most because it is mathematical. Mathematics are no language for most humans. Chess like everything demanding conscious cognitive work is best stored and processed with language. Your method gives the first linguistic approach I see.
      

    Medical textbooks say I should be dead since April 2002.
    Dum spiro spero. Smiley
    Narcissm is the humans primary disease.
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #7 - 05/19/19 at 10:35:16
    Post Tools
    No. 62 C. E. C. Tattersall
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White to play and win.

    Leon_Trotsky was asking about corresponding squares in another thread
    https://www.chesspub.com/cgi-bin/chess/YaBB.pl?num=1554857629. So I will use this example to explain my personal thinking method around these endgames. My whole idea is to be able to solve these over the board (i.e. without marking the corresponding squares on a piece of paper). To help with this, I think of the chessboard as a globe. (If you find my method idiosyncratic, remember that it is personal. Feel free to come up with your own method. Anything that helps you get to the correct solution is fine.)

    Let's start from this analysis diagram:
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    • First I label the North and South poles: f5/f7 is the North pole, b5/b7 is the South pole. These are actually the first corresponding squares. If black cannot reach these corresponding squares in time, the white king gets in.
    • Next I find the lines of longitude for each player. b5-c4-d3-e4-f5 for white, b7-c7(c8)-d8-e7(e8)-f7 for black. Obviously if black's route is longer than white's, black will just lose to a king-march by white (assuming no counterplay). Otherwise, these should all be corresponding squares as well.
    • Then I identify the equator, which is the mid-point of the longitudes. In the diagram, the kings stand on the equator. I find it helps to mentally place the kings here before doing any calculations. Your mileage may vary.
    • The last thing I look for is defensive maneuvering space. These are different ranks (usually ranks) that black has available on the longitude. Here c7(c8) and e7(e8) are examples. Often if black has a 3x2 zone of ranks and files to maneuver in, then once the king reaches this zone, white cannot win. But in this example, black's d7 pawn prevents this zone.

    So far this is all preliminary thinking, and at this point we would continue noting the coordinate squares, building them up logically around the longitudes above. But the virtue of this method is that most positions that come up in actual games are *not* won by coordinate squares, but by triangulation. In a true coordinate squares problem, once black achieves "opposition", white is unable to lose a tempo, whereas triangulation is synonymous with losing a tempo. Anyway, my preliminary steps above are partially to organize my thoughts for coordinate squares, but more importantly to look for a way to short-circuit the whole thing and win in an easier way.

    In my analysis diagram, black to move just loses. Black's king has to step to one side, and white's king goes the other way, when black can no longer reach the pole. White to move can exploit black's lack of maneuvering space to give the move to black.
    1.Kd2. Again, black has to choose a side.
    • 1...Kc8 2.Ke3 (threatening 3.Ke4, which black needs to answer with 3...Ke7, so) 2...Kd8 3.Kd3.
    • Or 1...Ke8 2.Kc3 Kd8 3.Kd3. Q.E.D.

    From the first diagram, Tattersall's variation with 3...Kc7 follows precisely this path. 1.d5! Kc7 2.d4 Kb7 3.Kc3 Kc7 4.Kd2! (his exclams) Kc8 5.Ke3 Kd8 6.Kd3 (see analysis diagram) Ke7 (6...Kc7 7.Ke4 Kd8 8.Kf5 Ke8 9.Kg6 Ke7 10.Kg7) 7.Kc4 Kd8 8.Kb5 Kc7 9.Ka6 Kc8 10.Kxb6

    I mentioned before "assuming no counterplay", and this is something that always has to be checked. Tattersall's early moves are designed to prevent black's counterattack with ...Ka6 and ...b5, and his play is effective. But the computer shows that white can allow this and still win.
    1.d5! Kc7 2.d4 Kb7 3.Kc3 Ka7 4.Kd2 Kb7 5.Ke3 Ka6 6.Ke4 (Tattersall gave 6.Kd3!) b5 7.axb5+ Kxb5 8.Kf5 Ka4 9.Kg6 Kxa3 10.Kxh6 a4 11.Kxg5 Kb2 12.h6 a3 13.h7 a2 14.h8=Q a1=Q 15.Qb8+ Kc1 16.Qxd6 with "+5" according to Shredder. The strange central pawn formation shields the white king from checks. I think even I could win that particular Q vs Q ending.

    So the computer noticed one dual, but when I looked at this years ago I had another one in mind.
    1.d5 Kc7 (Against 1...Ka6 white can play 2.d4 Kb7 3.Kd3 Ka6 4.Ke4 b5 with Shredder's variation above. After 1...Ka7 2.Kd4 Ka6 3.Ke4 b5 it's even worse, 4.axb5+ Kxb5 5.Kf5 Ka4 6.Kg6 Kxa3 7.Kxh6 a4 8.Kxg5 Kb3 9.h6 and with the d4 square empty black doesn't even get a queen) 2.Kd4! White's route c4-d4-e4 is just as short as c4-d3-e4, meanwhile white retains a spare tempo. 3...Kd8 3.Ke4 Ke7 4.Kf5 Kf7 5.d4 Kg7 6.Ke4 +-.
      

    Tattersall-0062.pgn ( 1 KB | 127 Downloads )
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #6 - 05/18/19 at 17:10:00
    Post Tools
    No 61 C. E. C. Tattersall (no date)
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White to play and win.

    Tattersall likes king and pawn endings. Only the last of his nine compositions has any pieces. Exactly what kinds of endings are in these 1000? I was curious, so I wrote a script to process the FENs and count the different piece configurations. It probably took me more time to format the table here on chesspub.
    #
    Side A Side B
    ====== =========================== =======================
    97
    Pawns
    27
    1 Knight --
    12
    1 Knight 1 Knight
    18
    2 Knights --
    3
    more Knights
    35
    1 Bishop --
    34
    1 Bishop 1 Knight
    40
    1 Bishop 1 Bishop
    68
    more Bishops etc.
    28
    1 Rook --
    45
    1 Rook 1 Rook
    7
    more Rooks only
    16
    1 Rook 1 Knight
    55
    1 Rook 1 Bishop
    15
    1 Rook 2 minor pieces
    23
    1 Rook + 1 Knight 1 Rook
    17
    1 Rook + 1 Bishop 1 Rook
    120
    more Rooks etc.
    16
    1 Queen --
    29
    1 Queen 1 Queen
    33
    1 Queen minor pieces
    19
    1 Queen 1 Rook
    3
    1 Queen 2 Rooks
    16
    1 Queen 1 Rook + 1 Knight
    14
    1 Queen 1 Rook + 1 Bishop
    6
    1 Queen
    more Rooks etc.
    22
    1 Queen + 1 Knight 1 Queen
    18
    1 Queen + 1 Knight 1 Queen + 1 Knight
    22
    1 Queen + 1 Bishop 1 Queen
    13
    1 Queen + 1 Bishop 1 Queen + 1 Knight
    7
    1 Queen + 1 Bishop 1 Queen + 1 Bishop
    8
    1 Queen + 1 Rook 1 Queen + 1 Knight
    1
    1 Queen + 1 Rook 1 Queen + 1 Bishop
    9
    1 Queen + 1 Rook 1 Queen + 1 Rook
    104
    more Queens etc.

    Edited:
    In the table, I changed two instances of "Knight" to "Bishop".
      

    Tattersall-0061.pgn ( 0 KB | 124 Downloads )
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #5 - 05/17/19 at 05:19:05
    Post Tools
    No. 58 C. E. C. Tattersall (no date)
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White to play and win.
    "Founded upon a position that occurred in actual play."

    Tattersall gives
    • 1.g3? f3 2.Kg1 Kxe7 3.Kf2 Ke6 4.Kxf3 Kxe5 5.h4 Ke6 6.Kf4 Kf6 =
    • 1.Kg1 Kxe7 2.Kf2 Ke6 3.Kf3 Kxe5 4.h4 Ke6 5.Kxf4 Kf6 6.g3 +-

    I bet the the 1.g3? variation is similar to what happened in the game.

    This is the position right before 6.g3.
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White has two ways to proceed.
    1. Hold back on g2-g3, advance the queenside pawns until they are blocked by black, and at the end play g2-g3, winning by zugwang.
    2. Play g2-g3, then use the theory of 3 vs 3 to run black out of pawn moves, finally winning by zugzwang.

    For the composer, these are equivalent, and by using method 2 he is in effect saying "I know this 3 vs 3 is a loss for the player who moves first." If the reader did not know this fact, then method 1 would muddy the waters, leaving the reader wondering if the pawns could have maneuvered differently. But for the practical player, these are *not* equivalent; because if white makes a mistake in method 1, white still wins, but if white makes a mistake in method 2, white *loses*. So my advice is, hold on to every spare tempo as long as possible. Use method 1.
    Edited:
    I wrote white *loses*. In this endgame, that's not correct. White can retreat without losing. But in the general case, zugzwang might cost a whole point.


    Anyway, after 6.g3 Tattersall analyzes 6... a6 7. c3 (notably not symmetrically with 7.a3; 7.a3 would be analogous to Fine's move in No. 96) and then he gives four tries for black. Having checked his variations with the computer, I can say they are "perfect". At no time does he allow black any chance. See the pgn for the details.

    This is the position right before 4...Ke6.
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    Black to move and draw. ("cook" by an ordinary chessplayer)

    ...Ke6 here is *completely* ridiculous. I guess the composer just "knows" that with black to move, the 3 vs 3 is going to run into zugzwang anyway. So like in the method 1/method 2 choice above, he chooses the most artistic way. But any normal person (not a composer), would at least run the queenside pawns first, get into the inevitable zugzwang, and *then* play ...Ke6. If there is nothing better. But of course there is something better.

    "?" means the evaluation changes. "!" means the only move to "preserve the tempo". If there is more than one such move, all are given.

    4...a5 = (4...c5? +-) 5.c4 (5.a4 =) 5...b6 6.b3! c6 7.a3! a4 8.bxa4 (8.b4? -+) 8...c5 9. a5! bxa5 10.a4! Kd4 = 11.Kxf4 Kxc4 12.Kxf5 Kb3 13.g4 hxg4 14. Kxg4 = {tablebase} There are more details in the pgn.

    The whole point is that in the 3 vs 3 duel, while black cannot choose to lose a tempo, black *can* choose the final pawn structure. By choosing a pawn structure with an *isolated* white pawn on c4, black gets a passed c-pawn in the quickest possible way. Actually, I initially thought black would win, but the computer set me straight. It's only -/+.
      

    Tattersall-0058.pgn ( 2 KB | 112 Downloads )
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #4 - 05/17/19 at 01:49:15
    Post Tools
    Excellent memory. I only looked in the book, but after your post I went to my other sources. I have three slightly different BCE corrections by Sam Louie (all dated 1990) and another version by Paul Crane / Rev. David Chew (no date, but my letter to Paul Crane was in 1988). Both collections say that 2...b6 is losing, and refer to the same issue and page that you cited. There are lists of "CL&R References to Basic Chess Endings" in Chess Life & Review: May 1979, page 285, and April 1974, page 273. The one from 1979 has the same citation.

    Quote:
    Reuben Fine wrote:
    It is of course impossible for anybody to know even a small part of what is contained in this book by heart. But the exact amount of specific knowledge is relatively unimportant; what counts is how well the principles are grasped. For this reason I have throughout tried to set up typical positions (for these are merely shorthand for general principles) and have always preferred helpful rules to mathematical exactitude. (my emphasis)
    --Basic Chess Endings (1941), page 572

    I agree with the typical positions part (although if we remember typical positions that has to count as specific knowledge, yes?). But the "helpful" rules are actually the opposite of helpful. The problem is that what is memorable is not the same as what is correct. Furthermore, what is memorable impinges on our decisions even if the rational mind knows it is not correct. This is why propaganda and slogans are effective -- rational people are somewhat affected even though they try not to be, and irrational people get quite deranged because they don't even have rationality as palliative.
      
    Back to top
     
    IP Logged
     
    kylemeister
    God Member
    *****
    Offline



    Posts: 4901
    Location: USA
    Joined: 10/24/05
    Re: Tattersall A Thousand Chess Endings
    Reply #3 - 05/16/19 at 06:27:41
    Post Tools
    The inaccuracy of that "by simply copying his opponent's moves" statement is something that stuck in my memory from a Larry Evans Q&A in Chess Life way back.  I see from Google Books that it was in 1976 -- vol. 31, p. 148.  (The Tsheshkovsky-Browne game which you said was in one of your first issues was vol. 31, p. 559.)

    I'm also reminded of an article the year before by Tim Krabbé:  On Tattersall No. 336 -- The Saavedra Position (vol. 30, p. 680).
      
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #2 - 05/16/19 at 04:55:14
    Post Tools
    Next up (tomorrow) is Tattersall No. 58, but while analyzing it I realized my understanding of 3P vs 3P was defective. So first a digression into Fine (1941) Basic Chess Endings. I also checked the new edition Fine / Benko (2003), but Benko did not making any changes to this particular example.

    Fine no.96
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *

    Quote:
    Reuben Fine wrote:

    Starting from the symmetrical position (No. 96) the person who does not move always retains an extra tempo, by simply copying his opponent's moves. (my emphasis)

    The highlighted part is what I remembered, and it was messing me up, because for the case of 3 vs 3 it is not true. There is no general rule of whether black should copy or not copy white's moves, and this is what I was finding as I analyzed Tattersall.

    In order to get a legal pgn, I created a zugzwang position with kings, and copied Fine's analysis to it. I also flipped a-h from Fine's diagram to make it easier to compare with Tattersall No.58, when we get to that.
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *

    Quote:
    Reuben Fine wrote:

    1. c3 c6 2. b3 b6 3. a3 a6 4. c4 a5!
    (4...c5? 5. b4! a5 6. bxa5 bxa5 7. a4)
    5. b4
    (5. a4 c5)
    (5. c5 bxc5 6. a4 c4! 7. bxc4 c5)
    5...a4 6. b5 c5
    or 6. c5 b5

    Note that on the fourth move, Fine has black answer 4.c4 with 4...a5! (his exclam), in parentheses showing the "copying" move as losing. This is clearly a case of "do as the GM does, not as the GM says".

    Quote:
    Reuben Fine wrote:

    From this last (sic) variation we can draw the all-important conclusion that an exchange loses the tempo. (his emphasis)

    As far as I can tell, this second statement is correct, as long as it is a simple exchange of PxP followed immediately by ...PxP. Fine's variation 1. c3 c6 2. b3 b6 3. a3 a6 4. c4 a5 5. c5 bxc5 6. a4 c4 7. bxc4 c5 shows that an "exchange" of pawns by sacrifice and counter-sacrifice does not change the tempo.

    Putting Fine's analysis in the computer was an eye-opener. I thought that black had more choices early, and that it was only later that the moves mattered. The computer showed just how critical each choice is. Even Fritz 6 on my consumer grade laptop is able to calculate to the end, or at least after a couple of pawns have been committed. The real shocker is that Fine's 1.c3 c6 2.b3 b6 actually loses for black! So much for copying. There follows 3.c4 c5 4.a3 a6 reaching Fine's parenthetical note 4...c5?  2...c5 is the only move there. Another point, which I already knew from analyzing Tattersall, is that in Fine's variation 1. c3 c6 2. b3 b6 3. a3 a6 4. c4 c5? 5. b4! black can exchange 5...cxb4 6.axb5. Now black has a potential outside passed pawn, and for this reason I generally would prefer to start with 1.a3 instead of 1.c3. But in this exact position it doesn't matter, 6...a5 7.c5! bxc5 8.bxa5! and the tablebase confirms a win for white, although queen endings are not my favorite.
      

    Fine-096.pgn ( 0 KB | 106 Downloads )
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Re: Tattersall A Thousand Chess Endings
    Reply #1 - 05/15/19 at 04:51:58
    Post Tools
    No. 36 C. E. C. Tattersall (no date)
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White to play and win.

    This one holds no mysteries, provided you know Fahrni - Alapin (8/2k5/p1P5/P1K5/8/8/8/8 w - - 0 1). But since both dates are unknown, this composition may have preceded the more commonly cited example, which would be interesting indeed. See the pgn attachment for Tattersall's solution. I will have much more to say about the next endgame.
      

    Tattersall-0036.pgn ( 0 KB | 114 Downloads )
    Back to top
     
    IP Logged
     
    an ordinary chessplayer
    God Member
    *****
    Offline


    I used to be not bad.

    Posts: 1667
    Location: Columbus, OH (USA)
    Joined: 01/02/15
    Tattersall A Thousand Chess Endings
    05/14/19 at 04:40:31
    Post Tools
    This came out in two volumes in 1910 and 1911. I have the Hippocrene reprint. The copyright page says "Introduction to Hippocrene Edition [c]1973 Frank Brady". That exactness is admirable. These days I see quite a few reprints of public domain works where the new edition implies a copyright over the whole work. I didn't find a free ebook, if anybody knows of one I would appreciate a link.

    The work includes nine studies by C. E. C. Tattersall himself. I thought it would be interesting to look at them, both from a chessic viewpoint, and also to form an impression of the books.

    No. 13  C. E. C. Tattersall (no date)
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    * * * * * * * *
    *
    White to play and win.

    Tattersall's solution runs: Quote:
    1.e4! fxe3
    ( 1...f3 2.c5! 2...dxc5 ( 2...Kf4 3.cxd6 Kg3 4.d7 f2 5.d8=Q ) 3. Ke3 c4 4. Kxf3 )
    ( 1... Kf6 2. Ke2 Kg6 3. Kf3 Kg5 4. c5 dxc5 5. d6 )
    2.Kxe3 Kf6! 3.Kd4 Ke7 4.Kc3 Kd7 5.Kb4 Kc7 6.Ka5! Kb7 7.Kb5 Kc7 8.Ka6
    "Were the pieces further to the left White could not win."


    Okay then. The first thing to notice is that at the end of the 2...Kf4 variation (I put it in italics), black is queening with check. The tablebase confirms it's a win, but this is hardly ideal. The second question is can white win in another way? It seems yes, I found: 1.Kc3 Ke4 2.c5 Kxd5 3.cxd6 Kxd6 4.Kd4. This is pretty simple, and it works on other files -- except when white starts with a g-pawn, in which case at the end black can turn it into an h-pawn.
      

    Tattersall-0013.pgn ( 0 KB | 108 Downloads )
    Back to top
     
    IP Logged
     
    Bookmarks: del.icio.us Digg Facebook Google Google+ Linked in reddit StumbleUpon Twitter Yahoo