Understanding Rook Endgames: WWW - Reading.ac - Uk/centaur
Understanding Rook Endgames: WWW - Reading.ac - Uk/centaur
Understanding Rook Endgames: WWW - Reading.ac - Uk/centaur
Article
Accepted Version
It is advisable to refer to the publisher’s version if you intend to cite from the
work. See Guidance on citing .
www.reading.ac.uk/centaur
CentAUR
Central Archive at the University of Reading
Reading’s research outputs online
Understanding Rook Endgames
Guy Haworth
Reading, UK1
Understanding Rook Endgames (Müller and Konoval, 2016) or URE is the latest contribution of many to
endgame lore by Gambit Publications, and is the first such work totally checked by 7-man and sub-7-man
‘EGT’ endgame table verification. It is available not only in paperback but on Apple/Android devices
courtesy of the Gambit Chess Studio app, a particularly useful tool here for the mobile reader following
the analysis.
Table 1
maxDTx data for the featured sub-8-man rook endgames.2
The authors scarcely need any introduction. Karsten Müller is the major and most generous purveyor of
endgame wisdom via magazines, training DVDs and books (2001, 2008, 2016). Yakov Konoval was the
first computer scientist to create a program efficient enough to generate 7-man EGTs in reasonable time.
His fruitful partnership with Marc Bourzutschky not only addressed all 7-man endgames but set new
standards in the independence of EGT-verification. New depth records were repeatedly set, in this case to
the Depth to Conversion metric DTC. New chessic knowledge and insights were also created and many
games and studies were highlighted where optimal moves had previously been missed. Having said that,
the existence of definitive EGTs benchmarks also highlights the excellence of endgame play at the top
level. The 7-man findings of Bourzutschky and Konoval (2006-13) appeared often in the magazine EG
and were reviewed in the ICGA Journal (Haworth, 2005-13). A page on the authors themselves would
have been an interesting addition.
1 33, Alexandra Rd., Reading, Berkshire, RG1 5PG, UK. Email: [email protected]
2 DTM Depth to Mate. DTZ Depth to Zeroing of the move-count. The DTZ50 metric considers the 50-move-draw rule
and indicates possibly-modified depths: frustrated wins, fw, are ‘draws’. DTZ50 also gives depths to frustrated wins.
DTZ50 and therefore DTZ50 have been calculated for all 6-man endgames whereas DTZ has not.
Table 2
The structure of URE’s KRPPKRP analysis.
Some 10% of games arrive at a rook endgame, and the main focus of this new book is on KRPPKRP to
which 160 pages are devoted. The preparatory material necessarily includes KRPKR (14pp), KRPPKR
(23pp), KRKP (4pp) and KRPKRP (11pp), and refers to the earlier rook-endgame work of Müller (2001,
2008) and Nunn (1992, 2009, 2010). The last endgames highlighted are KQRPKQR and KRRPKRR (7pp).
Each chapter is laced with history, corrections to past analysis, new discoveries and exercises for the
reader. Closing out, there are chapters on principles of play, theoretical endgames, historic games and
solutions to the exercises. There are comprehensive indexes on players, composers and analysts: all world
champions across 130 years from Steinitz to Carlsen have contributed examples of play.
Examples of longest wins are listed as usual and will fascinate those interested in the most profound, subtle
and inscrutable play, especially as the initial positions are unlikely to occur on the board. See Tables 1 and
3 here, and Haworth (2017) for extended statistics and lines of play. The practical player will be more
assisted by the three tables of statistics, showing the distribution and win/draw profile of actual games
across the various parts into which the endgames are divided. KRPKR for example is simply studied in
terms of the four files for the pawn: a/h, b/g etc. The b/g files feature the most games (18,830 of 60,802)
and the highest likelihood of winning (48%).
KRPKRP is divided according to whether there are passed pawns or not. Winning chances are 31% on
average with maxDTC (per Pp-file arrangement) varying from 25 moves for KRP(a)KRP(b) to 41 for
KRP(a)KRP(c).
Devising a taxonomy for KRPPKRP is the data-mining grand challenge solved by this book. There are
144 configurations of the pawns3 and these are discussed under ten themes based mainly on the relative
positions of the pawns’ files, see Table 2 which is based on the book’s p66. Sub-configurations are again
3 Exercise for the reader? Black pawn on files a-d; White pawns on aa/ab/…/hh; 4 * C(2+7, 2) = 4 * (9 * 8/2) = 144.
usually in terms of specific files for the pawns – ab-a (the most common, 46% wins), bc-a (63% wins),
gh-a (69% wins) etc. Even where pawn-files are not used, readers should easily be able to find the relevant
part of the KRPPKRP chapter.
Table 3
Some highlighted rook endgame positions, with references to URE.
The remainder of this review aims to capture the range and depth of research, and the flavour of the book’s
text, using just six Table 3 positions (#02 and #22-26) as illustrated in Figure 1.
No book on rook endgames is complete without the misnamed KRPKR ‘Lucena’ position which is traced
back (Nunn, 1992, #179; Winter, 2016) to Salvio (1634) rather than to de Lucena (1497). This shows a
common winning technique: the pawn when off the edge allows the king and rook to line up together
behind it to exploit the opposing rook’s lack of diagonal mobility. “1. Re4. The rook moves up the board
to provide a shield for the king. 1. … Rh1 2. Kf7 Rf1+ 3. Kg6 Rg1+ 4. Kf6 Rf1+ (4. … Kd6 5. Re6+
Kd7 6. Re5 Rg2 7. Rg5 +-) 5. Kg5 Rg1+ 6. Rg4 +-. This technique is often referred to as ‘building a
bridge’, with White’s king and rook forming the pillar in the middle. The Lucena win can also be used
with a bishop’s or central pawn.”
KOMODO-STOCKFISH shows that even a top class chess engine can be confounded by a zugzwang. “71.
Rh3? (71. Kd5! Rc3 72. Ke4! Rc5 73. Rh3) Rc6 72. Ke4 Rc5 and now White is to move in the mutual
zugzwang. 73. Rd3 Kxh7 74. Rd4 Kg7 75 Rxb4 [KRPKR] Kf6 76. Kd4 Rg5 77. Kc4 Ke6 78. Rb5 Rg1
79. Kc5 Kd7 80. Rb7+ Kc8 81. Rb6 and the game was later drawn.” There are other examples in the
‘computer games’ §4.14 of engines struggling with fortresses and positional draws.
a b c
d e f
Fig. 1. Table 3 positions: a) #02 ‘Lucena’, b) #22 KOMODO-STOCKFISH, c) #23 Lauronen-Kivipelto,
d) #24 Levenfish and Smyslov, e) #25 Botvinnik-Fischer and f) #26 Euwe-Grünfeld.
Lauronen-Kivipelto includes the DTC-deepest KRPPKRP position which has occurred over the board:
dtc = 61 compared with the maxDTC of 79 moves.4 “34. a4? The win starts with 34. Ke2! Rh4 35. a3!
Kg7 36. Rb2! Ra4 37. Ra2! 34. ... Rh1+? (34. ... Rh5=) 35. Ke2 Ra1 36. Rb4 Ra3 37. f4? (37. Rf4) 37.
... Kg7 38. Kd2 Kg6 39. Kc2 Kf5 40. Kb2 Rh3 41. a5 Kg4 42. a6 Rh6 43. Ra4 Rh8 44. Kc3 f5 45. a7
Ra8 46. Kd3 Kf3 47. Kd2 Rd8+ 48. Ke1 Re8+ 49. Kf1 Ra8 50. Kg1 Rg8+ 51. Kf1 Ra8 52. Ke1 Re8+
53. Kd2 Rd8+ 54. Kc3 Ra8 55. Kd3 Kg4? (55. ... Kg3=) 56. Ke3 Re8+ 57. Kf2 Ra8 5s8. Kg2 Kh4 59.
Kf3 Kh5 60. Ra6 1-0.”
The position from Levenfish and Smyslov (1986, 1989), originally published in 1957, has already been
analysed by Nunn (2014) in his EGT-assisted revisit of that book. The URE text: "According to Levenfish
and Smyslov, White wins, but Black has a nice draw: 1. ... Rc6+ (1... g4? 2. Kxh6 g3 (2. ... Rc6+ 3. Kh7 Kb8
4. Rf4 g3 5. Rg4 Rc3 6. Kg6 +–) 3. Rg7 Rc3 4. Rg4 Ka7 5. Kg5 +–) 2. Kh7 Kb8 3. Rg7 Kc8? This move
loses as does 3. ... Rc7? 4. Kxh6 Rxg7 5. Kxg7 g4 6. h6 +–. Black can hold by 3. ... g4 4. Rxg4 Kc7 5. Rg6
Rc5!! 6. Kxh6 Kd7! = or 3. ... Rd6 4. Rg6 Kc7 5. Kxh6 g4 6. Kh7 Rd7+ 7. Kg8 Rd8+ 8. Kf7 Rd5!! (Levenfish
and Smyslov missed this tactical nuance) 9. h6 Rd7+ 10. Ke6 Rd6+ 11. Kf5 Rxg6 and White doesn't have
hxg6. The move 8. … Rd5!! was also found by John Nunn. 4. Rg6 Kb7 5. Kxh6 g4 6. Kh7 Rc7+ 7. Kg8
Rc8+ 8. Kf7 Rc5 9. h6 Rc7+ 10. Kg8 Rc8+ 11. Kh7 Rc7+ 12. Rg7 g3 13. Kh8! +–.”
From the Historic Games chapter, the only Botvinnik-Fischer confrontation. "One of the most famous
adjourned games. Fischer had sealed 45. ... Rc5. 46. Rf7 Ra5 47. Rxh7. The Soviet team had analysed all
REFERENCES
Bleicher, E. (2017). http://www.k4it.de. EGT-query service on Nalimov-style sub-6-man DTC/Z EGTs.
Bourzutschky, M. and Konoval, Y. (2006-13) Findings from 7-man DTC EGTs. EG issues 11.162,
17.185/6, 18.188/190 and 19.191.
De Lucena, L. R. (1497). Repetición de Amores y Arte de Ajedrez con 101 Juegos de Partido. Online copy
at http://preview.tinyurl.com/hojcn5v.
Dvoretsky, M. (2003). Dvoretsky’s Endgame Manual. Russell Enterprises. 4th edition, 2014.
Fiekas, N. (2017). https://syzygy-tables.info/. Query service on de Man’s ‘syzygy’ DTZ50 EGTs.
Haworth, G. McC. (2005-13) Reviews of MB/YK ‘EG’ articles on 7-man DTC EGT discoveries. ICGA
Journal issues 28.4, 32.4, 34.2, 35.2, 36.1 and 36.2.
Haworth, G. McC. (2017). http://centaur.reading.ac.uk/65694/. This review plus supplementary data and
pgn file.
Левенфиш, Г., Смыслов В. (1986). Теория ладейных окончаний, (The Theory of Rook Endings), 3rd
revised and extended edition. YoYo Media, Moscow.
Levenfish G. and Smyslov, V. (1989). Rook Endings. 2nd revised edition. Batsford. ISBN 978071-
3458091.
Müller, K. (2014). Chess Endgames. DVD series. Chessbase.
Müller, K. and Lamprecht, F. (2001). Fundamental Chess Endings, Gambit, esp. chapter 6.
Müller, K. and Konoval, Y. (2016). Understanding Rook Endgames. Gambit. ISBN-13 978191-0093818.
Müller, K. and Pajeken, W. (2008). How to Play Chess Endgames. Gambit. ISBN-13: 978190-4600862.
MVL team (2017). http://tb7.chessok.com/probe. DTM-evaluation service for sub-8-man endgames.
Nunn, J. (1992). Secrets of Rook Endings. Gambit. Second edition (1999). ISBN-13: 978190-1983180.
Nunn, J. (2009). Understanding Chess Endgames, Gambit. ISBN-13: 978190-6454111.
Nunn, J. (2010). Nunn’s Chess Endings Volume 2, Gambit, esp. chapter 8. ISBN-13: 978190-6454234.
Nunn, J. (2014). Discoveries in R+2P v R+P Endings. ICGA Journal, Vol. 36, No. 3, pp. 139-142.
Salvio, A. (1634). Il Puttino. Online copy at http://preview.tinyurl.com/h8ckttg.
Tamplin, J. (2017). http://chess.jaet.org/endings/ DTC/Z/Z50 EGT-query service.
Winter, E. (2014). The Lucena Position. http://www.chesshistory.com/winter/extra/lucena.html.
Zakharov, V. (2016). http://tb7.chessok.com/. The ‘MVL’ query-service on the Lomonosov sub-8-man
DTM EGTs.