Player # 1 (2200) - Player # 2 (2200)
The  "Goldsby Position" 
 Pensacola,  FL  (2),  (USA)
 01.1975 

 [A.J. Goldsby I] 

This position ... or something very much like it ... arises from a study of an ending,  (K+Q  vs.  K+R); 
which was given in the book, "Practical Chess Endings."  (by GM Paul Keres.)  

YEARS of study, (maybe even DECADES!!!!);  had convinced me that Keres ... 
had GREATLY under-estimated the resources of this position. 

I also have the book,   [The] "Secrets of Pawnless Endings."  (by GM John Nunn.) 
He covers many SIMILAR positions, but not this EXACT position.  (As far as I can determine.) 

(This analysis is also  NOT  found in the series of books, 
"The Encyclopedia Of Chess Endings." {See the section beginning on page # 54, Volume IV.} )

  --->  WHITE TO MOVE AND WIN.     

(Published in  "Chess Life,"  April, 1988. In Pal Benko's column; "Endgame Laboratory."  Pg. # 54. 
  Benko also analyzes this idea in the May {'88} issue of the magazine, as well.) 

1.Qg6!!, {Diagram?} 
Absolutely the best move. 

I was the FIRST to discover this move ... and  the whole procedure which follows. 

I have a book in RUSSIAN on pawnless endings, and the starting position is shown there. 
They do NOT show this continuation, however. 

     [ Many computers want to play the move: 1.Qb4+, ('!?') {Diagram?} 
       but can make little or no progress. 
       (Note: July, 2004: With table-bases I can safely say that Qb4+ or Qc5+ might just work.) 

       I have done literally DOZENS of tests. I even ran this position on a SUN-based 
       computer, with multiple processors. I was also running the latest software for that time. 
       (Nimzo 8.0) Even this machine failed to make any meaningful progress from the 
       starting position. ].  

1...Rd1, {Diagram?}  
The best move here ... and even many Masters were unable to prove a win in this position.  

I was also the FIRST person ... as far as I know ... to  PROVE  a win with K+Q versus 
a King and a separated Rook.  

(This move actually shortens the solution by just a few moves, but ... I am trying to illustrate 
 an idea. And besides, even strong Masters, when faced with this idea over the board, have 
 failed to find an answer. And originally - before computers became powerful enough to 
 assist in this ending - I really thought Rd1 was the best defense.) 

*********

     [  The defense that lasts the greatest number of moves is probably the line that begins 
        with: 1...Kd8!?; {Diagram?} but White wins by forcing Black into the "Philidor" 
        position in about 15-20 moves.  

        The main line runs:  2.Qb6+ Kc8;  3.Qb5! Kc7;  4.Qc5+ Kb7;  5.Ke6! Rh7!?;  
        6.Qd5+!? Ka7;  7.Kd6! Rh2!?;  8.Kc6! Rh6+; {Diagram?}  
        A deep analysis of this position shows that this is probably forced. 
        9.Kc5 Ra6;  10.Qe4!, {Diagram?} This is probably best.  (10.Qd7+!?)  
        10...Rb6; {Diagram?}  This is just about forced.  11.Qd4!, {Diagram?} 
        Nice. Now White applies another of the basic positions ... 
        which I call the 'RAM.' 

        (One of my students just purchased the brand-new Fritz 8.0. 
          He is running on a Pentium 4 with 250 MB of RAM. {No endgame module.}) 

        Fritz 8.0  chooses Qa4+!?, instead of the better move.   (11.Qa4+!?)    

        11...Rb7; {Diagram?}  This is probably forced.  12.Kc6+ Ka8;  13.Qa1+!, {D?} 
        The correct move here.   (13.Qd5!?)   13...Kb8[]; {Diagram?} This is 100% forced. 
         (13...Ra7?; 14.Qh8#)   14.Qa5! , ("+/-") {Diagram?} 
        and although it has taken 14 [long] moves, we have now arrived at ... 
          "The Philidor Position!"   
        (There is no need to look any further. If you have studied your earlier examples, 
          you know how to win.) ].  

*********

2.Qg5+!, {Diagram?} 
The best move out of MANY different alternatives. 
(I was the very FIRST person to show this maneuver, and prove it was a win!! 
 See the 'Chess Life' column quoted previously.)  

     [ Most players - and MANY early computer programs - want to play the move: 
        2.Qg7+, {Diagram?} in this position. ].   

2...Kf8; {Diagram?} 
This is forced ... a King move to ANY light square leads to a fork that will win 
Black's Rook. 

     [ For example, the Rook is lost after:  2...Ke8??;  3.Qh5+, ("+/-") {Diagram?} 
        Or 2...Kf7??;  3.Qh5+, ("+/-") {Diagram?} ].    

3.Qf4+!, {Diagram?} 
Again the best move ... many players choose Qh6+ in this position. 

     [ 3.Qh6+ ('!?') ].  

3...Ke7; {Diagram?} 
Again - forced. 

 (Any King move to a light-colored square will result in the loss of the Rook.) 

     [ Moving the King to any light square will lose the Rook:  3...Ke8?;  4.Qa4+ Rd7;  
       5.Ke6!, ("+/-") {Diagram?} and White wins the Rook.  

       Black cannot run away to the corner here ...  any step on the g-file is disastrous; 
       e.g.  3...Kg7??;  4.Qg4+, {Diagram?} forking King and Rook. ("+/-") ].  

In the next few moves ... the pattern continues.  
4.Qb4+! Kd8;  5.Qb3!!, {Diagram?} 
This is the most accurate ... and the  most  forcing. 

(Many students, computers etc, have all suggested  Ke6  here as superior. But that is 
 just plain silly. Many GM's have checked this position, and found my move to be best. 
 {Not to mention a virtual super-computer!}  At best, Ke6 can transpose to my idea. 
 At worst, it will simply add more moves to the solution.) 

     [ 5.Qh4+!? ].  

5...Re1+!?; {Diagram?} 
Originally I thought this was forced, but Black could actually last (maybe) a little longer 
with a simple retreating move. 

The best move - the one which lasts the longest - is actually  5...Rd7.  But most players, 
even strong Masters, reject ...Rd7; as too passive to be any good! 

Also - my original idea - was to find a  CLEAR  and  FORCED  win ... for those positions, 
where the King and Rook became separated!!  (As in the game I witnesses many years 
ago in the Championship of a major New York Chess Club. See my article that precedes 
my analysis of the Philidor Position.) 

***

     [ Maybe better is: 5...Rd7!?; {Diagram?} and Black might last a few more moves 
       than he does. (But most players reject this, knowing passive defense will not help 
       Black much in these types of positions.) The main line - for the insatiably curious - 
       is: 5...Rd7!?; ('?!')  6.Ke6!, Re7+;  7.Kf6, Rc7;  8.Qb6!, Kc8;  9.Ke6!?, Rb7!?; 
       10.Qa6!?, Kb8;  11.Kd6!?, Rb1;  12.Qa4!!, Rb7;  13.Qa5!, Rb1; ('?!') {Diagram?} 
       Apparently this is forced.  14.Qd8!+, Kb7;  15.Qe7+, Kb8;  16.Qe8+, Ka7; {Box?} 
       This could be forced. 17.Qf7+!, Rb7;  {Box.} This is best, according to several computers, 
       the basic idea is if Black waits ... his Rook will either be picked off, or he will be forced 
       to play this in an even worse position.  18.Qf2+!, Kb8;  This is forced as well. (The corner 
       is a much inferior square.)  (18...Ka6??; 19.Kc6 "+/-")   Now White wins simply, by very 
       quickly applying another one of the basic techniques or positions. 19.Qf4!!, Rc7[]; {Diag?} 
       This is completely forced.  20.Qe5!, {Diagram?} Another of our basic positions. 
       ("The RAM.")  20...Rb7;  21. Kc6+!, Ka7;  22.Qa1+!, Kb8;  23.Qa5. ("+/-") {Diag?} 
       And now, by FORCE ... we have reached another of our basic positions, 
       "The PHILIDOR Position."  
       (I am not 100% convinced of the accuracy of all of these moves. But I cannot check 
        them at the moment - mainly my friend, with the VERY powerful computer, is simply
        too busy - and cannot give me any free time.)  

       If we have done our studies, we know this to be a relatively simple win. 
       (Note: This line is  NOT  re-playable, and is here just for your general information.) ].  

***

6.Kd6!, (a new position)  {See the diagram, just below.}  
Establishing ..............   "The {actual} GOLDSBY Position."   

White wins by force.  

 The actual ...  "GOLDSBY Position." 
 A discovery that changed the course of endgame history. 
(I was the FIRST ... to prove the win with  K+Q vs. K and separated Rook.)

***

--->  Notice the power of this position.   Black's King is nailed the edge of the board. The White King immediately faces him, creating all kinds of threats. The White Queen guards the square that the Black Rook can give a check on. AND!!!! ..... THE WHITE QUEEN THREATENS MATE ON  BOTH  (!!!!) SIDES OF THE BLACK KING!!  This was my original idea. I was the  FIRST  to show one could FORCE a win in these types of situations. Additionally, my analysis which I sent to dozens of authors, clearly showed this idea had been MISSED in practically   ALL   the major (real, over-the-board) games that were played in this type of ending!! (Prior to my discovery.) 

***

     [ The move: 6.Kf4?, {Diagram?} accomplishes very little. ].  

6...Kc8; {Diagram?} 
This is pretty much forced. 

     [ Worse are:  6...Re8; 7.Qb8#;   Or  6...Ke8; 7.Qg8# ].  

White now picks off the Black Rook.  
7.Qc3+ Kb7;  8.Qxe1, {Diagram?} 
White wins easily.  

This winning method was the result of over 20+ years of research and analysis. 
 (But I first had this basic idea circa 1975.)  

 Note: To be honest, GM Pal Benko found the superior 4.Qh4+!! This uses my idea - but much more 
  economically. You can actually shave quite a few moves off my method with his technique.  :p 

Copyright (©) A.J. Goldsby I.  Copyright (c) A.J.G; 2002. 
Copyright (©) A.J. Goldsby, 1992-2001.  

  1 - 0