Winboard Forum

Hi Chess960 friends,

again I point to the fine PDF document from the Frankfurt Chess Tigers.

You can find and download the PDF document (in German language) from the very
interesting Chess Tigers' home site at: http://www.chesstigers.de/download.php
( please search there for the link: Chess960 Startpositionen und Regeln ).

Regards, Reinhard.

Hi Andrew,

well, you know I have suggested another implementation. But when I see you programming that two tables solution that directly, I am happy for that one page solution of all Chess960 positions obviously to be usable and understandable. I hope, that also your example would help to convince more engine programmers to support Chess960 too. But of course there is indeed something more necessary than the initial arrays.

Thank you for publishing your solution as an example here.

Reinhard.

P.S.: it would be better, to replace the code line:

if (Code >= 960) Code = 959;

by

Code: Select all

Code %= 960;

because SP 0 and SP 960 are identic.
P.P.S.: if you like it, you could mention my name as the inventor of that Chess Tigers' positioning scheme.

How about this one.

Code: Select all

void GenFRCPosition(char *fen) {
   char piece[9],*p[8];
   int i,s,k,b1,b2;

   for (;;) {
      memset(piece,'x',sizeof(piece));
      piece[8]='\0';
      for (i=0;i<8;i++)      p[i] = &piece[i];
      k =   (genrand()%6) + 1;
      *p[k] = 'K';
      for (i=k;i<8;i++)      p[i] = p[i+1];   
      s =   (genrand()%(7-k))+k;
      *p[s] = 'R';
      for (i=s;i<7;i++)    p[i] = p[i+1];
      s =   genrand()%k;
      *p[s] = 'R';
      for (i=s;i<6;i++)    p[i] = p[i+1];      
      s =   genrand()%5;
      *p[s] = 'B';
      for (i=s;i<4;i++)      p[i] = p[i+1];
      s =   genrand()%4;
      *p[s] = 'B';
      for (i=s;i<3;i++)      p[i] = p[i+1];
      for (i=0;i<8;i++) {
         if (piece[i]=='B') {
            b1 = i; i++; break;
         }
      }
      for (;i<8;i++) {
         if (piece[i]=='B') {
            b2 = i; break;
         }
      }
      if (b1 & b2 & 1)   continue;
      if ((~b1) & (~b2) & 1)   continue;
      break;
   }
   s =   (genrand()%3);
   *p[s] = 'N';
   for (i=s;i<2;i++)   p[i] = p[i+1];
   s =   (genrand()%2);
   *p[s] = 'N';
   p[s] = p[s+1];
   *p[0] = 'Q';   
   char bpiece[9];
   for (i=0;i<8;i++)   bpiece[i] = piece[i]+(97-65);
   bpiece[8]='\0';   
   sprintf(fen,"%s/pppppppp/8/8/8/8/PPPPPPPP/%s w KQkq - 0 1",bpiece,piece);
}



This is maybe a little nicer

Code: Select all

void GenFRCPosition(char *fen) {
   char piece[9],*p[8];
   int i,s,k,b1,b2;

   for (;;) {
      memset(piece,0,sizeof(piece));
      for (i=0;i<8;i++)      
         p[i] = &piece[i];
      k =   (genrand()%6) + 1;
      *p[k] = 'K';
      for (i=k;i<8;i++)      
         p[i] = p[i+1];   
      s =   (genrand()%(7-k))+k;
      *p[s] = 'R';
      for (i=s;i<7;i++)    
         p[i] = p[i+1];
      s =   genrand()%k;
      *p[s] = 'R';
      for (i=s;i<6;i++)    
         p[i] = p[i+1];      
      s =   genrand()%5;
      *p[s] = 'B';
      for (i=s;i<4;i++)      
         p[i] = p[i+1];
      s =   genrand()%4;
      *p[s] = 'B';
      for (i=s;i<3;i++)      
         p[i] = p[i+1];
      for (i=0;i<8;i++)
         if (piece[i]=='B') {
            b1 = i; i++; break;
         }
      for (;i<8;i++)
         if (piece[i]=='B') {
            b2 = i; break;
         }
      if ( 1 == ((b1&1) + (b2&1)) )
         break;
   }
   s =   (genrand()%3);
   *p[s] = 'N';
   for (i=s;i<2;i++)   
      p[i] = p[i+1];
   s =   genrand()%2;
   *p[s] = 'N';
   p[s] = p[s+1];
   *p[0] = 'Q';   
   char bpiece[9];
   for (i=0;i<8;i++)   
      bpiece[i] = piece[i]+(97-65);
   bpiece[8]='\0';   
   sprintf(fen,"%s/pppppppp/8/8/8/8/PPPPPPPP/%s w KQkq - 0 1",bpiece,piece);
}



Hi Sune,

please don't be too angry, when I place some critic here. But I see a problem in your code. It is that you are mixing up randomizing and creating a well defined starting position.

A randomizing process simply could consist in randomly creating a number between 0 and 959. But the creation of a specified starting position has to be always and worldwide identic.

Some time ago I published my routine for that. It could make sense to repeat it here. It is covering CRC positions also. Moreover a simple test routine has been encountered, too.

Regards, Reinhard.

Code: Select all

//========================================
// Valid CRC / Chess960 Position Generator
//========================================
// Reference Implementation, (C) 2005 by
// Reinhard Scharnagl, Munich, Germany
//----------------------------------------
// Correction 2005-Feb-28 (GC-Nearness)
//========================================

#include < string.h>
#include < stdio.h>

#define TXT_LIM 160

static char FenZone[TXT_LIM];

// insert a symbol into FEN-String
// -------------------------------
// color could be:
// col <  0 => not specified
// col == 0 => bright square
// col == 1 => dark square
void PlaceIntoFEN
  (int cntFree, char symbol, int fieldColor)
{
  for (int pos = 0, free = 0; ; ++pos) {
    if (fieldColor < 0 || ((fieldColor ^ pos) & 1)) {
      if (!FenZone[pos] && cntFree == free++) {
        FenZone[pos] = symbol;
        break;
      }
    }
  }
}

// generating of FEN strings
// -------------------------
// nr could be
// nr >= 0 creating Chess960 position (1 ... 960)
// nr <  0 creating CRC position    (1 ... 48000)
const char *GetFen(int nr)
{
  // knight distributions over 5 free squares
  static const int knight_pos[10] = {
     3, // xx--- (binary encoded)
     5, // x-x--
     9, // x--x-
    17, // x---x
     6, // -xx--
    10, // -x-x-
    18, // -x--x
    12, // --xx-
    20, // --x-x
    24  // ---xx
  };

  // clear the working area
  int bit, pos = TXT_LIM;
  while (--pos >= 0) { FenZone[pos] = '\0'; }

  // test whether CRC is requested
  bool isCRC = (nr <= 0);

  if (isCRC) {
    nr = -nr;
    bool q_first = ((nr % 2) != 0);
    nr /= 2;
    PlaceIntoFEN(nr % 5, q_first ? 'q' : 'a', 0);
    nr /= 5;
    PlaceIntoFEN(nr % 5, q_first ? 'a' : 'q', 1);
    nr /= 5;
  }

  PlaceIntoFEN(nr % 4, 'b', 0);
  nr /= 4;
  PlaceIntoFEN(nr % 4, 'b', 1);
  nr /= 4;
  PlaceIntoFEN(nr % 6, isCRC ? 'c' : 'q', -1);
  nr /= 6;
  pos = knight_pos[nr % 10];
  for (bit = 5; --bit >= 0; ) {
    if ((pos & (1 << bit)) != 0)
      PlaceIntoFEN(bit, 'n', -1);
  }

  PlaceIntoFEN(2, 'r', -1);
  PlaceIntoFEN(1, 'k', -1);
  PlaceIntoFEN(0, 'r', -1);

  int width = isCRC ? 10 : 8;
  char *pC = &FenZone[width];
  *pC++ = '/';
  for (pos = width; --pos >= 0; ) {
    *pC++ = 'p';
  }
  for (pos = 4; --pos >= 0; ) {
    *pC++ = '/';
    if (width >= 10) {
      *pC++ = '1';
    }
    *pC++ = (char)('0' + width % 10);
  }
  *pC++ = '/';
  for (pos = width; --pos >= 0; ) {
    *pC++ = 'P';
  }
  *pC++ = '/';
  for (pos = 0; pos < width; ++pos) {
    *pC++ = FenZone[pos] ^ ('a'^'A');
  }
  strcpy(pC, " w KQkq - 0 1");

  return FenZone;
}

// check if FEN is valid for CRC
// -----------------------------
bool IsValidCRC(const char *pFen)
{
  // to be avoided GC position
  static const char *gcArray = "rnbqckabnr";
  // pawn covering pieces (like a rook)
  static const char *covNear = "rcqk";
  // pawn covering pieces (like a bishop)
  static const char *covDiag = "baqk";
  // pawn covering pieces (like a knight)
  static const char *covDist = "nac";

  int size = (int)(strchr(pFen, '/') - pFen);
  int diff = 0;
  for (int n = size; --n >= 0; ) {
    // different to GC?
    if (pFen[n] != gcArray[n]) {
      ++diff;
    }
    // unprotected pawns?
    if (strchr(covNear, pFen[n]))
      continue;
    if ((n+1) < size && strchr(covDiag, pFen[n+1]))
      continue;
    if ((n-1) >=   0 && strchr(covDiag, pFen[n-1]))
      continue;
    if ((n+2) < size && strchr(covDist, pFen[n+2]))
      continue;
    if ((n-2) >=   0 && strchr(covDist, pFen[n-2]))
      continue;
    return false;
  }
  // GC-near position?
  if (diff < 3 && size == (int)strlen(gcArray)) {
    return false;
  }

  return true;
}

// test output
// -----------
int main(void)
{
  puts("\nfirst Chess960 positions");
  for (int nrFRC = 0; ++nrFRC <= 10; ) {
    printf("(%03d) %s\n",
      nrFRC, GetFen(nrFRC));
  }

  puts("\nfirst CRC positions");
  int cntValid = 0;
  for (int nrCRC = 0; ++nrCRC <= 48000; ) {
    const char *pFEN = GetFen(-nrCRC);
    bool valid = IsValidCRC(pFEN);
    if (nrCRC <= 32) {
      printf("(%05d %s) %s\n",
        nrCRC, valid ? "ok" : "--", pFEN);
    }
    if (valid) {
      ++cntValid;
    }
  }

  printf("\n%d valid CRC arrays\n", cntValid);

  return 0;
}


Hi Reinhard,

Yeah I noticed you passed an integer but I have no real usage for that.

The idea as I see it, is to pick a random position when you want a new
game, so you'd have to pick the integer randomly anyway.

Should one still for some reason want to save or forward the opening
position to others, it would probably be easier to just use the generated
fen directly.
For example the PGN format has such a header tag for FENs.

But truth be told, I see this as just a special case of a more general
game design. I'm making an engine that can play "chess" on
a 64x64 board!

I haven't yet decided how to define castle rules in the general case though,
I'm not sure I even want to tackle that. You can imagine that an opening
position should always be shuffled here though.

What do you think?

Code: Select all

63  n q c e q r e e m r m n m n b c r e m n m b q r c b q c q m c n q q e b q m m e b r c e b n c q c k e b e c q n e c n q b q r q
62  b b q n e c m q b c m e m e r c m n r c r r r n e m m e q b e e r n e m b q n m n r b q q c n e n n e b c b e b b m m n n q r b
61  q n q n c m e m r n q e n b e c c m q c m q r q r c e e c m r n r q b c m m q q n e c m r b e q q c q r b r r b m n n n b r r c
60  r r e r n m m c e m n e q e n b e b c q r e r b m m n e e e e b b r e c r q r m c r m b c m m c n c n n q b n e q r n q c n r c
59  b c r m c c n e c n n c c c e n n m b q r e n b c b q b e q n m q r q c e q b q n q n q e q c c m n b n m c q n m m m n r b m c
58  q b e e c c m c e b n c n b r n b r m q r e e n q m n e e c q e n m r q m b c m c b r r q r n c m b e b r e e b c e e e n q n c
57  e m e n r e c r n e c n q c n n r r r e e b q n m e b q q r r c q b m m n c c c r m q b r n c n e n e b c e c r q b e c e r r e
56  e m m n q n e r e n b m n b b b e e b c c c m e q q m m e r q m b c b e c c b n m e b e c e e q r r c m b e e m e e b e c n c e
55  e q r m n b c c c e e e m e n c q b e b n q c b b b n b r q b n b m c e q n b n c b q b c n r m n e n c e c n e e c r n e m r b
54  e q r r b c m r r m q n c q b c q q m r e q q c n m b n n c m n m c q b q e b c q q q e b n b e n r m m r m c r q r b b b n r m
53  c b b b n c c q e e q q b c m q c n b r b m n r c c e q b n m e b m b e c b n c c e b c c m m q e q q m m c e n c e q b c q r c
52  r b b r r b m b c m m n e r m q c e n c b r c e q r r e r c r e b c q m b q n r e n r q n m c b r b r b c m r m n r c r b e m q
51  e c e m c c r b e n c n r c e b q e e n q b e b n q c n c m r b r c n e n e e e c e q r b c e e m n b n r b e r e m q c b r c n
50  n r n q m q c c r e r e m c c q m c m b c m e m m m m n q m m m e b n b b n n q b m q n b r r r m e m c n n r m n c b m e b c e
49  b e c q q c m c c q e n m m e e e r n e e q b q n e b n m n n n e r c q b m b m n q r q m e m e n b r q c r m b b r c q m r q r
48  e b e c c b r n q r m m q c q r b q e m m c r m m r n n r m b c m r r m c c q e n r m m m e r m n b c b n b r b e r e e q e r r
47  b r b e b m r e c b e e n b n e b r m m r r q n r n m q m q b m q m q n m m n n b b b m c n e m q n b m c b m c n b c c e c c m
46  r c m n e c m n b n r m n n c c n e n n r r n r n c m r r n m b q n b m m q m r q r r c q n c m m b n b n c q b q c m e n n n n
45  q e e r m b c c r c b r q m r b q b n n m e n q q b e m n e c c b r b q n m e q c r m q r e n b m b r c m n r b e e b c b m c m
44  b n b b m q q m b q m c q c c n r b m m r e e r e q b c e c r r c e e b q m r c m n r m q b b m q q e n n n m m e r b c r m r q
43  q r r b n r e m n m q b r n b b b r e b c c b e b e n q m m c n n n r n b r q m r c q b m m n q q e q q q c b r e m m q m m q n
42  e q r m c q r q r n q r n n c m r m c c b n b r c m e c n m n b n n c r q r c q n n r q m q e m r r n n q n c b r n n c q e m e
41  b e n b e q b r b n c r m c e b e m r r q n q n q m e e c e c e q r b e q c c e n c n r q r n n c n q r m r r c r n r n e b m r
40  e m m e b r r q m r b n b e n n c e q q e r m e r e e m c n b c c b r c n r c b r c r e q b c q n m c m e e b q n e n q r m m c
39  m r e b n m e r b e b n c m n n q e b r m n e q n e e m c c n n q q c e q m q n e r q b r m r r n e r m c e c q c c e q c c n b
38  c m n b m n e c b n n r b b m m c n b n m r e b e b q q q q r b c c m m q n q m m b q n m m b r e b e q b m n q b m e r n b n m
37  m q q m q e c b r e m m e q e q r r q m b b c c c b e c r c c c c m m q r b q m n n q q m n c b q q m m r n c b m n m b q b r n
36  q m q n r n q e m c q e m m n c q r r e e q c q c n q n c n m c r m e b m c b r c c q m m e q r c e r b q r m e r e r m n r c r
35  p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p
34                                                                                                                                 
33                                                                                                                                 
32                                                                                                                                 
31                                                                                                                                 
30                                                                                                                                 
29                                                                                                                                 
28  P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P
27  Q M Q N R N Q E M C Q E M M N C Q R R E E Q C Q C N Q N C N M C R M E B M C B R C C Q M M E Q R C E R B Q R M E R E R M N R C R
26  M Q Q M Q E C B R E M M E Q E Q R R Q M B B C C C B E C R C C C C M M Q R B Q M N N Q Q M N C B Q Q M M R N C B M N M B Q B R N
25  C M N B M N E C B N N R B B M M C N B N M R E B E B Q Q Q Q R B C C M M Q N Q M M B Q N M M B R E B E Q B M N Q B M E R N B N M
24  M R E B N M E R B E B N C M N N Q E B R M N E Q N E E M C C N N Q Q C E Q M Q N E R Q B R M R R N E R M C E C Q C C E Q C C N B
23  E M M E B R R Q M R B N B E N N C E Q Q E R M E R E E M C N B C C B R C N R C B R C R E Q B C Q N M C M E E B Q N E N Q R M M C
22  B E N B E Q B R B N C R M C E B E M R R Q N Q N Q M E E C E C E Q R B E Q C C E N C N R Q R N N C N Q R M R R C R N R N E B M R
21  E Q R M C Q R Q R N Q R N N C M R M C C B N B R C M E C N M N B N N C R Q R C Q N N R Q M Q E M R R N N Q N C B R N N C Q E M E
20  Q R R B N R E M N M Q B R N B B B R E B C C B E B E N Q M M C N N N R N B R Q M R C Q B M M N Q Q E Q Q Q C B R E M M Q M M Q N
19  B N B B M Q Q M B Q M C Q C C N R B M M R E E R E Q B C E C R R C E E B Q M R C M N R M Q B B M Q Q E N N N M M E R B C R M R Q
18  Q E E R M B C C R C B R Q M R B Q B N N M E N Q Q B E M N E C C B R B Q N M E Q C R M Q R E N B M B R C M N R B E E B C B M C M
17  R C M N E C M N B N R M N N C C N E N N R R N R N C M R R N M B Q N B M M Q M R Q R R C Q N C M M B N B N C Q B Q C M E N N N N
16  B R B E B M R E C B E E N B N E B R M M R R Q N R N M Q M Q B M Q M Q N M M N N B B B M C N E M Q N B M C B M C N B C C E C C M
15  E B E C C B R N Q R M M Q C Q R B Q E M M C R M M R N N R M B C M R R M C C Q E N R M M M E R M N B C B N B R B E R E E Q E R R
14  B E C Q Q C M C C Q E N M M E E E R N E E Q B Q N E B N M N N N E R C Q B M B M N Q R Q M E M E N B R Q C R M B B R C Q M R Q R
13  N R N Q M Q C C R E R E M C C Q M C M B C M E M M M M N Q M M M E B N B B N N Q B M Q N B R R R M E M C N N R M N C B M E B C E
12  E C E M C C R B E N C N R C E B Q E E N Q B E B N Q C N C M R B R C N E N E E E C E Q R B C E E M N B N R B E R E M Q C B R C N
11  R B B R R B M B C M M N E R M Q C E N C B R C E Q R R E R C R E B C Q M B Q N R E N R Q N M C B R B R B C M R M N R C R B E M Q
10  C B B B N C C Q E E Q Q B C M Q C N B R B M N R C C E Q B N M E B M B E C B N C C E B C C M M Q E Q Q M M C E N C E Q B C Q R C
 9  E Q R R B C M R R M Q N C Q B C Q Q M R E Q Q C N M B N N C M N M C Q B Q E B C Q Q Q E B N B E N R M M R M C R Q R B B B N R M
 8  E Q R M N B C C C E E E M E N C Q B E B N Q C B B B N B R Q B N B M C E Q N B N C B Q B C N R M N E N C E C N E E C R N E M R B
 7  E M M N Q N E R E N B M N B B B E E B C C C M E Q Q M M E R Q M B C B E C C B N M E B E C E E Q R R C M B E E M E E B E C N C E
 6  E M E N R E C R N E C N Q C N N R R R E E B Q N M E B Q Q R R C Q B M M N C C C R M Q B R N C N E N E B C E C R Q B E C E R R E
 5  Q B E E C C M C E B N C N B R N B R M Q R E E N Q M N E E C Q E N M R Q M B C M C B R R Q R N C M B E B R E E B C E E E N Q N C
 4  B C R M C C N E C N N C C C E N N M B Q R E N B C B Q B E Q N M Q R Q C E Q B Q N Q N Q E Q C C M N B N M C Q N M M M N R B M C
 3  R R E R N M M C E M N E Q E N B E B C Q R E R B M M N E E E E B B R E C R Q R M C R M B C M M C N C N N Q B N E Q R N Q C N R C
 2  Q N Q N C M E M R N Q E N B E C C M Q C M Q R Q R C E E C M R N R Q B C M M Q Q N E C M R B E Q Q C Q R B R R B M N N N B R R C
 1  B B Q N E C M Q B C M E M E R C M N R C R R R N E M M E Q B E E R N E M B Q N M N R B Q Q C N E N N E B C B E B B M M N N Q R B
 0  N Q C E Q R E E M R M N M N B C R E M N M B Q R C B Q C Q M C N Q Q E B Q M M E B R C E B N C Q C K E B E C Q N E C N Q B Q R Q

    0 1 2 3 4 5 6 7 8 9101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263



Hi Sune,

I have to confess, that I am not able to understand the intention of your posting. May be you felt as if I have tread on your shoes. But that has not been what I intended. I simply want to point out that there is a difference in creating any random starting array or to derive it from a (randomly chosen) identification number.

Regards, Reinhard.

Hi Reinhard,

I don't understand why you start talking about treading on someone's shoes. I wasn't aware I appeared offended, if so I have certainly given you the wrong impression.

I was just showing you my FRC generator, I am making a GUI and have a need to generate legal FRC positions (and all kinds of other positions for my 64x64 board game).

Since you bring it up though, I admit I don't fully understand your way of
doing it with the integer and numbering scheme.
Is there an advantage here relative to the more traditional use of a FEN?

Chees,
From the non-offended Sune

Hi Sune,

Sune wrote:Since you bring it up though, I admit I don't fully understand your way of doing it with the integer and numbering scheme.
Is there an advantage here relative to the more traditional use of a FEN?

remember that Chess960 is not only played by computers. You have to make it easier for OTB players to start and finally write down their games. Because of this the WNCA (World New Chess association), founded in August 2003, has accepted my numbering scheme of the 960 different starting arrays. Their honourable activities for Chess960 in connection with the Frankfurt Chess Tigers is great and lead to a nearly global acceptance of this system. I simplified it then to be documented on simply one sheet of paper. This enables Chess960 players and clubs to distribute easily the Chess960 rules and positions to anybody who would be interested. If human players decide to have a game of Chess960 they will randomly chose a number between 1 and 960, possibly by the use of the RAND key of a calculator, or a special dice as proposed in my book or at my homesite. With that new scheme they could simply derive the initial array from the drawn number.

There is no need to avoid or forget the extended FRC FEN proposal I wrote two years ago.

Reinhard.

Hi Reinhard,

Okay I see your point now with the OTB play, FENs are not so easily written and read by humans.

I think though that we have to invent even simpler ways of setting up
positions than using a table. Maybe such a table is too complex for
children or you might not have it on you...

A simple way could be to let the humans randomize the position as a
part of the game.
You start the clocks and white puts down the king, presses the clock, then
black puts the queen and so forth.
When all the pieces are set normal play can begin.

One advantage here, is that if both players wants to play
regular chess, they can just decide to generate that position.

-S.

Hi Sune,

Sune wrote:I think though that we have to invent even simpler ways of setting up positions than using a table. Maybe such a table is too complex for children or you might not have it on you...

indeed, I also have suggested in my German language book on Chess960 some simple methods to manually find a random starting position. But if it is necessary to communicate games or results to organizations e.g. for to compute rankings, gathering data for statistics, is is necessary to have a worldwide standard.

Reinhard.