Comments/Ratings for a Single Item

You are right! There is a double move when you use the start position Fairy-Max thinks is standard, but if you set up a position through the GUI the double move disappears?!? This is a bug in Fairy-Max' position setup. Only virgin Pawns have the right to double move, but the test whether a Pawn put on the board during sets up is virgin assumes virginity for second-rank Pawns, rather than using the rank where it would put the Pawns itself. So far the only variants where Pawns did start on other ranks than 2nd were Makruk/ASEAN and Grant Acedrex, and neither of these have a double push. So the bug could remain undetected.

What about Grand chess from where apothecary chess 1&2 take inspiration, wouldn't it have the same pawn problem?

H.G.

I tried to set up the second game the same way and I sadly failed.

Here is what I got:

Initial position:

r2zeel2r/1qnbkabnc1/pppppppppp/10/10/10/10/PPPPPPPPPP/1QNBKABNC1/R2ZEEL2R w - - 0 1

.ini file

/settingsFile=settings.ini
/saveSettingsFile=settings.ini
;
/cp
/fcp="fmax.exe"
/fd="./Fairy-Max"
/scp="fmax.exe"
/sd="./Fairy-Max"
;
/variant=apothecary2
/size=middling
/autoLogo true
;
/showTargetSquares=true
/pieceMenu=false
/sweepPromotions=true

fairy .ini file

// Large-board variant
Game: apothecary2 # PNBRQW.A..C......GKpnbrqw.a..c......gk # elven
10x10=3
7 3 4 5 10 11 9 4 3 7
7 3 4 6 10 12 8 4 3 7
p:100 -16,24 -16,6 -15,5 -17,5
p:100 16,24 16,6 15,5 17,5
n:340 14,7 31,7 33,7 18,7 -14,7 -31,7 -33,7 -18,7 45,6 51,6 -45,6 -51,6
b:420 15,3 17,3 -15,3 -17,3
e:370 15,7 17,7 -15,7 -17,7 34,7 30,7 -34,7 -30,7 48,6 -48,6 3,6 -3,6
l:320 47,7 49,7 -47,7 -49,7 19,7 13,7 -13,7 -19,7 16,6 1,6 -1,6 -16,6
z:300 46,7 50,7 -46,7 -50,7 35,7 29,7 -29,7 -35,7 15,6 -15,6 17,6 -17,6
R:725 1,3 16,3 -1,3 -16,3
A:1000 15,3 17,3 -15,3 -17,3 14,7 31,7 33,7 18,7 -14,7 -31,7 -33,7 -18,7
C:1100 1,3 16,3 -1,3 -16,3 14,7 31,7 33,7 18,7 -14,7 -31,7 -33,7 -18,7
Q:1200 1,3 16,3 15,3 17,3 -1,3 -16,3 -15,3 -17,3
k:-1  1,7 16,7 15,7 17,7 -1,7 -16,7 -15,7 -17,7
k:-1  1,7 16,7 15,7 17,7 -1,7 -16,7 -15,7 -17,7
#
# P& fmWfceFifmnD
# N& NmG
# L& CmW
# Z& ZmF
# E& FAmH
# C& RN
# A& BN
# K& K

Where am I wronG?

- What about Grand chess from where apothecary chess 1&2 take inspiration, wouldn't it have the same pawn problem?

It would, if Fairy-Max would play it. But it doesn't. Mainly for the reason that promotion there is only to captured pieces, and Fairy-Max does not keep track of what is captured. It always uses a fixed promotion piece.

Anyway, I uploaded a fixed version of Fairy-Max.zip, to the same link as before. This allows the double-push even after setting up the position. The lack of a double push could also be the reason that it was reluctant to advance Pawns, as a single push doesn't earn as much bonus as a double push.

Game: apothecary2 # PNBRQW.A..C......GKpnbrqw.a..c......gk # elven
10x10=3
7 3 4 5 10 11 9 4 3 7
7 3 4 6 10 12 8 4 3 7

One thing that is wrong is that not all pieces you define occur in the 'pieceToChar' string in the first line of the definition. In this line each position corresponds to a piece image that WinBoard has, and the letter you write in that position defines for which piece that image will be used. The engine sends this to WinBoard whe you select the variant. Because there is no Z, E and L in this string, WinBoard thinks there are no pieces with these names in the variant. So it does not understand the FEN that contains these letters.

Another problem is that piece number 7 is used as the promotion piece, and in your case that would be the Zebra. Also the 3rd and 4th line of the definition are the white and black pieces on the back rank of Fairy-Max' internal start position. In your piece definitions pieces 12 and 13 are Kings, meaning that white would start without a King. (But setting up the position through WinBoard would cure that.)

You could change it to:

Game: apothecary2 # PNBRQ.EA....Z...C...LKpnbrq.ea....z...c...lk # elven
10x10=3
7 3 4 5 10 12 9 4 3 7
7 3 4 6 10 13 8 4 3 7
 

and move the Rook definition to the 7th place to make that promotion piece.

You should also scale down the piece values a bit. It is not only the ratio that counts. Fairy-Max uses the absolute total to decide when it is safe to start running with its Pawns. These high piece values drive up the total so much that Pawn moves will be discouraged enormously, in the opening.

I did all that and it works, but I'm afraid that now I took away the zebra as a promotion choice as I moved the zebra to the 8th rank!

This is not the diagram. Fairy-Max never has promotion choice. It can only promote to piece #7, and it thinks the opponent can only promote to piece #7 too. It will accept any promotion to a non-royal piece with value not ending in a '3' as input, however.

Best is probably to not exclude any pieces (i.e. make none of the values end in '3'), and leave it to the honesty of the user not tho choose A, C or Q on 8th and 9th rank.

Well first, I decided to change the promotion rule to:

8th rank:less than a rook

9th rank:also to rook

10th rank: also to the 3 strong pieces

The reason for the initial rule was to not always promote to queen, but we got always promote to rook (I mean practically). The new rule is more flexible, but I'm afraid that it still leads to many rook promotions.

Second: the pawn double move still doesn't work, or it is something I did wrong?

Oh! Ok, Silly me!

You might still be using the old Fairy-Max executable. Unfortunately I did not alter the version number, so it is not easy to see. It could of course also be that I goofed in uploading the new version. If you set up a position with a Pawn on the 2nd rank (select Mode -> Edit Position, move back a Pawn, and select Edit Game again), does it allow that Pawn to do a double push?

Weirdly, no! Even pawns set back to the 2nd rank don't get the 2move!

OK, I uploaded a version of the fixed Fairy-Max now that prints as version number 5.0b2. You could try that to know for sure you are using a fixed version. But in preparing it I noticed something EXTREMELY WEIRD, which might be the same thing as that you are facing:

I am using Windows 8.1. When I replaced the old Fairy-Max with the new one, it seems both continued to exist, and double-clicking the file would run the new one, but WinBoard would run the old one. If I renamed the file, the new one got renamed, and both names would work in WInBoard, the reamed one printing the new version, the old name (no longer existing in the folder) would print the the old version number.

[Edit] It seems that the old one it is running is in the same folder as winboard.exe. I also had an fmax.exe there. When I renamed that to yet another name, I get an error message when I use the old name. But I can still choose between the old (in the WinBoard folder) and the new (in the sub-folder Fairy-Max) by just changing the nameof the .exe. I don't understand how it can run files in different folders by just changing the file-name.

The thing with the executables happens to me, too. I can't get the new version working. But I'm willing to give it up, as the new promotion rule chages the game significantly and it's unimplementable in Fairy-Max. And I still haven't found a way to write the aanca and griffin in Skaak II.

I managed on my own the easy task of rewriting the weirdPromotion script:

<script>function WeirdPromotion(x1, y1,x2, y2, promo)
{
  if((board[y1][x1] & 15) != 1) return promo; // moved piece is not a Pawn
  if(y2 == 9 || y2 == 0) return ((promo & 15) == 1 ? 9 : promo); // on last rank: cannot stay Pawn
  if((y2 == 7 || y2 == 2) && ((promo & 15) > 6))return board[y1][x1]; // rook or larger than Rook: remains Pawn
  if((y2 == 8 || y2 == 1) && ((promo & 15) > 7))return board[y1][x1]; // larger than Rook: remains Pawn
  return promo; // choice was acceptable
}
</script>

<script>function WeirdPromotion(x1, y1,x2, y2, promo)
{
  if((board[y1][x1] & 15) != 1) return promo; // moved piece is not a Pawn
  if(y2 == 9 || y2 == 0) return ((promo & 15) == 1 ? 11 : promo); // on last rank: cannot stay Pawn
  if((y2 == 7 || y2 == 2) && ((promo & 15) > 7))return board[y1][x1]; // rook or larger than Rook: remains Pawn
  if((y2 == 8 || y2 == 1) && ((promo & 15) > 8))return board[y1][x1]; // larger than Rook: remains Pawn
  return promo; // choice was acceptable
}
</script>

That change looks OK, but note that the 9 and the 11 in case one select a Pawn on the last rank stand for a WHITE Queen, so that black in that case would promote to an opponent piece. I later corrected that in the script I posted, but by then you had already copied it. The proper way would be to write promo+8 or promo+10, because promo in that case would be 1 (white Pawn) or 1025 (black Pawn).

What exactly is not working with Fairy-Max now? Can you run the one with version number 5.0b2 in WinBoard? For me this version workedfine in both Apothecary 1 and 2.

I am pretty sure that doing Aanca orGriffin is beyond Sjaak's capabilities.

I've made the script chages as you've said. I'm still trying to run 5.0b. I redownloaded from the same link, pasted over what there was with copy and replace in both game directories and double pawn move doesn't work, I think because I'm not running 5.0b. Still working on that.

Actually 5.0b is installed, but not working in winboard!

Now I get it, It is fairy-max 5.0b, but when I paste the new inital position pawns loose their virginity!

The fixed version (that would consider Pawns set up on the 3rd rank virgin) is 5.0b2 . You can double-click fmax.exe to see what version number it prints. If that prints 5.0b the download must have gone wrong. (Perhaps because your browser still caches the old download?) If it prints 5.0b2, but WinBoard still says 5.0b (e.g. in the title bar), there is something wrong in picking the right version to run on your system.

It says 5.0b. I think I haven't dowloaded the proper version, I'll retry!

I haven't downloaded the proper version and I don't know how to do that!

When it says 5.0b2 you ca be certain it is the fixed version. The first fixed version I made was also called 5.0b, but since you still had problem, that leaves doubt which version you were using.There are many ways mixups can occur when downloading new versions (like browser caching).

Is the link the same one?

This link:

http://hgm.nubati.net/Fairy-Max.zip

provides only FairyMax 5.0b, not Fairy-Max 5.0b2. I'm sure of that. I'm not sure where should I find Fairy-Max 5.0b2.