Syzygy Tablebase Names: A very stupid exercise.

[AndrewGrant]

So, I needed some code in python to generate a list of all the Syzygy file names. KBvK, KQQQvKQ, and so forth. Syzygy has a particular naming scheme. Q > R > B > N > P. The side with more pieces is on the left of the v. The weaker side, if equal pieces, is on the right side.

Here is my solution...

Code: Select all

from itertools import combinations_with_replacements

def build_tablebase_names():

    letters = ['', 'Q', 'R', 'B', 'N', 'P']; candidates = []
    for LHS in combinations_with_replacement(letters, 4):
        for RHS in combinations_with_replacement(letters, 4):
            candidates.append('K{}vK{}'.format(''.join(LHS), ''.join(RHS)))

    def valid_filename(name):
        for i, letter in enumerate(letters[1:]):
            name = name.replace(letter, str(9 - i))
        lhs, rhs = name.replace('K', '9').split('v')
        return 2 < len(lhs + rhs) <= 6 and int(lhs) >= int(rhs)

    return list(filter(valid_filename, candidates))

Extremely unnecessary, but hey it does exactly as desired, without any sort of deeply nested loops. Heres the answer that rests inside the Syzygy code base: https://github.com/AndyGrant/Ethereal/b ... #L676-L733

[AndrewGrant]

A (mostly uninteresting?) observation as well. Determining the number of files that would be used for the table bases up to N pieces is as follows below. Although note that for N=20, KQQQQQQQQQQQQQQQQQQvK would exist, although is illegal.

Code: Select all

2 0
3 5
4 35
5 145
6 510
7 1511
8 4031
9 9751
10 21941
11 46251
12 92503
13 176483
14 323553
15 572263
16 981023
17 1634775
18 2656510
19 4218785
20 6562555

[Ajedrecista]

Hello Andrew:

A (mostly uninteresting?) observation as well. Determining the number of files that would be used for the table bases up to N pieces is as follows below. Although note that for N=20, KQQQQQQQQQQQQQQQQQQvK would exist, although is illegal.

Code: Select all

2 0
3 5
4 35
5 145
6 510
7 1511
8 4031
9 9751
10 21941
11 46251
12 92503
13 176483
14 323553
15 572263
16 981023
17 1634775
18 2656510
19 4218785
20 6562555

I think that what you are looking for is in the following site:

Number of endgames in Chess with any starting position.

Column 'Endgames Cumulative', then substracting 1 if you are not interested in K vs k (2 pieces). Only legal subsets are counted in that page, so KQQQQQQQQQQQQQQQQQQvK does not suit there. If you are not looking for legal subsets only, you still can check the number of legal and illegal subsets with your figures and those of that web.

Regards from Spain.

Ajedrecista.

[AndrewGrant]

I think that what you are looking for is in the following site:

Number of endgames in Chess with any starting position.

No, I'm not looking for anything in particular. Just a better solution than copy pasting ~510 file names for later use.

[Dann Corbit]

Here are the legal names for two through ten men (I guess kk is not very interesting since every element is 1/2):

[dangi12012]

So, I needed some code in python to generate a list of all the Syzygy file names. KBvK, KQQQvKQ, and so forth. Syzygy has a particular naming scheme. Q > R > B > N > P. The side with more pieces is on the left of the v. The weaker side, if equal pieces, is on the right side.

dont now if you are interested after all this time but:

what you need here is the python equivalent of

Code: Select all

std::next_permutation

and skip every item that is *Not in order* (Q > R > B > N > P)

[AndrewGrant]

As the title suggests, no. "A very stupid exercise", but a cute compact solution.