CCRL Discussion Board

I see you have been busy! You might have noticed that after generating the tablebase, the last thing the program prints before asking you to enter a position is two octal numbers. These are codes for 'maximin' positions, the longest forcible mate. For instance with N+F this is 36 moves. (For these e...

..., and a subset of cases where the king of the side not to move is in check. The key here is the word 'subset'. I know how Nalimov treats indexing of the King locations, which would exclude all positions where the side not to move is in check by the opponent King . But this is reciprocal, so it s...

That could explain why the are different for the two stm. But then I don't understand how the total of won + lost + draw + broken could be different for wtm and btm. That should just be 64^N (times whatever factor you get from symmetry reductions).

I looked at the krnnkr.tbs and kqnnkq.tbs files posted on this website,to figure out if these end-games are generally won or not. But there is something I don't understand: The files contain wtm lost/draw/won/broken and btm lost/draw/won/broken position counts. But when I add them up, the number of ...

Well, since you seem to have a serious interest in this, I decided to do something that had been on my to-do list, namely upgrade the user-friendliness of that EGT generator a little. I uploaded the source code of the improved version to http://hgm.nubati.net/conv4x.c . Compared to the previous vers...

Oh, you are right! This is not strictly as it should be (the word 'void' is required before 'zgen'). Your compiler seems to be a whole lot more critical than mine; even my most recent version that I was using last week still has this, and I never get an error message for it. (I am using gcc under Cy...

Do you generate everything as one big table, or do you generate the smaller tables before the larger tables? If the latter, there is not much difference. I generate one big table. The original generator just iterated by DTM, first all mates (with any number of pieces), then all mat-in-1 by unmoving...

I usually build a table together with all possible successor tables, (stored in the broken positions), as my generator is purely RAM based, and doesn't do any disk acces at all. So it has to start from scratch every time. But even when you build table by table, the problem I sketched would occur whe...

OK, doing only forward moves would indeed avoid many of the problems that troubled me. And it would already greatly speed up the process if you could limit the generation of forward moves to positions that at least have a chance to become decided, because they are conceivably a predecessor of some p...

I am not familiar with building tablebases for checkers (or in general games with mandatory capture). At first glance it seems the 'last piece moved' is part of the game state (like e.p. rights in chess). But you don't need to remember the square, just the piece. So in a 3+3 men tablebase, that woul...

You would make life enormously more easy for yourself if you used an existing graphical chess interface like WinBoard. Then you would not have to bother with any graphics at all, but can limit yourself to writing engines that communicate just through reading text from the standard input (normally ke...

The generator posted on my website (as C code) is reasonably 'fairy friendly', as in fact my main reason for building it was to study end-games with unorthodox pieces. E.g. it takes the way the pieces move from tables, and any piece can be royal. A limitation is that it cannot do Pawns. Another limi...

Sorry for the late reply; I have not visited this forum in ages! The way I did it was by modifying my simple tablebase generator (posted on my website) in the mentioned way: adding a single pass through it where I mark all positions which satisfy a certain condition on material (namely non-bare vs b...

Is your ubuntu machine running 32 bit or 64 bit? I'm fairly sure this line alone with cause major problems on 64 bit: tb = (char*)((int)p+4095&~4095); /* align with cash line */ as the cache alignment code will definitely be truncating p in some situations under linux where an int will be 32 bi...

In my case it was the (FILE *) pointer to the stats file that was corrupted because the terminating 0 of the name string was written into it. This led to an immediate crash, as the first thing the generator does is write the name of the EGT to the stats file. I guess the optimization level also affe...

Unfortunately my long post had not stuck untder the ctrl-c when the server failed, so let me just say this:

MAX should be difined as MEN+2 rather than MEN+1, for the array holding the name of the EGT not to overflow.

OK, so te problem is that it crashes. this is stange, as I am using gcc too (albeit on a 32-bit machine under cygwin). I have no compilation warnigs for the version on my webpage (when I compile as 'gcc -O2' that is; when I use -Wall, I get a lot of pedantic warnings about use of parantheses). And i...

I tried running the code from your website and after changing SIZE and alignment: #define SIZE (1<<6*MEN) for(tb=p;(int)tb&4095;++tb); /* align with cash line */ I am not sure what you intend to accomplish by changing SIZE. I don't expect the program to remain consistent if you do that. SIZE is...

If you don't mind me asking, what are the memory requirements of your in memory structures like compared Nalimov space usage? (If I was to use the code I'd probably write something that saved the index to disk, but I'd still be interested in the relative space differences). The generator uses 1 byt...

I have a version of the generator th source of which is on my website that does DTC. I could send you this source, if you are interested. (It is just a very small modification compared to the dtM verion that is published.) It only does pawnless endgames, though, and it doesn't do anything with the t...

This is what puzzled me: it is already hard to imagine 5-1 positions tht are not a guaranteed 1-0 no matter who has the move. Having the move can hadly be a disadvantage when you can afford to sacrifice anything an still be left with a won end-game. How is the situation for 4-1 end-games. Can't the ...

How do you defne a Zugzwag? f the same position has a shorter DTM with the losing side to move than with the winning side to move?

Well, I can do all (and have done very many) pawnless 3, 4 and 5-men with not-too-strange pieces on 8x8. (They must have 8-fold or 4-fold symmetric, position-independent moves, and capture by replacement, not necessarily in the same way as they move.) I can also do 3-men end-games on arbitrary even ...

No reactions, so it seems you are out of luck. I don't have them either. The help-mates would be very easy to do, though. It would just be a matter of skipping the verification step in the normal generation process, and declaring every potentially lost position immediately lost (except for te cycle ...

I am also reviewing the in-RAM building process now that I started working again on 7-men generation, as even when disk storage is used, as part of the process will still take place in RAM. One new idea to reduce the number of L1 cache misses is to apply white retro-moves not to individual positions...