Bored with orthodox Chess?

[Greg Strong]

I was still too busy settng up my live Spartan games. This has been done now, though, and currently ChessV is playing Fairy-Max at

http://80.100.28.169/spartan/spartan.html

Nice! Good job with this

[Roger Brown]

Thanks for the valuable help anyway. Current Winboard is awesome and its graphics are too, so i wonder how people whine/complain about bad Winboard graphics etc etc.

Hello George,

We (H.G. and I) wonder about that too!

Later.

[hgm]

OK, a fixed version is uploaded to the beta installer now.

The problem was that at some point I had changed WinBoard to only create bitmaps from the font pieces that it actually needed in the current variant, because converting all pieces led to very weird display problems on computers with small video memories. But then of course I have to recalculate on every variant switch. But the re-calculation was only done when the number of board ranks or files changed (because calcuating the bitmaps was part of the general routine to calculate a new board layout. Which was OK when you switched from normal to gothic, but not in general.

[George Tsavdaris]

OK, a fixed version is uploaded to the beta installer now.

OK many thanks!
Of course you mean this link right?
http://hgm.nubati.net/WinBoard-4.5.0.exe

Does this include fixes(or it's the same issue so it is solved automatically) to the other problem i reported here and here?

[George Tsavdaris]

OK, a fixed version is uploaded to the beta installer now.

OK many thanks!
Of course you mean this link right?
http://hgm.nubati.net/WinBoard-4.5.0.exe

Does this include fixes(or it's the same issue so it is solved automatically) to the other problem i reported here and here?

Hmm as it seems it does! So graphic problems have vanished! Hooray.

If only Microsoft where as quickly in identifying, understanding and solving bugs as you. I have already send them a bug report about it(about the @ issue in shortcuts) and they replied saying something like:
You can try not to put @ if this creates problems. Try - or +.

Thank you Microsoft for the enlightening answer!! It really helped.

[hgm]

30 Games played, no illegal moves, no crashes. A lot of losses though!

Hopefully there will be a small tournament?

What is the current status of the various Spartan engine developments?

Richard: Is a Catalyst binary available for download somewhere?
Gregory: Is ChessV now all finished? You mentioned something about "enabling the bitboard generation" in your e-mail.
Do you still plan to convert Quadrox as well?

[Daniel Shawul]

Hey, I did not forget about this.
It can play chess variants now. Checkers too. Name is "Nebiyu" meaning Prophet.
Othello and Go are in progress but I need to study the rules carefully.
If I were working on Spartan alone I would have been done long ago.

Questions : What is the maximum number of different types of fairy pieces that can appear on a single variant at the same time ?
Are they all substitution of the orthodox pieces like a "hoplite" for a "pawn" ?
Anyone has perft for chess variants especially Spartan ?

Thanks

[hgm]

It depends a little on if you consider black pieces that are mirror images of white pieces as the same piece type or not. (E.g. in Fairy-Max white Pawns and plack Pawns must be defined as different piece types.) Shogi has many such asymmetric pieces (Pawn, Knight, Lance, Silver, Gold).

Shogi is the worst serious variant I know for this anyway, because there are many piece types to start with (8, of which 5 asymmetric), and then most piece types promote to yet other piece types. This also raises the problem of pieces that move the same, but are different in other respects (like reverting to another unpromoted form when captured, such as a Crazyhouse Queen that in reality is a promoted Pawn). I guess you would want to define those as different piece types too. Shogi then has 14 piece types (of which 9 asymmetric).

In Fairy-Max I allow 15 piece types, shared by black and white. The latter is not smart, btw; I am thinking of decoupling the white and black piece types completely in a future version, and during initialization fill the black move-generator tables with the mirrored move of the white.
Apart from Shogi (which Fairy-Max cannot play anyway, because it is a drop game), I have never found an interesting variant where 15 was not enough, even when you have to define asymmetric pieces twice. Of course there are giant Shogi variants (e.g. on a 36x36 board) with many dozens of piece types, but I don't think anyone ever succeeded to play those.

[PK]

good news: I have managed to reach the stage where my variant engine (currently playing only orthodox chess, shatranj and spartan chess) has move generator and search, but no eval to speak of. If no new bugs show up, I should have something working on the Sunday evening. so please sign in Oberon (named after the fairy prince from Midsummer Night's Dream)

[hgm]

Ah, that is great news. I am currently studying the possibility to convert the new engine I last worked on in March, which then only did normal Chess. The 0x88 test is a bit tricky, though, because of the jumping Spartan pieces, and the fact that I use 8-bit capture codes. So it is a bit of a puzzle to squeeze it in. I think I found a solution now. This would require me to adapt the meaning of the bits to the variant: in Shatranj, where the Alfil & Ferz partcipate, I would need different assginment from Spartan, which would indicate these two moves by the same bit for the benefit of the Lieutenant.