WhiteTower: As for Chess960 / FRC there will be necessary 960 * 400 years to have the same opening knowledge extension, supposed the favored starting array will be exchanged all 400 years. For CRC it will last another 25 times longer.

Nasmichael: Any well designed improvement could be noticed by detecting an initial witch hunting by traditionalists against it. Regarding that Chess960 (FRC) and 10x8 Chess (CRC, GC) will have a very promising future.

Pythagoras: Thank you for that hint!

Thad: there seems a great event just about to be created - so I just have read at http://www.gothicchess.com/news.html.

There just has been a live game, where SMIRF's problems to develop pieces' mobility have been shown:

[Event "Live test game 20 min + 15 sec"]
[Site "GC live"]
[Date "2005.11.05"]
[Time "22:21:13"]
[Round "A"]
[White "Gothic Inventor"]
[Black "Smirf Beta"]
[Result "1-0"]
[Annotator "RS"]
[SetUp "1"]
[FEN "rnbqckabnr/pppppppppp/10/10/10/10/PPPPPPPPPP/RNB QCKABNR w KQkq - 0 1"]

1. c4 Nc6 {(11.16) -0.176} 2. Nh3 g6 {(11.14) -0.162} 3. Nc3 Bd4 {(12.01)
-0.262} 4. g4 Nh6 {(11.09) +0.004} 5. e3 Bxc3 {(12.14) +0.170} 6. dxc3 f5
{(12.12) +0.123} 7. Bd5 e6 {(14.00) +0.053} 8. g5 exd5 {(12.01=) +0.898} 9.
gxh6 dxc4 {(12.06) +0.777} 10. hxi7 Axi7 {(13.00) +0.629} 11. e4 h6 {(13.00)
+0.555} 12. exf5 gxf5 {(12.28) +0.178} 13. Cxe8+ Qxe8 {(12.01=) -0.266} 14. Ae2
Ne5 {(11.10) +0.432} 15. O-O Nd3 {(10.34) +0.391} 16. Bf4 d6 {(11.11) +0.469}
17. Qd2 Nxf4 {(12.01=) -0.203} 18. Axf4 O-O {(13.00) +0.078} 19. Rae1 Qd8
{(12.01) -0.037} 20. Ni5 Aj5 {(13.01=) -0.441} 21. Ag6+ Kj8 {(14.00) -0.371}
22. Rhg1 j6 {(13.00) -1.283} 23. Rg3 Qg8 {(12.27) -2.475} 24. Re7 jxi5 {(14.01)
-7.059} 25. Qe2 Ah3+ {(12.00) -41.21} 26. Rxh3 h5 {(03.25) -10.90} 27. Rj3+ Qj5
{(02.00?) -2.664} 28. Rxj5# 1-0

a) there is an only time-restricted beta of SMIRF at
http://www.chessbox.de/Compu/schachsmirf_e.html
b) you can download a reduced but always working Gothic Vortex Copper via
(a well known web address)
c) a free version of ChessV could be found via
http://sourceforge.net/projects/chessv/
d) a beta of the newcomer Fereshte has been seen at
http://home.ripway.com/2005-10/493246/chess/index.html

Anencephal: I have sent a C++ file to you here, but it is not keeping its formatting. We should communicate by email. Anyone who wants it could ask me for that CRC/Chess960 numbering scheme.

Reinhard.

P.S.: also have a look at http://www.chessbox.de/xfen_e.html or http://www.chessbox.de/beta.html

Anencephal: Thank you to give us that download possibility! I loaded and started it. The first impression is, that here is a new sophisticated program with an own GUI, what is double work, which should be honoured! The first view detected a surprising amount of functionalities. Some remarks:

a) You seem to have implemented a different numbering scheme, why? I could supply you with a standard position generating subroutine.

b) Your program seems not yet absolutly stable. It crashes when answering to following:

[Event "Testgame 1m/move"]
[Site "ChessBox"]
[Date "????.??.??"]
[Round "1"]
[White "Smirf Beta"]
[Black "Fereshte 0.8"]
[Result "*"]
[Annotator "R.S."]
[SetUp "1"]
[FEN "nrakcqnrbb/pppppppppp/10/10/10/10/PPPPPPPPPP/NRA KCQNRBB w KQkq - 0 1"]

1. Nf3 {(11.03) +0.246} c6 2. i4 {(11.11)
+0.631} h5 3. ixh5 {(13.02+) +1.012} *

Reinhard.

P.S.: c) Your program uses a permanent brain (or at least 50% processor capacity). This should be off switchable when playing against another program at the same computer.

d) it could help to have any sound, when a move has been calculated.

Anencephal: of course CRC programs are of interest !

Pythagoras: Thank you for performing that tournament! I simply want to explain, that because of SMIRF's multi-reiterative approach the size of the used cache is very important. It should be set to 80 MB or 160 MB if possible.

Pythagoras:
a) Good to see you continuing your engine testing.

b) If you would edit your posting end replace the line break inserted by brainking in the X-FEN line by a space, we would become able to load the game by cut and paste.

c) SMIRF now has come to the version 1.29, but it still will be improved.

d) If you would be interested in SMIRF's special Perft results (completely analyzing moves e.g. whether they are check threats or even mates) you could see SMIRF's current results here:

FEN: rnbqckabnr/pppppppppp/10/10/10/10/PPPPPPPPPP/RNBQ CKABNR w KQkq - 0 1

Ply Nodes all (x) (e.p.) all (+) (#) Prom. Castl. Sec.
------------------------------------------- ------------------------------------
1 28 0 0 0 0 0 0 0
2 784 0 0 0 0 0 0 0
3 25283 387 0 120 0 0 0 0
4 808984 14812 0 4019 0 0 0 0
5 28946187 907953 468 304360 398 0 0 1.0
6 1025229212 36740962 14017 11052055 17277 0 0 36.8
------------------------------------------- ------------------------------------

Sorry, it seems to be impossible to format text here ...

Luke Skywalker: a) As seen in http://brainking.com/de/ArchivedGame?g=1111972 CRC is sufficiently complex even without unprotected pawns. So having 48000 candidates it seems to be a good idea to select those about 40% more stable starting arrays.
b) The king could castle with the c-side Rook and becoming placed at the third file (c) from left, or with the i-side Rook and becoming placed at the second file from right (i). The Rook will find its place then on the King's inner neighbour square.

Caissus: the most important thing in CRC is to get a survey of the board situation. Else it could become a sort of minesweeper.

at http://www.chessbox.de/Compu/schachsmirf_e.html free testable including October.

Anencephal: With a small program I generated following:

first CRC positions
(00001 --) aqbbcnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/AQBB CNNRKR w KQkq - 0 1
(00002 ok) qbbacnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/QBBA CNNRKR w KQkq - 0 1
(00003 --) abbqcnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/ABBQ CNNRKR w KQkq - 0 1
(00004 ok) qbbcnanrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/QBBC NANRKR w KQkq - 0 1
(00005 ok) abbcnqnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/ABBC NQNRKR w KQkq - 0 1
(00006 --) qbbcnnrakr/pppppppppp/10/10/10/10/PPPPPPPPPP/QBBC NNRAKR w KQkq - 0 1
(00007 --) abbcnnrqkr/pppppppppp/10/10/10/10/PPPPPPPPPP/ABBC NNRQKR w KQkq - 0 1
(00008 --) qbbcnnrkra/pppppppppp/10/10/10/10/PPPPPPPPPP/QBBC NNRKRA w KQkq - 0 1
(00009 --) abbcnnrkrq/pppppppppp/10/10/10/10/PPPPPPPPPP/ABBC NNRKRQ w KQkq - 0 1
(00010 --) baqbcnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BAQB CNNRKR w KQkq - 0 1
(00011 --) bqabcnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BQAB CNNRKR w KQkq - 0 1
(00012 ok) bbqacnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBQA CNNRKR w KQkq - 0 1
(00013 --) bbaqcnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBAQ CNNRKR w KQkq - 0 1
(00014 ok) bbqcnanrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBQC NANRKR w KQkq - 0 1
(00015 ok) bbacnqnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBAC NQNRKR w KQkq - 0 1
(00016 --) bbqcnnrakr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBQC NNRAKR w KQkq - 0 1
(00017 ok) bbacnnrqkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBAC NNRQKR w KQkq - 0 1
(00018 --) bbqcnnrkra/pppppppppp/10/10/10/10/PPPPPPPPPP/BBQC NNRKRA w KQkq - 0 1
(00019 ok) bbacnnrkrq/pppppppppp/10/10/10/10/PPPPPPPPPP/BBAC NNRKRQ w KQkq - 0 1
(00020 --) bacbqnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BACB QNNRKR w KQkq - 0 1
(00021 ok) bqcbannrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BQCB ANNRKR w KQkq - 0 1
(00022 --) bbcaqnnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCA QNNRKR w KQkq - 0 1
(00023 ok) bbcqannrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCQ ANNRKR w KQkq - 0 1
(00024 ok) bbcnqanrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN QANRKR w KQkq - 0 1
(00025 ok) bbcnaqnrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN AQNRKR w KQkq - 0 1
(00026 ok) bbcnqnrakr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN QNRAKR w KQkq - 0 1
(00027 ok) bbcnanrqkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN ANRQKR w KQkq - 0 1
(00028 --) bbcnqnrkra/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN QNRKRA w KQkq - 0 1
(00029 ok) bbcnanrkrq/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCN ANRKRQ w KQkq - 0 1
(00030 ok) bacbnnqrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BACB NNQRKR w KQkq - 0 1
(00031 ok) bqcbnnarkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BQCB NNARKR w KQkq - 0 1
(00032 ok) bbcannqrkr/pppppppppp/10/10/10/10/PPPPPPPPPP/BBCA NNQRKR w KQkq - 0 1
...
21821 valid CRC arrays

Anencephal: well this additional rule has two reasons: a) to have more balanced starting arrays, b) to avoid a need for a 7-choice, if the position would be created by the use of traditional dice. 48000 is the number of possible CANDIDATES for starting positions, but only those could be selected with protected Pawns. SMIRF would do that for you. Or you could start a CRC game at BrainKing.

Anencephal: How did you get this number? SMIRF is selecting valid candidates among those 48000 numbers. Almost every second one would be illegal. Thus the selecting process is essential.

Pythagoras: Well, there is an alternative method for experienced users. After having SMIRF installed once you could expand the file http://www.chessbox.de/Down/TestPackB.zip into your SMIRF program folder. This ZIP will contain the new non DLL parts of the project. Simply overwrite everything there. But be aware, that all is still beta. Version 1.21 probably will be released tomorrow. (P.S.: now it exists.) (P.P.S.: now 1.23.) (P³.S.: now 1.24.)
The DLLs are packed into a (probably) constant file http://www.chessbox.de/Down/ConstDLL.zip. This would enable you e.g. to install SMIRF onto a memory stick by expanding both ZIP files into one new folder.

Pythagoras: a) SMIRF 1.21 is not yet sufficiently fast. But it will play your desired moves:
(P.S.: same is valid for version 1.20 ;-) )

01:42.2 (15.01=) +2.598 34...Cj4+ (=/=) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Rxi3+ 38.hxi3 Cj3+ 39.Kh1 Cxj2+
01:39.7 (15.00-) +2.536 34...Cj4+ (=/=) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
01:30.4 (14.01=) +2.817 34...Cj4+ (=/=) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
01:29.4 (14.01+) +2.817 34...Cj4+ (=/=) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
01:27.4 (13.02=) +2.561 34...Cj4+ (Cj5) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
01:27.3 (13.02+) +2.561 34...Cj4+ (Cj5) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
00:56.1 (13.02+) +1.418 34...Cj4+ (Cj5) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Rxi3+ 38.hxi3 Cj3+ 39.Kh1 Cxj2+
00:53.1 (13.02+) +0.847 34...Cj4+ (Cj5) 35.Ki1 Bg2 36.Nxf7 Bh3 37.Cxc5 Rxi3+ 38.hxi3 Cj3+ 39.Kh1 Cxj2+
00:44.6 (13.00-) +0.785 34...Cj5 (=/=) 35.Nxf7 Bg2 36.j3 Cj4+ 37.Kj2 Ch3+ 38.Ki2 Rf8 39.Ne5 Rxf4
00:30.7 (12.01=) +0.927 34...Cj5 (Cj4+) 35.Nxf7 Bg2 36.j3 Cj4+ 37.Kj2 Ch3+ 38.Ki2 Rf8 39.Ne5 Rxf4
00:28.5 (12.01+) +0.927 34...Cj5 (Cj4+) 35.Cxc5 Bg2 36.Ki1 f6 37.Rg1 Cj3+ 38.Ki2 Cj4+ 39.Kj1 Bh3+
00:27.2 (12.01+) +0.642 34...Cj5 (Cj4+) 35.Cxc5 Bg2 36.Ki1 f6 37.Rg1 Cj3+ 38.Ki2 Cj4+ 39.Kj1 Bh3+
00:21.3 (11.03=) +0.499 34...Cj5 (Cj4+) 35.Cxc5 Cj4+ 36.Kj1 Rxi3 37.hxi3 Cxj2+ 38.Kxj2 Bh1 39.Rxh1 f6
00:20.8 (11.03+) +0.499 34...Cj5 (Cj4+) 35.Cxc5 Bg2 36.Ki1 Cj3+ 37.Ki2 Cj4+ 38.Ki1 Ch5 39.e4 Ch3+
00:16.1 (11.01=) +0.499 34...Cj4+ (=/=) 35.Ki1 Ch3+ 36.Ki2 Bg2 37.Cxc5 Cj4+ 38.Ki1 Ch5 39.e4 Ch3+
00:12.8 (11.00-) +0.960 34...Cj4+ (=/=) 35.Ki1 Ch3+ 36.Ki2 Bg2 37.Cxc5 Ch5 38.Ki1 Cxe5 39.fxe5 Rxi3+
00:09.8 (11.00-) +1.246 34...Cj4+ (=/=) 35.Ki1 Ch3+ 36.Ki2 Bg2 37.Cxc5 Ch5 38.Ki1 Cxe5 39.fxe5 Rxi3+
00:05.7 (10.01=) +1.389 34...Cj4+ (=/=) 35.Ki1 Ch3+ 36.Ki2 Bg2 37.Cxc5 Ch5 38.Ki1 Cxe5 39.fxe5 Rxi3+
00:03.2 (10.01+) +1.389 34...Cj4+ (=/=) 35.Kj1 Bg2 36.Cxc6 Bh3+ 37.Ki1 Ci2+ 38.Kh1 Cxj2+ 39.Kg1 Cj1+
00:02.7 (10.00-) +1.326 34...Cj4+ (=/=) 35.Kj1 Ch3 36.Ki2 Bg2 37.Cxc5 Ch5 38.Rd1 Cxe5 39.fxe5 Rxi3+
00:01.9 (10.00-) +1.658 34...Cj4+ (=/=) 35.Kj1 Ch3 36.Ki2 Bg2 37.Cxc5 Ch5 38.Rd1 Cxe5 39.fxe5 Rxi3+
00:01.3 (09.01=) +1.842 34...Cj4+ (=/=) 35.Kj1 Ch3 36.Ki2 Bg2 37.Cxc5 Ch5 38.Rd1 Cxe5 39.fxe5 Rxi3+
00:01.0 (08.01=) +1.414 34...Cj4+ (=/=) 35.Kj1 Ch3 36.Ki2 Bg2 37.Cxc5 Ch5 38.Rd1 Cxe5 39.fxe5 Rxi3+
00:00.6 (07.01=) +1.501 34...Cj4+ (=/=) 35.Kj1 Ch3 36.Ki2 Bg2 37.Cxc5 Ch5 38.i4
00:00.3 (06.02=) +2.026 34...Cj4+ (Ci5) 35.Kj1 Ch3 36.Cxc5 Cxh2+ 37.Ki1 Cxe2 38.Nxf7 Ri5 39.j4
00:00.3 (06.02+) +1.429 34...Cj4+ (Ci5) 35.Kj1 Ch3 36.Nxg4 Bf5 37.e4 Bxg4 38.Ri1
00:00.1 (05.00-) +1.111 34...Cj4+ (Ci5) 35.Kj1 Ci4 36.Ki1 j5 37.Nxf7 Cxi3+ 38.hxi3 Rxi3+
00:00.0 (04.03=) +1.551 34...Cj4+ (Ci5) 35.Kj1 Ci4 36.Ki1 j5
00:00.0 (02.06=) +2.334 34...Cj4+ (=/=) 35.Kj1 Rxi3 36.hxi3 Cxj2+ 37.Kxj2 Bh1

b) that behaviour mainly depends on the piece values.

c) SMIRF sees that advantage mostly in the form of a slight increase of positional evaluation, it is also related to b).

Pythagoras: you can of course use that version, you want to test. But because SMIRF's development is going on, there will exist new versions each time I fixed some errors or I implemented new ideas. So in short there will be one with better fork handling and slightly improved speed (about 5%). Thank you very much for your tournament. I always take some experiences from that!

Pythagoras: now there is a new Smirf Beta version 1.20 with a slightly improved behaviour, especially during the opening.

Pythagoras: you have started an interesting testing approach! It has already helped me to identify a bug. So I wonder for what I should hope - at least to find more bugs?

Pythagoras: Well, I have erased a bug, which enabled bad values to be accumulated in the cache. Thus there is a new beta 1.19.

Pythagoras: I have tried it with the initial low cache setting ... the cache size you mentioned is not possible in 1.17, check it within the options menu.

Pythagoras: I cannot have replayed those game moves. Could you please tell me, what is your processor, speed and the used cache size?

jolat: because SMIRF consists of two parts (GUI + Engine) there are different version numbers for both. The version number of the Engine is displayed in the caption of the GUI, the version of the GUI could be found via Help -> About ...

jolat: the beta is always subject to change. you will find a current beta to be downloaded at http://www.chessbox.de/Compu/schachsmirf_e.html

WhiteTower: You are wrong. Castling is a well defined process. To do it in a performant way (generating / executing) needs a matching environment and data structure. SMIRF is able to also support Fischer castlings in 8x8 and 10x8, also symmetrical castling like in Janus Chess already is supported. But the inverse castling of MBC does not match into those optimized structures. Remember, in SMIRF there is only ONE engine flexibly able to play a lot of 8x8 and 10x8 variants. If performance would be irrelevant, including new variants might be easy. But such is not the current challenge.

I actually argue for the SIMPLE solution, to play MBC the mirrored way, because that fits into SMIRF's structure. So the problem could easily be solved by changing the point of view. Otherwise you will have to define also new types of FEN and PGN, to make MBC games persistant.

Walter Montego: as I already have tried to explain, SMIRF is only able to play the mirrored MBC array, because of the intended unorthodox castling. I do not intend to implement a third method of castling yet, because that would affect very central parts of the program.

andreas: thank you, Andreas, for your feedback! SMIRF is not designed to play Blitz well. It should develop positional ideas and therefore needs some time for a good answer. But technically it plays Blitz, too. I have read your suggestions, there will be an analyse mode soon. You will have noticed, that SMIRF is always making a sort of multivariant modus by its move-pair iterative deepening. If a second best move seems to be about as good as the best one, then in the PV it will follow the best move in "()". SMIRF GUI + Engine is still a one man project. So it will need time to enrich the package with new elements.

Pythagoras: what should I say? I am not a lawer. In any case Smirf will not automatically produce a GC starting array without having a GC license. Thus the GC button will be disabled then, too.

Pythagoras: Actually a GothicChess key could not be published generally, before agreements with Ed Trice will have been made. But those things are in progress, so I hope.

But in a VARIANTS.PGN attached to the current beta there is a loadable alternative MBC (Embassy), but only in mirrored view because of its unusual original castling. Maybe it also could be an interesting starting array.

Replace beta keys by

User=Test Smirf free until 2005/10/31
KeyGUI=6c5747c2 5917bb40
KeyEngine=e207f259 d58a36f3

and give me some feed back - thank you.

PS: Download from http://www.chessbox.de/Compu/schachsmirf_e.html

Pythagoras: I hope to have that fixed until tomorrow.

Concerning Gothic and Janus Chess: Smirf does not support an opening library yet. Thus a learning opponent will outperform SMIRF soon. It actually is optimized for CRC and Chess960.

Pythagoras: Well, everybody is invited to participate the beta testing. At http://www.chessbox.de/Compu/schachsmirf_e.html there is a download possibility. There will be in short a new beta version again testable free and unrestricted during October. But the problem is, that several people seem to have difficulties to delete previous Smirf's INI files before installing the new beta. Smirf cannot delete that file itself, because it would later contain the personal key, which should be preserved. I am thinking for a solution to make it more fail save.

Walter Montego: well, I am just about to fix an error inserting CRC PGN's into SMIRF's GUI from Brainking directly by cut and paste. So there soon will be a new beta, and I will describe here how to start beta testing (for those who are interested).

The only tournament Smirf has participated has been in last November, where it tooks place four of eight in the Gothic Chess World Championship. But Smirf just has been born that time. Today it should be stronger.

Walter Montego: Then I have to apologize to not have made the beta handling more secure.

Walter Montego: indeed, but the new version seems to have produced some sort of frustration, I am not aware of.

Walter Montego: I am not yet ready with SMIRF. As always specified SMIRF has been a beta. And unfortunately I have got only very few feed back. So I have to test and fix and to improve without nearly any detailed resonance. When I am convinced that SMIRF would be ready to be distributed, I will announce that. Eventually there again will be a new serie of betas with a limited testing period. I will decide tomorrow.

jolat: No, if SMIRF would not use its full power during the applied testing period, it would be an error. But to examine that, I need more details as I have explained before.

Walter Montego: That is the way Shareware works. You could test its general abilities. But without having a personal key its strength is of course limited. Otherwise nearly nobody then ever would spend some money for SMIRF's development.

If you have installed the newer beta into the same folder as the old one, and you have NOT deleted the old INI file, then the new key would never have become valid. Because of that I have written such hints on my beta download page.

SMIRF has provided a two month valid free testing key, which just is expiring with October. If someone wants to test SMIRF's full abilities, he could request a personal limited testing key. But that is intended to be an exception.

rabbitoid: Well, it is hard to make a distant analysis for that situation. The cache used should be sized between 40 MB and half of the available memory. And having two programs playing one against another at a single computer without several processors will make no sense if any of those programs would be pondering or (as possible in SMIRF) has activated its permanwent brain. Equal play could be awaited only at equally distributed resources.

Nevertheless SMIRF is no magic or divine program. It will be always subject for improvements.

jolat: To inspect the cache size simply watch under options, where the cache could be doubled or halved. And the other program should know, whether it would be pondering or not.

jolat: well, you could see it yourself. Whenever Smirf's evaluation gets very low compared to the previous one, there it might have been a better move. SMIRF has not been designed primarily for blitz. So it actually could find its master there. Nevertheless I wonder about the circumstances. What has been the cache size? Was the other program always running (pondering)?
P.S. when I try to reconstruct it, on my computer I set 23 sec/move to get the first two answers. Already the third one would be f5. So there must be something what slows down SMIRF at your computer. Either you have other programs running or you have set the cache size very low.

jolat:
Your version is the current beta.

If you would post the PGN (Edit -> Copy as -> PGN, then insert it here), I will check it.

jolat: there are (hidden) possibilities to download earlier versions, thus it is important to know details.

Hard to say without having the game. Maybe Zillions is better. ;-)

jolat:
Which version (via Help -> About ...), Cache size, any of the programs pondering?

When a testing period has expired or no user key is present, SMIRF answers immediately (low quality).

Who ever wanted to know, which Gothic Chess program might be the strongest, actually could watch a long-time fight between Gothic Vortex Gold and SMIRF Beta. Average thinking times are floating between 1/2 hour and several hours, thus the playing level is of course impressing. It is also interesting to see the evaluation functions differ sometimes more than one Pawn unit. You will find more on this and also interesting comments at: http://s13.invisionfree.com/Gothic_Chess_Forum/index.php

In contrast to claimed unspecified advantages of Black within Janus Chess I am convinced of White being about 3/10 Pawn units ahead in its beginning. Similar statements have been done by chess players first being faced to some Chess960 / FRC starting arrays. Those statements simply reflect the absence of opening knowledge, thus merely prove to be caused by psychological effects without a real base.

mahavrilla:
Castling is different. It is done symmetrically. The a-side castling is written as O-O instead of O-O-O. As far as I know there are no patents on Janus. The representation of the Archbishop is replaced by a pointed dog, called Janus.