HalfPawn: Maybe, what time frame? But no sorrow, current private beta version probably will not been given to download as done prior, and the spreaded version will expire end of March. Up to now only 10 Euros have been donated for my SMIRF GUI + Engine as a donationware. So there could not be a huge user crowd after this month. If SMIRF, it must have been an old version. Current SMIRF would have played much superior and surely won.

HalfPawn: Not the SMIRF I play. It does something like that and I expect to be losing shortly.

HalfPawn: There are different opinions to piece average exchange values. Current SMIRF values will equalize: R+B+P/3 == Q, so it will depend on additional positional effects, whether such a trade will be recommended or not. To call such trades "dumb" is without any argument. Normally such a trade is not deciding the outcome of a 10x8 game.

P.S.: Another proof for the experts being here. If you really should be interested in testing the real SMIRF, you simply should play serious games with it.

HalfPawn: Let's watch again average piece value proposals. There are two known relevant number sets: a) within SMIRF, b) within Gothic Vortex.

For me it seems to be better implemented acually within SMIRF, watching such games like: http://www.exactachess.com/modules.php?name=Forums&file=viewtopic&p=498#498

SMIRF Engine:
How this game is conclusive i don't understand!?!
Conclusive about what? About piece values? I don't see the point....

Pythagoras: Sometimes captions will give you a hint. Here: if "dumb" moves are sufficient to start a discussion on piece values, it would obviously make more sense to discuss other programs instead of SMIRF, discovering SMIRF being that superior.

SMIRF Engine: .
.
.
I don't understand nothing from what you said!
Anyway i have the feeling that you believe i've said that Smirf plays dumb trades. It wasn't me that said that. But anyway if you want my opinion Smirf makes bad trades. Very bad ones. Of course i don't claim it is doing them often or not i just say it does them from time to time. And i believe that it would be stronger if it didn't make that bad trades....

Pythagoras: Indeed SMIRF still is a beta. Because it is my first playing serious Chess program, there still a lot of weaknesses are within it. Watching the SMIRF project one easily could notice that progresses will be made. But SMIRF's strength actually is already sufficient to fear no 10x8 opponent program. Thus it is not on SMIRF to change its evaluation model.

SMIRF Engine: .
.
.
I have watched the Smirf progress at 10x8 variants all these years and i've said, as you know, that i'm just impressed! You are doing a hell of a job with it that's why i wish you to continue with it and forget all these GO-thing.
The most interesting thing about Smirf for me is its ability to play, always speaking about 10x8 Chess, the opening with a non-computeristic way! It doesn't play the openings like a human but not like computer too. If you leave a computer alone in the opening it will screw up the game. Not Smirf! I guess this has to do with your approach.... Of course many many weaknesses are there in the opening play too but not as many as in all the other 10x8 programs i know....

I guess that Smirf's strength (that was the case in the latest Smirf beta i can play with) at 10x8 variants is sufficient for having no fear against the other programs. But i'm sure you don't speak about 8x8 Chess variants, right?
For example and since you are the only one that can test it, how it can do against Spike 1.1 at FRC?

About GO: Yes the GO-programming world is still unexplored and many innovative ideas can be implement. This obviously will give to the GO programmers the sense of a mystery, that they're swimming into uncharted places with many discoveries waiting them. And this is true. But considering the difficulty of GO-programming i don't think it's a good idea to abandon a successful project like Smirf and dive into the unknown......

Pythagoras: "For example and since you are the only one that can test it, how it can do against Spike 1.1 at CRC?"

P.S.: Everyone could download SMIRF and test it, it is not restricted until the end of this month.

Actually I do not play much 8x8 games against other programs. But because of the results from some special test sets, SMIRF's 8x8 strength seems to be somewhere around 2600 Elo.

Actually SMIRF has come to a stage, where a complete redesign would make sense. But there are very few fans supporting the SMIRF project but much more critics. So I am unsure to take that task.

HalfPawn: As far as I know I am no longer welcomed at gothicchesslive. During the last big GC event there end of 2005 there have been difficulties to rebuild broken games, if both players would not sufficiently cooperate. I preferred to resign then instead to play endless kindergarden games. My attempt to make a proposal how to fairly handle such unlikely situations had been ignored. Instead I have been called someone, who would leave tournaments, if it would not go his way. The truth is, that any victory or loss is losing any worth, if not gained fairly. That is also the reason, why I have resigned some games here at brainking, even though mostly having advantage.

My SMIRF has been available unrestrictedly until end of March. Everybody was able to test and compare it with his own program. That has not been possible for me. No actual 2006 version of ChessV or Gothic Vortex could be used by me for testing purposes. This is showing to everybody, where the fear is residing.

HalfPawn: Why do you do nothing here but taunt players?

SMIRF, I assume you have already decided to ignore HalfPawn. He seems to be here just to stir up trouble.

HalfPawn: "Since you "ran away" from the tournament, ..." No, it has been no tournament, because its rules were not sufficient to cover the raised problems. Instead of completing the rules, pressure has been put on me. So the only method for me to end this chaos had been to resign.

... you should challenge either ChessV ... so is it still alive? On its sourceforge site forum posts will not been answered at all.

... or Vortex ... As a reaction of a current posting of Ed: "You are mentally ill"? Well, the world is obviously strange.

... it was announced that there is a 64-bit parallel processing version of Gothic Vortex ... Well, so much power seems to be necessary to beat a new born single CPU 32-Bit amateur program SMIRF.

SMIRF Engine: The truth is, that any victory or loss is losing any worth, if not gained fairly. That is also the reason, why I have resigned some games here at brainking, even though mostly having advantage.

And i say again:
Why you propose that i didn't play fair.....?
Why you still insist on the wrong statement that Smirf had the advantage at these games....?


Yes seeing SMIRF's 8x8 abilities seperatedly would do harm to its new concept, but i prefer this than having to test games manually to see its strength. After an UCI implementation you will see Smirf appear in many Chess rating lists and tests. I believe this is more important from the possible aforementioned harm..... A massive number of games will help exponentialy you, to improve Smirf!

Pythagoras: One example of resigned (you know the reason) games:
http://brainking.com/de/ArchivedGame?g=1370112

SMIRF evaluates for your side as follows:

02:48.1 (10.01=) -2.000 18...Nd7 (Be6) 19.Bf4 Ne5 20.exf5 Bxf5 21.Ne4 Bxe4 22.Cxe4 Qe6 23.Bxe5 Bxe5 24.bxa5 O-O-O
01:44.5 (09.02=) -2.014 18...Nd7 (Be6) 19.Bf4 Ne5 20.Qxa5 Rxa5 21.bxa5
01:43.2 (09.02+) -2.129 18...Nd7 (Be6) 19.Bf4 Ne5 20.Qxa5 Rxa5 21.bxa5
01:36.3 (09.01=) -2.131 18...Be6 (Nd7) 19.exf5 Bxf5 20.f4 Ng4 21.Ne4 Bxe4 22.Cxe4 axb4 23.Cxd6+ cxd6
01:13.9 (08.47=) -1.910 18...Be6 (Nd7) 19.exf5 Bxf5 20.f4 Ng4 21.Nb7 Qd7 22.Nxa5 Me6 23.Nxc6 Rxa4 24.Nxe7
00:41.7 (08.05=) -1.910 18...Be6 (Ng4) 19.exf5 Bxf5 20.f4 Ng4 21.Nb7 Qd7 22.Nxa5 Me6 23.Nxc6 Rxa4 24.Nxe7
00:29.2 (08.03=) -1.910 18...Be6 (Bd7) 19.exf5 Bxf5 20.f4 Ng4 21.Nb7 Qd7 22.Nxa5 Me6 23.Nxc6 Rxa4 24.Nxe7
00:25.1 (08.01=) -1.910 18...Be6 (d3) 19.exf5 Bxf5 20.f4 Ng4 21.Nb7 Qd7 22.Nxa5 Me6 23.Cd5
00:09.6 (07.02=) -1.926 18...Be6 (d3) 19.Bf4 Bd7 20.Bxe5 Bxe5 21.Nxd7 Mxd7 22.exf5 gxf5 23.Cxc6 Cxc6 24.Qxc6 Qxc6 25.bxa5
00:09.1 (07.02+) -2.000 18...Be6 (d3) 19.exf5 Bxf5 20.Bxh6 ixh6 21.Cxc6+ Nxc6 22.Qxc6+ Qxc6 23.bxa5
00:07.0 (07.01=) -2.002 18...d3 (Be6) 19.Nxd3 Nxd3+ 20.exd3 Qxd3 21.exf5 Bxf5 22.Cxc6+ Cxc6 23.Qxc6+ Kf7 24.Qxc7+ Md7
00:03.0 (06.01=) -1.934 18...d3 (Be6) 19.exd3 axb4 20.Qxa8 Qxc5 21.Cf4 Nf3+ 22.Kd1 Bc3
00:02.6 (05.03=) -1.875 18...d3 (Be6) 19.exd3 Nxd3+ 20.Nxd3 Qxd3 21.exf5 Qxf1+ 22.Rxf1 axb4 23.Qxa8 gxf5
00:01.2 (05.01=) -1.875 18...d3 (Bd7) 19.exd3 Nxd3+ 20.Nxd3 Qxd3 21.exf5 Qxf1+ 22.Rxf1 axb4 23.Qxa8 gxf5
00:01.0 (04.18=) -1.875 18...d3 (Bd7) 19.exd3 Nxd3+ 20.Nxd3 Qxd3 21.exf5 Be6
00:00.9 (04.03=) -1.875 18...d3 (Ra7) 19.exd3 Nxd3+ 20.Nxd3 Qxd3 21.exf5 Be6
00:00.6 (04.01=) -1.875 18...d3 (Ng4) 19.exd3 Nxd3+ 20.Nxd3 Qxd3 21.exf5 Be6
00:00.3 (03.01=) -1.875 18...d3 (Ng4) 19.exd3 Nxd3+ 20.Nxd3 fxe4
00:00.2 (02.25=) -1.875 18...d3 (Ra7) 19.Nxd3 Nxd3+ 20.exd3 Qxd3 21.exf5 Me6+
00:00.0 (02.20=) -1.875 18...d3 (=/=) 19.Nxd3 Nxd3+ 20.exd3 Qxd3 21.exf5 Me6+
00:00.0 (02.00?) -1.621 18...Qxc5 (=/=)

SMIRF+UCI: Already I have so much ideas and improvements in my head, that I would not yet need help for that. Nevertheless I could use extern experiences and strength estimations. It seems more important (if I would do anything big at all), to make SMIRF 64 Bit aware. That is not easy because its GUI is built with Borland Builder, which is still not supporting 64 Bit for C++.

SMIRF Engine: .
.
.
Yes Smirf shows an advantage for itself but as i already said and "proved" (remember the last game i was black and Smirf believed it was winning with even +5.000 scores but finally lost) this doesn't mean nothing! Smirf simply can't see the inevitable.... It's far away from its horizon!


While we were speaking about Smirf, i took the chance and played just for fun 2 games against Rybka 1.1. Yeah i know the test against the monster is not fair but anyway.
Time control was 10 seconds per move with 256 Hash for both. In the first game i played the first 4 opening moves for Smirf according to mainbook.Arena while in the second game Smirf played the opening by itself. Result of course 2-0 and i didn't do that for any other reason just for fun....


[Event "SmirfGUI Computerchess Game"]
[Site "USER-403C8545D2"]
[Date "2006.03.31"]
[Round "?"]
[White "Smirf 146"]
[Black "Rybka 1.1"]
[Result "0-1"]
[Annotator "Test SMIRF until 2006-03-31"]

{The time limit has been passed.} 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 d6 5.
Nc3 {(11.02=) +0.773} Nf6 6. Be3 {(10.01) +0.684} a6 7. a4 {(09.00) +1.047} Be7
8. h3 {(09.02+) +0.906} O-O 9. Bd3 {(09.00) +0.936} Bd7 10. Qf3 {(10.00)
+0.805} Nc6 11. Nxc6 {(09.01) +0.715} Bxc6 12. O-O {(08.18=) +0.850} Nd7 13.
Qh5 {(08.01) +0.717} Ne5 14. f4 {(08.18=) +0.469} Nxd3 15. cxd3 {(10.00)
-0.084} g6 16. Qg4 {(09.01) +0.062} d5 17. exd5 {(10.00) -0.125} Bxd5 18. a5
{(09.01) -0.105} Rc8 19. Bb6 {(09.01) -0.014} Bc5+ 20. Bxc5 {(11.01=) -0.693}
Rxc5 21. f5 {(10.21) -0.838} Bc6 22. fxg6 {(09.29) -0.881} Rg5 23. gxh7+
{(11.01) -2.066} Kh8 24. Qe2 {(11.01) -2.508} Rxg2+ 25. Qxg2 {(12.00) -2.586}
Bxg2 26. Kxg2 {(12.00) -2.693} Qxd3 27. Ra4 {(11.05) -2.602} Qxh7 28. Rf3
{(11.00) -2.703} Rd8 29. Kh1 {(11.00) -2.541} Rd3 30. Rxd3 {(13.00) -3.557}
Qxd3 31. Rh4+ {(12.01) -3.973} Kg7 32. Rg4+ {(11.01) -3.973} Kf8 33. Kh2
{(12.00) -3.996} Qf5 34. Rg3 {(12.01) -4.250} Qxa5 35. Rg2 {(11.01) -4.404}
Qe5+ 36. Kh1 {(12.01) -5.215} b5 37. Rg4 {(11.23=) -5.377} a5 38. Rg1 {(12.25)
-5.646} b4 39. Na4 {(11.21=) -6.848} Qd5+ 40. Kh2 {(13.00) -7.980} Qa2 41. Rg3
{(12.05=) -8.289} 0-1


[Event "SmirfGUI Computerchess Game"]
[Site "USER-403C8545D2"]
[Date "2006.03.31"]
[Time "01:01:16"]
[Round "?"]
[White "Rybka 1.1"]
[Black "Smirf 146"]
[Result "1-0"]
[Annotator "Test SMIRF until 2006-03-31"]

{The time limit has been passed.} 1. d4 d5 {(10.03=) -0.195} 2. c4 dxc4
{(10.01) +0.039} 3. e4 Nc6 {(10.01) +0.420} 4. Nf3 Nf6 {(10.03) +0.242} 5. Nc3
Bg4 {(10.08) +0.047} 6. Be3 e5 {(10.12) +0.211} 7. d5 Bxf3 {(10.15=) -0.289} 8.
gxf3 Na5 {(10.00) -0.383} 9. Qa4+ c6 {(10.01) -0.766} 10. b4 b5 {(11.01=)
-0.160} 11. Qxa5 Qxa5 {(12.00) -0.432} 12. bxa5 Bb4 {(11.01) -0.662} 13. Bd2
cxd5 {(12.01=) -0.842} 14. Nxb5 Bxd2+ {(12.26) -1.191} 15. Kxd2 Rd8 {(12.00)
-1.191} 16. Nxa7 dxe4+ {(11.00) -1.598} 17. Kc2 Rd5 {(12.01) -0.969} 18. a6 Rc5
{(12.00) -0.418} 19. Rb1 O-O {(12.08) -0.123} 20. fxe4 Nxe4 {(12.02) -0.436}
21. Bg2 f5 {(12.01) -0.658} 22. Rb5 Rc7 {(13.01) -0.957} 23. Rb7 Rc5 {(13.00)
-1.312} 24. Nb5 Nxf2 {(12.00) -1.131} 25. Rg1 Kh8 {(11.01) -1.746} 26. a7 e4
{(10.07=) -3.311} 27. Rb1 Rcc8 {(11.05) -4.039} 28. Nd6 Ra8 {(11.19) -4.559}
29. Rb8 g6 {(11.01) -10.23} 1-0

Pythagoras: Time control was 10 seconds per move with 256 Hash for both

Well, SMIRF is not a blitzer by design. Nevertheless it cannot stand against such 8x8 programs yet, even version MS-158. Remember, it is my first approach. I am already happy to gain such advantages as against you. ;-)

SMIRF Engine: .
.
.
LOL! Such advantages would definitely mean a win at 8x8 Chess for Smirf against me, but at 10x8 where material is less important than having an attack, i can plan better than Smirf so even +5.000 evaluations can't be trusted....


Anyway i will play tomorrow with some weaker opponents like Fritz 5.32, Comet B68, List 504 and Horizon 4.1 in order to find out of where approximately is its strength.
Considering the limited time i will have for this small test, what time controls do you suggest for this? I want to use the _shortest_ time controls possible that you believe Smirf will play best(as it is not a blitzer as you said).
Since tomorrow is the last day i can use Smirf i want to at least explore it a bit better....

Pythagoras: Well, my current SMIRF MS-158 is again noticable stronger than the still to be download version MS-156. Real fans could make a serious project donation e.g. via PayPal to get a permanent key, of course valid also for coming versions. So additional testing is not at all absolutely impossible. I would suggest a testing at least with SMIRF's preset "rapid" timing. Playing Chess960 games would be a fair approach.