For me it was in the early 80's when Frans and I just had started. With one of Frans' programs that was sold commercially the program could castle its king out of trouble when it had a rook on A2 and a king on E2. Or Ra3 and Ke3. Or Ke2 and Rh2 etc.
And all your troubles are over.
The first is just taking advantage of a Critter 1.4's learning flaw in where the user can teach it that any line is best by using excluding moves, allowing the scenario that shows the opening position was mate in 3:
Critter 1.4 Session File
50/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (76.059) 490
51/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (78.487) 493
52/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (80.915) 496
53/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (83.343) 499
54/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (85.771) 504
55/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (88.199) 506
56/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (90.627) 506
57/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (93.055) 508
58/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (95.483) 507
59/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (97.911) 497
60/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (100.339) 496
61/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (102.767) 498
62/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (105.195) 500
63/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (107.623) 505
64/04 0:00 +M3 1.e3 f6 2.Bc4 g5 3.Qh5+ (110.051) 507
But that's dull compared to the next one. In the process of going from Rybka 3 to Rybka 4 Beta, Persistent Hash functionality was broken, so the engine would crash when trying to use it. Eventually I found a workaround, so that Rybka 4 Beta wouldn't crash when turning ON persistent hash. I was very excited that I managed to make it work, but then...
Rybka 4 Beta: Persistent Hash ON
10.00 0:00 -81.78 1.Bf3 1
11.00 0:00 -81.78 1.Bf3 1
12.00 0:00 -81.78 1.Bf3 1
13.00 0:00 -81.78 1.Bf3 1
14.00 0:00 -81.78 1.Bf3 1
15.00 0:00 -81.78 1.Bf3 1
16.00 0:00 -81.78 1.Bf3 1
16.00 0:04 +M10 1.Bf3 (1.998.848) 465
Yeah, Bf3 is a move that mates in 10 from the opening position, who would have thought it?!
You just need to be a bit flexible with the "Bishop in squares of the same color" rule.
2)I think that it may be interesting to see the mating line after Bf3 because I see no forced mate for white in 10 moves.
Maybe white can win in 10 moves if white continue to play not by the chess rules but we need to see some line of 19 plies.
> How do you know that Rybka meant Bc1-f3 and not Bf1-f3 or maybe some bishop that is not on the board to f3?
I assume the engine sent the GUI c1f3 with a mate in 10 score (note: I didn't make a picture, that part of the board was actually what the GUI was showing). But indeed, Vas claimed this was a display bug from the GUI, and that it should have shown the f1 Bishop jumping over the pawn (thus the quote on my sig "It's easier for a bishop to jump over a pawn than to change square color!").
> I think that it may be interesting to see the mating line after Bf3 because I see no forced mate for white in 10 moves.
Yes if the pieces can make such illegal moves
Anyway, don't over-think it, the engine was just seeing garbage.
When engines go psycho.
Black: Ka8, Nb8, Ba7
This is a Modul classic of the 1990's. MChess played 1. Nxb8 here, announcing mate in 3. But in fact 1. Qxb8+ Bxb8 2. Nb6 is mate in 2. If you execute 1. Qxb8+ however, MChess claimes a draw because of insufficient material after 1... Bxb8. If you execute that move, it will play a random move like 2. Nf8, missing the mate in 1 (because it thinks every move is 0.00). Finally, when given the mating move 2. Nb6, MChess went beserk and the main line showed something like 2. ... -/--. Priceless Modul comment: 'A very weird move, indeed! But do you know a better one?'.
It is very funny to see a few of the current top programs also missing this. F.e. Houdini 1.5 has exactly the same behaviour (without TB of course).
try to setup a position with more kings and look if you can move and check mate the black king and the GUI allows the result
without GUI you can still talk with an UCI engine ...
Just Start the UCI engine and give the commands which a GUI might have given in the DOS console you see there...
...And Engine responds, to you!! ... Amazing!!
To get a taste of what I am saying just start a bare UCI engine by double clicking on it and type go infinite and see the infinite analysis happening in the DOS console in front of your eyes!!
If we know UCI protocol we might be able to play with an UCI engine in the DOS console itself!
This is Very New and Funny stuff for me!:)
> If we know UCI protocol we might be able to play with an UCI engine in the DOS console itself!
If the console doesn't respond to your commands properly no chance the GUI will :-)
*This actually took me a while as I first assumed the physical board was out of step with its internal board (no screen, just the moves displayed on an old fashoned calculator like display), you had to ask where each piece was to verify the board - which was slooow. Once the position was verified I took the king and the computer just carried on regardless!?!
This is reminiscent of the fight between King Arthur and The Black Knight in the Holy Grail. After Arthur cuts off the Black Knights arms and legs, the knight calls after him as he's leaving "All right, we'll call it a draw".
i dont know if i must laugh or cry...for your female friend,must be understood
She hung her knight and I captured it mentioning that it was hanging. She replied "I didn't want that knight". I was so sad for the knight
girls and chess must have a restrain order between them
Marry me Tatiana Kosintseva!! :-P
this is a proposition of marriage? Tatiana,wherever you are,answer him...and invite me to the wedding
> After half a minute of concentration, she took my pawn from d5 and captured hers on c6.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill