Only BUP wins

[vb4]

Crafty announces mate in 2,724.

Well, that certainly wins the prize for the deepest mate announcement ever.
And certainly, it is a mate in 2,724 or less (plus it has the right key move)

It probably should have been 27.24 and the decimal was just left off. Just my 2 cents.

Sure, but how do you make .24 of a move?

Why didnt I think of that <s>

Bob, please comment on this:

New game
4Q3/Pq4pk/5p1p/5P1K/6PP/8/8/8 w - - 0 1

Analysis by Crafty 23.1 64 bit:

1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qa2 4.Bb7 Qd2 5.Bf3 Qd4 6.Be2 Qd5 7.Qg6+ Kg8
+- (5.00) Depth: 11 00:00:00 205kN
1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qa2 4.Bb7 Qd2 5.Bf3 Qd4 6.Be2 Qd5 7.Qg6+ Kg8
+- (5.00) Depth: 11 00:00:00 243kN
1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qd3 4.Qf7 Qd8 5.Qd7 Qb6 6.Bd5 Qd4
+- (5.06) Depth: 12 00:00:00 408kN
1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qd3 4.Qf7 Qd8 5.Qd7 Qb6 6.Bd5 Qd4
+- (5.06) Depth: 12 00:00:00 455kN
1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qa2 4.Qf8 Qc4 5.Be8 Qg8 6.Qxg8+ Kxg8 7.Kg6 Kf8 8.Bf7 Ke7 9.Kxg7 Kd6
+- (5.27) Depth: 13 00:00:00 569kN
1.a8B Qb3 2.Be4 Qc4 3.Bc6 Qa2 4.Qf8 Qc4 5.Be8 Qg8 6.Qxg8+ Kxg8 7.Kg6 Kf8 8.Bf7 Ke7 9.Kxg7 Kd6
+- (5.27) Depth: 13 00:00:00 626kN
1.a8B
+- (5.67) Depth: 14 00:00:00 801kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bd5 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5
+- (6.45) Depth: 14 00:00:00 870kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bd5 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5
+- (6.45) Depth: 14 00:00:00 957kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bd5 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5
+- (6.43) Depth: 15 00:00:00 1220kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bd5 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5
+- (6.43) Depth: 15 00:00:00 1516kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bb3 Kh7 7.Bd5 Kh8 8.Kg6 h5 9.Kxh5 Kh7
+- (6.65) Depth: 16 00:00:00 1985kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bb3 Kh7 7.Bd5 Kh8 8.Kg6 h5 9.Kxh5 Kh7
+- (6.65) Depth: 16 00:00:00 2494kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bb3 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5 Kh7 10.Bg6+ Kg8
+- (6.45) Depth: 17 00:00:01 3756kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bb3 Kh7 7.Bf7 Kh8 8.Kg6 h5 9.Kxh5 Kh7 10.Bg6+ Kg8
+- (6.45) Depth: 17 00:00:02 7978kN
1.a8B
+- (6.85) Depth: 18 00:00:02 8136kN
1.a8B
+- (7.85) Depth: 18 00:00:02 8384kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kg8 6.gxf6 gxf6 7.Kxh6 Kf8 8.h5 Ke7 9.Kg6 Kd6 10.Kxf6
+- (8.58) Depth: 18 00:00:03 10313kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kg8 6.gxf6 gxf6 7.Kxh6 Kf8 8.h5 Ke7 9.Kg6 Kd6 10.Kxf6
+- (8.58) Depth: 18 00:00:05 13044kN
1.a8B
+- (8.98) Depth: 19 00:00:05 14019kN
1.a8B
+- (9.98) Depth: 19 00:00:05 14156kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Ke8 8.Kxf6 h5 9.Bf7+ Kf8 10.Bxh5 Kg8 11.Bf7+ Kf8 12.Bg8
+- (10.42) Depth: 19 00:00:09 19960kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Ke8 8.Kxf6 h5 9.Bf7+ Kf8 10.Bxh5 Kg8 11.Bf7+ Kf8 12.Bg8
+- (10.42) Depth: 19 00:00:13 26205kN
1.a8B
+- (10.82) Depth: 20 00:00:14 26994kN
1.a8B
+- (11.82) Depth: 20 00:00:14 27180kN
1.a8B
+- (13.82) Depth: 20 00:00:14 27455kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 20 00:00:22 47334kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 20 00:00:32 57920kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 21 00:00:45 86706kN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 21 00:01:05 117mN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 22 00:01:31 182mN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 22 00:02:13 261mN
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 fxg5 6.hxg5 hxg5 7.Kxg5
-+ (-#2724) Depth: 23 00:03:02 463mN

[BBauer]

Here Crafty 23.2 gives a mate in 71.

starting thread 1
1 0.01 -6.11 1. ... b1=B (3Knps)
1-> 0.01 -6.11 1. ... b1=B
2 0.01 -1 1. ... b1=B!
2 0.01 -3 1. ... b1=B!
2 0.01 -M 1. ... b1=B!
2 0.01 -Mat71 1. ... b1=B <EGTB>
2-> 0.01 -Mat71 1. ... b1=B <EGTB>
3 0.01 -Mat71 1. ... b1=B <EGTB>
3-> 0.01 -Mat71 1. ... b1=B <EGTB>
4 0.01 -Mat71 1. ... b1=B <EGTB>
4-> 0.01 -Mat71 1. ... b1=B <EGTB>
time=0.01 mat=-1 n=41 fh=100% nps=1.0M
extensions=0 qchecks=0 reduced=0 pruned=0
predicted=0 evals=1 50move=0 EGTBprobes=1 hits=1
SMP-> splits=4 aborts=0 data=1/128 elap=0.01
terminating SMP processes.

mate in 71 moves.

Black(1): b1=B
time used: 0.01

Crafty doesn't give the moves. Is this feature gone?
kind regards
Bernhard

[M ANSARI]

Again I say that this game is totally insignificant. Yes in the final position under promotion to a bishop seems to get an extra .5 point, but there were tons of positions before that where the engine would have found much better moves that would have negated the need for the final position to arise. You are choosing a game by humans that by quickly analysing with Rybka you can see that it is way below the level of a top engine, with very weak moves being played by both sides. I want to see a game where an engine against another engine or an engine against a human, lost a half point due to bishop under promotion.

Don't get me wrong, I don't mind having this "feature" added in analysis mode, but really if it hurts performance even a tiny bit, it sure as hell is not worth implementing it in play mode.

If I go through the games of say a 500 game tourney, there are many many other things that continually pop up that are much much more important than bishop under promotion, and actually cost points ... this item simply never comes up and is not an issue IMHO.

[BBauer]

I would never buy a chess program that has not implemented the full set of chess rules. The argument 'this never occurs in real games' is old but poor. The fascinating thing in chess is that it is always new. New positions, new ideas. By the way, I do not believe that implementing bishop promotion costs anything significant.
kind regards
Bernhard

[Cubeman]

It still has implemented the full set of rules, so no reason not to buy.The rules state that when pawn reaches the 8th rank it has to be promoted to either Q,R,B,or N.It is completely your choice.Rybka accepts BU by the opponent so no worries there.
If implementing BU costs elo then I would rather do without BU.I suppose each person has their own view and can vote with thier wallet.

[mhalstern]

Vas knows what he is doing. Not allowing BUP probably adds a few elo points to Rybka. He should at least, offer BUP as a UCI option, with a readme disclaimer about the loss.

[BBauer]

Bishop under promotion is a possibility just like rook under promotion etc.
If promotion to bishop is the best move the engine should find it.

You wrote: Vas knows what he is doing.

That is a question.
In New In Chess 2009/7 you see Rybka 4 Aquarium advertised.
You can read:
"Rybka 4 Aquarium - Pre-Order Now!"
Rybka 4 is another big step forward for the world's strongest chess engine".

Certainly if he hides the true nps and the true depth he will know what he is doing

You wrote: Not allowing BUP probably adds a few elo points to Rybka.

This looks like nonsense to me. How many elo points?

kind regards
Bernhard

[beachknight]

DF11 is also fast.

Code: Select all

16: 4Q3/Pq4pk/5p1p/5P1K/6PP/8/8/8
4Q3/Pq4pk/5p1p/5P1K/6PP/8/8/8 w - - 0 1

Analysis by Deep Fritz 11:

1.a8B Qb3 2.Be4 Qc4 3.Qe6 Qb4 4.Bd5 Qf8 5.Qc6 Qb8 6.Qa8 Qxa8 7.Bxa8 Kg8 8.Bd5+ Kh7 9.Bc4 Kh8 10.Kg6 h5 11.Kxh5 Kh7 12.Bd5 Kh8 13.g5 fxg5 14.hxg5 
  +-  (5.61)   Depth: 19/28   00:00:01  856kN, tb=392
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Kg6 Kf8 6.Be6 Ke7 7.Kxg7 Ke8 8.Kxf6 Kf8 9.Bf7 h5 10.Bxh5 Kg8 11.Bf7+ 
  +-  (8.83)   Depth: 20/31   00:00:04  1535kN, tb=1472
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Kg6 Kf8 6.Be6 Ke7 7.Kxg7 Kd6 8.Kxf6 Kc6 9.Ke7 h5 10.gxh5 Kc5 11.f6 Kd4 
  +-  (9.19)   Depth: 21/33   00:00:07  2024kN, tb=2662
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.g5 Kf7 6.gxh6 Kf8 7.Kg6 Ke7 8.hxg7 Kd6 9.g8Q Ke5 10.h5 Kd4 11.Kxf6 Ke3 12.h6 Kf3 
  +-  (19.93)   Depth: 22/37   00:00:21  4258kN, tb=8292
1.a8B Qb3 2.Bc6 Qg8 3.g5 Kh8 4.gxf6 Qxe8+ 5.Bxe8 gxf6 6.Kxh6 
  +-  (#24)   Depth: 23/35   00:00:51  14542kN, tb=25480
1.a8B Qb3 2.Bc6 Qg8 3.g5 Kh8 4.gxf6 Qxe8+ 5.Bxe8 gxf6 6.Kxh6 
  +-  (#24)   Depth: 23/35   00:00:52  15118kN, tb=25786

(, chessengines20 30.03.2010)

[beachknight]

Crafty 23.01 cb is faster

Code: Select all

16: 4Q3/Pq4pk/5p1p/5P1K/6PP/8/8/8
4Q3/Pq4pk/5p1p/5P1K/6PP/8/8/8 w - - 0 1

Analysis by Crafty 23.01 cb:

1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kh8 6.Bf7 Kh7 7.Bd5 Kh8 8.Kg6 h5 9.Kxh5 Kh7 
  +-  (7.05)   Depth: 17/26   00:00:01  3721kN, tb=2142
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kh8 6.gxf6 gxf6 7.Kg6 Kg8 8.Kxf6 Kh8 9.Bf7 Kh7 10.Bg6+ Kg8 
  +-  (7.25)   Depth: 17/30   00:00:01  4771kN, tb=2341
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kh8 6.gxf6 gxf6 7.Kg6 Kg8 8.Kxf6 Kh8 9.Bf7 Kh7 10.Bg6+ Kg8 
  +-  (7.65)   Depth: 18/29   00:00:03  6479kN, tb=6087
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kg8 6.Bf7+ Kf8 7.gxh6 gxh6 8.Kxh6 Ke7 9.Kg7 Kd7 10.Kxf6 Kd6 
  +-  (8.02)   Depth: 18/33   00:00:05  8607kN, tb=6560
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kg8 6.Bf7+ Kf8 7.gxh6 gxh6 8.Kxh6 Ke7 9.Kg7 Kd7 10.Kxf6 Kd6 
  +-  (8.42)   Depth: 19/32   00:00:08  11620kN, tb=13976
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 Kg8 6.Bf7+ Kf8 7.gxh6 gxh6 8.Kxh6 Ke7 9.Kg7 Kd7 10.Kxf6 Kd6 
  +-  (9.42)   Depth: 19/32   00:00:08  11923kN, tb=14641
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Kd6 8.Kxf6 Kc5 9.Kg7 Kd4 10.Kxh6 Ke5 11.Bd5 
  +-  (10.02)   Depth: 19/35   00:00:14  16196kN, tb=14761
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Kd6 8.Kxf6 Kc5 9.Kg7 Kd4 10.Kxh6 Ke5 11.Bd5 
  +-  (10.42)   Depth: 20/30   00:00:23  21765kN, tb=31008
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Kd6 8.Kxf6 Kc5 9.Kg7 Kd4 10.Kxh6 Ke5 11.Bd5 
  +-  (11.42)   Depth: 20/30   00:00:23  22036kN, tb=32075
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.Qxg8+ Kxg8 5.Be6+ Kf8 6.Kg6 Ke7 7.Kxg7 Kd6 8.Kxf6 Kc5 9.Kg7 Kd4 10.Kxh6 Ke5 11.Bd5 
  +-  (13.42)   Depth: 20/30   00:00:23  22420kN, tb=32196
1.a8B Qb3 2.Bc6 Qc4 3.Bd7 Qg8 4.g5 Qxe8+ 5.Bxe8 hxg5 6.hxg5 fxg5 7.Kxg5 
  +-  (#22)   Depth: 20/36   00:00:38  39919kN, tb=32714

(, chessengines20 30.03.2010)

[beachknight]

this is a tablebase position

DF11:

Code: Select all

17: 8/8/5b2/8/K7/8/kp6/3N4 b - - 1
8/8/5b2/8/K7/8/kp6/3N4 b - - 0 1

Analysis by Deep Fritz 11:

1. -+  (-#71): 63...b1B 
2. =  (0.00): 63...Ka1 
3. =  (0.00): 63...Kb1 
4. =  (0.00): 63...b1Q 
5. =  (0.00): 63...b1R 
6. =  (0.00): 63...b1N 
7. =  (0.00): 63...Be5 
8. =  (0.00): 63...Bd4 
9. =  (0.00): 63...Bc3 
10. =  (0.00): 63...Bg5 
11. =  (0.00): 63...Bh4 
12. =  (0.00): 63...Bg7 
13. =  (0.00): 63...Bh8 
14. =  (0.00): 63...Be7 
15. =  (0.00): 63...Bd8 

(, chessengines20 30.03.2010)