Deep Blue Game 2

General discussion about computer chess...
Post Reply
induktio
Posts: 4
Joined: Mon Jan 02, 2012 5:40 am

Deep Blue Game 2

Post by induktio » Tue Jan 03, 2012 9:58 am

I recently stumbled across the wikipedia article on the second game of the match:
However, it was discovered in 2007 by Internet analysis[7] that after 45...Qe3 46.Qxd6 Re8 47.Qd7+ Re7 48.Qc6 Qxe4 49.d6 Qd3+ 50.Kg1 Re8 51.Ra1 there is no perpetual check and the attack continues. The critical move being for White to force Black's rook on e7, which is later attacked by 49.d6, but also for White to retract his queen to c6, which can later move to g2, if necessary, to block the perpetual check. Various other lines have been analyzed, such as 48...Qc1+ or 48...Qf4+, but White (Deep Blue) invariably ends up with a superior position. Given this analysis, all of Deep Blue's moves make sense; it is not apparent that the computer made any mistake.
To me this seems rather peculiar, as I understood the consensus over the years was that the final position was a draw. The linked video is hardly convincing either. So here's the position after 45. Ra6 for reference:
1r6/5kp1/RqQb1p1p/1p1PpP2/1Pp1B3/2P4P/6P1/5K2 b - - 14 45
I ran some analysis of the suggested line after 49. d6 (below) as it seems engines have differences over the next best move.
8/4rkp1/R1QP1p1p/1p2pP2/1Pp1q3/2P4P/6P1/5K2 b - - 0 49
Crafty 23.4 x64:

Code: Select all

9-> 537:37   1.17   1. ... Qxf5+ 2. Kg1 Rd7 3. Qxb5 Qe6
		    4. Qc5 Kg6 5. Rc6 e4 6. b5 e3 7. Qxc4
		    Qe5 8. Qg4+ Kh7 9. Qe2 Rb7 10. b6 Qe8
		    11. Rc5 Rxb6 12. Rd5 f5 13. c4 Rb1+
		    14. Kh2 f4 15. c5 Qe4 16. Rd1 Rxd1
		    17. Qxd1
30   721:04   1.41   1. ... Qxf5+ 2. Kg1 Rd7 3. Qxb5 Qe6
		    4. Qc5 Kg6 5. Rc6 e4 6. b5 e3 7. Qxc4
		    Qxc4 8. Rxc4 Rxd6 9. Rd4 Rb6 10. c4
		    Kf5 11. Kf1 Ke5 12. Rd7 f5 13. Ke2
		    Rg6 14. Kxe3 Rxg2 15. Rd5+ Ke6 16.
		    Kd4 Re2 17. Kc5
30   1127:12   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 <HT>
30-> 1127:12   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 <HT>            (s=2)
31   1700:39   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 8. Qc7 Kg8 <HT>
31-> 1982:42   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 8. Qc7 Kg8 <HT>
The score seems kinda drawish and the material is still dead even, so I don't see any win here for white. Shouldn't the wiki chapter be entirely removed for misinformation?

hyatt
Posts: 1242
Joined: Thu Jun 10, 2010 2:13 am
Real Name: Bob Hyatt (Robert M. Hyatt)
Location: University of Alabama at Birmingham
Contact:

Re: Deep Blue Game 2

Post by hyatt » Wed Jan 04, 2012 3:03 am

induktio wrote:I recently stumbled across the wikipedia article on the second game of the match:
However, it was discovered in 2007 by Internet analysis[7] that after 45...Qe3 46.Qxd6 Re8 47.Qd7+ Re7 48.Qc6 Qxe4 49.d6 Qd3+ 50.Kg1 Re8 51.Ra1 there is no perpetual check and the attack continues. The critical move being for White to force Black's rook on e7, which is later attacked by 49.d6, but also for White to retract his queen to c6, which can later move to g2, if necessary, to block the perpetual check. Various other lines have been analyzed, such as 48...Qc1+ or 48...Qf4+, but White (Deep Blue) invariably ends up with a superior position. Given this analysis, all of Deep Blue's moves make sense; it is not apparent that the computer made any mistake.
To me this seems rather peculiar, as I understood the consensus over the years was that the final position was a draw. The linked video is hardly convincing either. So here's the position after 45. Ra6 for reference:
1r6/5kp1/RqQb1p1p/1p1PpP2/1Pp1B3/2P4P/6P1/5K2 b - - 14 45
I ran some analysis of the suggested line after 49. d6 (below) as it seems engines have differences over the next best move.
8/4rkp1/R1QP1p1p/1p2pP2/1Pp1q3/2P4P/6P1/5K2 b - - 0 49
Crafty 23.4 x64:

Code: Select all

9-> 537:37   1.17   1. ... Qxf5+ 2. Kg1 Rd7 3. Qxb5 Qe6
		    4. Qc5 Kg6 5. Rc6 e4 6. b5 e3 7. Qxc4
		    Qe5 8. Qg4+ Kh7 9. Qe2 Rb7 10. b6 Qe8
		    11. Rc5 Rxb6 12. Rd5 f5 13. c4 Rb1+
		    14. Kh2 f4 15. c5 Qe4 16. Rd1 Rxd1
		    17. Qxd1
30   721:04   1.41   1. ... Qxf5+ 2. Kg1 Rd7 3. Qxb5 Qe6
		    4. Qc5 Kg6 5. Rc6 e4 6. b5 e3 7. Qxc4
		    Qxc4 8. Rxc4 Rxd6 9. Rd4 Rb6 10. c4
		    Kf5 11. Kf1 Ke5 12. Rd7 f5 13. Ke2
		    Rg6 14. Kxe3 Rxg2 15. Rd5+ Ke6 16.
		    Kd4 Re2 17. Kc5
30   1127:12   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 <HT>
30-> 1127:12   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 <HT>            (s=2)
31   1700:39   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 8. Qc7 Kg8 <HT>
31-> 1982:42   1.21   1. ... Qd3+ 2. Kg1 Re8 3. Ra1 h5 4.
		    Qc7+ Kg8 5. d7 Rf8 6. Qc8 Kf7 7. Rf1
		    Qd2 8. Qc7 Kg8 <HT>
The score seems kinda drawish and the material is still dead even, so I don't see any win here for white. Shouldn't the wiki chapter be entirely removed for misinformation?

Trust me, the position was a draw. We burned no telling how many computer hours that night. And we had lots of GM players helping (Bruce Moreland and I burned most of the time it seems). And when told, Kasparov spotted it almost instantly. The problem is always a perpetual here, not just an "equal position". Computers could simply not see the combination back then. Not sure about today, even though we are far faster... I certainly doubt any will see the draw in OTB time limits, maybe in hours or overnight...

induktio
Posts: 4
Joined: Mon Jan 02, 2012 5:40 am

Re: Deep Blue Game 2

Post by induktio » Wed Jan 04, 2012 7:04 am

hyatt wrote: Trust me, the position was a draw. We burned no telling how many computer hours that night. And we had lots of GM players helping (Bruce Moreland and I burned most of the time it seems). And when told, Kasparov spotted it almost instantly. The problem is always a perpetual here, not just an "equal position". Computers could simply not see the combination back then. Not sure about today, even though we are far faster... I certainly doubt any will see the draw in OTB time limits, maybe in hours or overnight...
Sure, it's a draw because of the perpetual. It seems one can shuffle the rooks and queens in so many ways without advancing anywhere, so maybe that's why computers have trouble figuring it out?

It's kind of amazing how long misinformation can persist on Wikipedia, now the chapter is removed though.

Post Reply