Author: Heiner Marxen
Date: 03:46:34 08/18/01
Go up one level in this thread
On August 13, 2001 at 15:17:48, leonid wrote:
>On August 13, 2001 at 11:53:23, leonid wrote:
>
>>Hi!
>>
>>This position is easy:
>>
>>B2N1qr1/B2Qb2q/Q2Rq3/Q2Q1q2/Q2Q2q1/Q2Q3n/R2Qq2k/1KN2rq1 w - -
>>
>>Actually original position was this that I changed in order of having more
>>accessible one.
>>
>>[D]B2B1qr1/N2Qb2q/Q2Rq3/Q2Q1q2/Q2Q2q1/Q2Q3n/R2Qq2k/1KN2rq1 w - -
>>
>>This last position look like very innocent, since it have initially only 93
>>moves, but take care! When I put it on one professional program, where special
>>mate solving engin was asked to do the work, response was deadly. It reacted on
>>"US" script like true fanatic from the Middle East. Beside committing
>>Selbstmord-Anschlag on the spot it killed my Windows ME as well.
>>
>>Please indicate your result.
>>
>>Thanks,
>>Leonid.
Hello Leonid!
Finally, after more than 78 hours (!) on my K7/600 (350 MB hash) Chest
provides two PVs for two key moves solving this as a mate in 9:
PV: Qxg1+ Qxg1 Qxe2+ Qff2 Qxf2+ Qgxf2 Q7xe6 Rxc1+ Qxc1 Qxd3+ Qaxd3 Qf3 Qxh3+
Kxh3 Qh1+ Qh2 Qxh2#
PV: Rxe6 Rxc1+ Qaxc1 Qxc1+ Qxc1 Qxd3+ Qaxd3 Qxd3+ Qxd3 Qb4+ Q4xb4 Qf3 Rexe2+ Rg2
Rxg2+ Qxg2 Rxg2#
1.Qxg1+ has been given by Gian-Carlo, already, but 1.Rxe6 (no check!) comes
as a surprise. Also, the behaviour of Chest comes as a surprise to me,
especially the large time necessary for the final depth=9 as compared to the
mere 45 minutes for the depth=8:
depth time EBF[T] EBF[N] hash-speed
# 1 0.00s 0kN 0.87 1- 0
# 2 0.00s 0kN 1.00 1- 0
# 3 0.03s 2kN [ 19.55] 0.96 94- 0
# 4 0.49s [ 16.33] 26kN [ 12.46] 1.02 1721- 0
# 5 5.79s [ 11.82] 272kN [ 10.44] 1.14 24491- 0
# 6 39.83s [ 6.88] 1741kN [ 6.40] 1.40 171448- 0
# 7 254.16s [ 6.38] 11262kN [ 6.47] 1.70 1091347- 1
# 8 2375.96s [ 9.35] 109788kN [ 9.75] 1.93 9380661- 1310495
# 9 282572.74s [118.93] 6162420kN [ 56.13] 1.76 889707645- 880959744
^^^^^^^ (this one is too small)
The last node count is the sum of several 32 bit counters, some of which
did overflow/wrap. Therefore the sum is too small, and the EBF by nodes
is also too small for depth=9.
But the times are accurate, and EBF=119 for the last depth is clearly much
larger than wanted.
The refutation table of Chest:
refu 1: Q5e5+ Qfxe5 [ 8-]
refu 2: Qg2+ Q4xg2 [ 8-]
refu 3: Qh1+ Qxh1 [ 8-]
refu 4: Q4f4+ Rxf4 [ 8-]
refu 5: Q4e5+ Qexe5 [ 8-]
refu 6: Qf2+ Qfxf2 [ 8-]
refu 7: Qxg1+ Qxg1 [ 7-]
solu 1: Qxg1 [ 8+]
solu 2: Rxg1 [ 8+]
solu 3: Kxg1 [ 3+]
solu 4: Nxg1 [ 1+]
refu 8: Qxe2+ Qgxe2 [ 8-]
refu 9: Q2f4+ Qfxf4 [ 8-]
refu 10: Nb5 Qxd3+ [ 8-]
refu 11: Nc8 Qxd3+ [ 8-]
refu 12: Nc6 Qxd3+ [ 8-]
refu 13: Bc7 Qb8+ [ 8-]
refu 14: Bb6 Qxd3+ [ 8-]
refu 15: Bxe7 Qb8+ [ 8-]
refu 16: Bb7 Qxd3+ [ 8-]
refu 17: Bc6 Qxd3+ [ 8-]
refu 18: Rxe6 Qxd3+ [ 7-]
solu 5: Qxd3+ [ 8+]
solu 6: Rxc1+ [ 8+]
solu 7: Q4xd4 [ 7+]
solu 8: Qxd2 [ 8+]
solu 9: Q1xd4 [ 8+]
solu 10: Qge4 [ 8+]
solu 11: Qe3 [ 8+]
solu 12: Qxd8 [ 6+]
solu 13: Qgf3 [ 7+]
solu 14: Q1g3 [ 8+]
solu 15: Q4g3 [ 7+]
solu 16: Qgf4 [ 5+]
solu 17: Qxd5 [ 8+]
solu 18: Qxe6 [ 7+]
solu 19: Bxa3 [ 7+]
solu 20: Bxd8 [ 7+]
solu 21: Bf6 [ 6+]
solu 22: Re1 [ 7+]
solu 23: Rd1 [ 7+]
solu 24: Rf2 [ 8+]
solu 25: Bh4 [ 8+]
solu 26: Qhf7 [ 6+]
solu 27: Qhh5 [ 7+]
solu 28: Qhh4 [ 8+]
solu 29: Qgh4 [ 7+]
solu 30: Qgh5 [ 7+]
solu 31: Qfh5 [ 7+]
solu 32: Q5f7 [ 7+]
solu 33: Q8f7 [ 7+]
solu 34: Nf4 [ 7+]
solu 35: Nf2 [ 7+]
solu 36: Ng5 [ 6+]
solu 37: Bd6 [ 7+]
solu 38: Bc5 [ 7+]
solu 39: Bb4 [ 7+]
solu 40: Bg5 [ 7+]
solu 41: Qhg7 [ 7+]
solu 42: Qh8 [ 7+]
solu 43: Qhg6 [ 7+]
solu 44: Qh1 [ 7+]
solu 45: Q1g2 [ 8+]
solu 46: Qgf2 [ 8+]
solu 47: Qgg6 [ 6+]
solu 48: Qgg7 [ 6+]
solu 49: Q4g2 [ 8+]
solu 50: Qe5 [ 7+]
solu 51: Q5f6 [ 7+]
solu 52: Qff2 [ 8+]
solu 53: Qfg6 [ 7+]
solu 54: Qfe4 [ 8+]
solu 55: Q8f6 [ 7+]
solu 56: Qfg7 [ 7+]
solu 57: Qef2 [ 8+]
solu 58: Qeg2 [ 8+]
solu 59: Rh8 [ 7+]
solu 60: Rg7 [ 7+]
solu 61: Rg6 [ 7+]
solu 62: Rg5 [ 7+]
solu 63: Rf3 [ 7+]
solu 64: Rf4 [ 6+]
solu 65: Qhh6 [ 7+]
solu 66: Qgg5 [ 6+]
solu 67: Qfg5 [ 7+]
solu 68: Qff4 [ 7+]
solu 69: Qff3 [ 8+]
solu 70: Qe8 [ 6+]
solu 71: Qfh6 [ 7+]
refu 19: Rc6 Qxd3+ [ 8-]
refu 20: Rb6 Qxb6+ [ 8-]
refu 21: Rb2 Q4xd4 [ 8-]
refu 22: Rc2 Q4xd4 [ 8-]
refu 23: Ra1 Qxd3+ [ 8-]
refu 24: Qxe7 Qxd3+ [ 8-]
refu 25: Qdc7 Qxd3+ [ 8-]
refu 26: Q7b7 Qxd3+ [ 8-]
refu 27: Qe8 Qxd3+ [ 8-]
refu 28: Q7c6 Qxd3+ [ 8-]
refu 29: Q7b5 Qxd3+ [ 8-]
refu 30: Q7xe6 Rf2 [ 8-]
solu 72: Qxd3+ [ 8+]
solu 73: Rxc1+ [ 8+]
solu 74: Q4xd4 [ 8+]
solu 75: Qxd2 [ 8+]
solu 76: Q1xd4 [ 8+]
solu 77: Qge4 [ 8+]
solu 78: Qe3 [ 8+]
solu 79: Qxd8 [ 7+]
solu 80: Qgf3 [ 8+]
solu 81: Qe8 [ 7+]
solu 82: Q1g3 [ 8+]
solu 83: Q4g3 [ 8+]
solu 84: Qxe6 [ 8+]
solu 85: Qxd5 [ 8+]
solu 86: Bxd6 [ 7+]
solu 87: Bxd8 [ 7+]
solu 88: Bf6 [ 7+]
solu 89: Re1 [ 7+]
solu 90: Rd1 [ 7+]
refu 31: Qdc8 Qxd3+ [ 8-]
refu 32: Q5xf5 Qxa2+ [ 8-]
refu 33: Qd5c5 Qxa2+ [ 8-]
refu 34: Qd5b5 Qxb5+ [ 8-]
refu 35: Q5xe6 Q4xd4 [ 8-]
solu 91: Qb5+ [ 8+]
solu 92: Qxd3+ [ 8+]
solu 93: Rxc1+ [ 8+]
refu 36: Q5c4 Qb5+ [ 8-]
refu 37: Qdb3 Qxb3+ [ 8-]
refu 38: Q5e4 Qb5+ [ 8-]
refu 39: Qf3 Qb5+ [ 8-]
refu 40: Q5c6 Qb5+ [ 8-]
refu 41: Q5b7 Qb5+ [ 8-]
refu 42: Q4e4 Qb6+ [ 8-]
refu 43: Qxg4 Qxd3+ [ 8-]
refu 44: Qd4c4 Qxd3+ [ 8-]
refu 45: Qd4b4 Qxb4+ [ 8-]
refu 46: Qf6 Qb4+ [ 8-]
refu 47: Qg7 Qb4+ [ 8-]
refu 48: Qh8 Qb4+ [ 8-]
refu 49: Q4c3 Qb4+ [ 8-]
refu 50: Q4b2 Qxd3+ [ 8-]
refu 51: Qa1 Qb4+ [ 8-]
refu 52: Q4e3 Qb4+ [ 8-]
refu 53: Q4c5 Qb4+ [ 8-]
refu 54: Qdb6 Qb4+ [ 8-]
refu 55: Q3e4 Qgxe4+ [ 8-]
refu 56: Q3xf5 Qhxf5+ [ 8-]
refu 57: Q3c2 Rxc1+ [ 8-]
refu 58: Q2c2 Rxc1+ [ 8-]
refu 59: Q2b2 Qxd3+ [ 8-]
refu 60: Qdd1 Qxd3+ [ 8-]
refu 61: Q2e3 Qxd3+ [ 8-]
refu 62: Qg5 Qxd3+ [ 8-]
refu 63: Qh6 Qxd3+ [ 8-]
refu 64: Qe1 Qxd3+ [ 8-]
refu 65: Q2c3 Qxd3+ [ 8-]
refu 66: Q2b4 Qxd3+ [ 8-]
refu 67: Q6b6 Qxd3+ [ 8-]
refu 68: Q6c6 Qxd3+ [ 8-]
refu 69: Qab7 Qxd3+ [ 8-]
refu 70: Qac8 Qxd3+ [ 8-]
refu 71: Q6b5 Qxd3+ [ 8-]
refu 72: Q6c4 Qxd3+ [ 8-]
refu 73: Qa5b5 Qxd3+ [ 8-]
refu 74: Qa5c5 Qxd3+ [ 8-]
refu 75: Q5b6 Qxd3+ [ 8-]
refu 76: Qac7 Qxd3+ [ 8-]
refu 77: Q5b4 Qxd3+ [ 8-]
refu 78: Q5c3 Qxd3+ [ 8-]
refu 79: Qa4b4 Qxd3+ [ 8-]
refu 80: Qa4c4 Qxd3+ [ 8-]
refu 81: Q4b5 Qxd3+ [ 8-]
refu 82: Q4c6 Qxd3+ [ 8-]
refu 83: Q4b3 Qxd3+ [ 8-]
refu 84: Qac2 Q4xd4 [ 8-]
refu 85: Qad1 Q4xd4 [ 8-]
refu 86: Q3b3 Qxd3+ [ 8-]
refu 87: Q3c3 Qxd3+ [ 8-]
refu 88: Q3b4 Qxd3+ [ 8-]
refu 89: Q3c5 Qxd3+ [ 8-]
refu 90: Qab2 Qxd3+ [ 8-]
refu 91: Ka1 Q1xd4+ [ 8-]
refu 92: Kb2 Q1xd4+ [ 8-]
refu 93: Kc2 Qxd3+ [ 8-]
Unfortunately, the above entries do not show the individual amount of
computation done for them, so we do not see much, what would explain
the EBF=119. (yet another detail to add to Chest)
Otherwise, the statistics from Chest do not provide surprises or clues.
I will have to study some parts of its search tree to understand what went
wrong in depth=9. This "US" has become an important test-case for Chest.
Playing the PV for the quiet solution move until after white's 6th move:
1.Rxe6 Rxc1+ 2.Qaxc1 Qxc1+ 3.Qxc1 Qxd3+ 4.Qaxd3 Qxd3+ 5.Qxd3 Qb4+ 6.Q4xb4
[D]B2B1qr1/N2Qb3/4R3/Q2Q4/1Q6/3Q3n/R3q2k/1KQ5 b - -
shows the hopeless situation for black quite drastically.
Black has sacrificed one of its pieces in every move, in order to say check,
but now checks by black do not prolong the mate any more.
I also instructed Chest to print the top 7 plies of the solution tree...
that part is not yet ready (after 98 hours total time) :-(
Have a nice day!
Heiner
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.