Author: Tim Foden
Date: 13:13:43 03/02/02
Go up one level in this thread
On March 02, 2002 at 12:02:09, Heiner Marxen wrote:
Thanks Heiner and Leonid for your comments. It does appear that there is a bug
in v2.16, and a bug has been found in play in the WinBoard forum where GLC
failed to mate with KQ vs. K!
I think I may have found the culprit (just a one line change), but I need to do
some more testing before I can be sure.
It now solves this mate in about 16mins, and finds the correct mate in 10 line.
Cheers, Tim.
>On March 02, 2002 at 10:54:41, leonid wrote:
>
>>On March 02, 2002 at 10:07:36, Tim Foden wrote:
>>
>>>On March 02, 2002 at 08:21:10, leonid wrote:
>>>
>>>>[D]3Q3R/Q2b2Q1/1Q1q1Q2/1qQqQq1B/Nr1q1rNB/2qkq3/1qnbnq2/Q2K2QR w - -
>>>>
>>>>Please indicate your result.
>>>>
>>>Green Light Chess v2.16, 48MB Hash, Athlon XP 1700+ says Mate in 9 after just
>>>over 13.5 minutes.
>>>
>>>Cheers, Tim.
>>
>>Hi, Tim!
>>
>>Are you sure about 9 moves solution? Or I tried some different position from
>>what is here, or you put wrong on your side. This position mine looked 9 moves
>>deep by brute force and found no mate. I don't know exactly shortest mate for
>>it. It was solved by 11 moves by selective and searched only 9 moves deep by
>>brute force. I can see 9 moves only during this night, since it took already 4.5
>>hours. It could be that Heiner, with speedier computer and hash, will say
>>shortest mate before this.
>>
>>Cheers,
>>Leonid.
>>
>
>Hi Tim, Hi Leonid,
>
>Chest confirms that there is not mate-in-9.
>Since that was my initial guess, I just restarted with depth=11.
>For depth=9 I needed 19.2 minutes (K7/600 350MB hash).
>The faster box is at work, not here at home, sorry.
>
>
>>>>anal
>>> Game stage: Opening
>>> Current eval: -0.685
>>> Ply Time Score Nodes Principal Variation
>>> 1 0.06 +15.973 13553 Nxb2+ Rxb2 2. Nxf2+
>>> 2 0.10 +16.152 38665 Nxb2+ Rxb2 2. Qbxd6
>>> 2 0.38 +16.152 218218 Nxb2+ Rxb2 2. Qbxd6
>>> 3 0.38 +16.552 220805 Nxb2+ Rxb2 2. Nxf2+ Qxf2 3. Qcxd6 {++}
>>> 3 0.64 +22.415 430008 Nxb2+ Rxb2 2. Nxf2+ Rxf2 3. Qbxd6 Qbxc5 4. Qfxf5+
>>> 3 2.26 +22.415 1781082 Nxb2+ Rxb2 2. Nxf2+ Rxf2 3. Qbxd6 Qbxc5 4. Qfxf5+
>>> 4 8.67 +22.126 7320735 Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Qbxb5+ Rxb5 4. Bxe2+
>>> Qxe2+ 5. Qxe2+ Rxe2 6. Qxf5+ Bxf5 7. Qg1xd4+ Nxd4
>>> 8. Qxd5 Rxd5 9. Qxb2
>>> 4 29.51 +22.126 24640k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Qbxb5+ Rxb5 4. Bxe2+
>>> Qxe2+ 5. Qxe2+ Rxe2 6. Qxf5+ Bxf5 7. Qg1xd4+ Nxd4
>>> 8. Qxd5 Rxd5 9. Qxb2
>>> 5 30.89 +22.526 25662k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Qbxb5+ Rxb5 4. Bxe2+
>>> Qxe2+ 5. Qxe2+ Rxe2 6. Qxf5+ Bxf5 7. Qg1xd4+ Nxd4
>>> 8. Qcxd4+ Qbxd4 9. Qa1xd4+ Qxd4 10. Qaxd4+ {ht}
>>> {++}
>>> 5 1:25 +32.728 69902k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Bxe2+ Rxe2 4. Qfxf5+
>>> Qd5e4 5. Qexe4+ Qexe4 6. Qxe4+ Kxe4 7. Qde7+ Qxe7
>>> 8. Qgxe7+ Qe5 9. Qxb2
>>> 5 8:06 +32.728 384947k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Bxe2+ Rxe2 4. Qfxf5+
>>> Qd5e4 5. Qexe4+ Qexe4 6. Qxe4+ Kxe4 7. Qde7+ Qxe7
>>> 8. Qgxe7+ Qe5 9. Qxb2
>>> 6 8:30 +33.128 403239k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Bxe2+ Rxe2 4. Qfxf5+
>>> Qd5e4 5. Qexe4+ Qexe4 6. Qxe4+ Kxe4 7. Q7g6+ Kf4
>>> 8. Qcxd6+ Kf3 9. Qxb2 {++}
>>> 6 13:24 +Mate09 782417k Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Bxe2+ Rxe2 4. Qcxd4+
>>> Qexd4 5. Qgxd4+ Nxd4 6. Qxb5+ Kc3 7. Qexd4+ Qxd4
>>> 8. Qfxd4+ {ht}
>>> 6 35:16 +Mate09 2669m Nxb2+ Qxb2 2. Nxf2+ Rxf2 3. Bxe2+ Rxe2 4. Qcxd4+
>>> Qexd4 5. Qgxd4+ Nxd4 6. Qxb5+ Kc3 7. Qexd4+ Qxd4
>>> 8. Qfxd4+ {ht}
>
>According to Chest, after 1.Nxb2+ Rxb2 there is a mate-in-8 to follow up,
>but not after 1.Nxb2+ Qxb2.
>
>Ok, lets follow that line, as far as you give it. After 7 moves, where your
>programs gives 8. Qfxd4+, what should be a mate-in-2, we have
>
>[D]3Q3R/Q2b2Q1/3q1Q2/1Q3q2/1r1q3B/2k5/1q1br3/Q2K3R w - -
>This is a mate-in-2 with 8.Q7a3+ Rb3 9.Qbxb3#, but after 8. Qfxd4+
>black may do 8...Qxd4 after which there is no mate move. After the other
>two legal black moves, 8...Rxd4 and 8...Kb3 there are mate moves available.
>
>So, something is wrong at your side, I guess. But since we have a mate,
>something else must be wrong, too. Lets go on:
>
>After 6 of your moves we have
>[D]3Q3R/Q2b2Q1/3q1Q2/1Q1qQq2/1r1n3B/2k5/1q1br3/Q2K3R w - -
>which is a mate in 3 with solution
> Qexd4+ Qxd4 Q7a3+ Rb3 Qbxb3#
> Rxd4 Qaxb2#
> Kb3 Qdxb2#
>Ok.
>
>After 5 of your moves we have
>[D]3Q3R/Q2b2Q1/1Q1q1Q2/1q1qQq2/1r1n3B/3k4/1q1br3/Q2K3R w - -
>and, according to Chest, no mate-in-4, as your line would need.
>Your program gave 6.Qxb5+ Kc3, which we see above, but what after
>6.Qxb5+ Qxb5 ? No mate in 3 is available, here.
>
>That should be enough stuff for you to search for a bug. Or you had a
>slightly different position. Check against the FENs above.
>
>Cheers anyhow,
>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.