Difficult tablebase/bitbase position

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Jouni
Posts: 3652
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Difficult tablebase/bitbase position

Post by Jouni »

[d]2k5/3P4/8/8/8/8/1r5p/R3K3 b Q -

Yes this is old position. Many current top engines are still helpless e.g. Rybka 3 and Naum 4.I only got complete solution from Chest(!):

1. ...+ Kd8 {+M18/18 0} 2. Ra8+ {-M17/17} Kxd7 {+M17/17} 3. Ra7+ {-M16/16}
Kd6 {+M16/16} 4. Ra6+ {-M15/15} Kd5 {+M15/15} 5. Ra5+ {-M14/14} Ke4
{+M14/14} 6. Rh5 {-M13/13} h1=Q+ {+M13/13} 7. Rxh1 {-M12/12} Rb1+ {+M12/12}
8. Kd2 {-M11/11} Rxh1 {+M11/11} 9. Kc3 {-M10/10} Rc1+ {+M10/10} 10. Kb4
{-M9/9} Kd5 {+M9/9} 11. Kb3 {-M8/8} Rc4 {+M8/8} 12. Ka3 {-M7/7} Kc5 {+M7/7}
13. Kb3 {-M6/6 0} Kb5 {+M6/6} 14. Ka3 {-M5/5} Rb4 {+M5/5} 15. Ka2 {-M4/4}
Ka4 {+M4/4} 16. Ka1 {-M3/3} Kb3 {+M3/3} 17. Kb1 {-M2/2 0} Rc4 {+M2/2} 18.
Ka1 {-M1/1} Rc1# {+M1/1} 0-1

Jouni
Top
User avatar
michiguel
Posts: 6401
Joined: Thu Mar 09, 2006 8:30 pm
Location: Chicago, Illinois, USA

Re: Difficult tablebase/bitbase position

Post by michiguel »

I assume that we are looking for a correct mate announcement with engines with 5 pc installed? Obviously, finding the correct move is easy even with no TBs.

Rybka and Naum had 5 pc TBs installed?

Miguel
Jouni wrote:[d]2k5/3P4/8/8/8/8/1r5p/R3K3 b Q -

Yes this is old position. Many current top engines are still helpless e.g. Rybka 3 and Naum 4.I only got complete solution from Chest(!):

1. ...+ Kd8 {+M18/18 0} 2. Ra8+ {-M17/17} Kxd7 {+M17/17} 3. Ra7+ {-M16/16}
Kd6 {+M16/16} 4. Ra6+ {-M15/15} Kd5 {+M15/15} 5. Ra5+ {-M14/14} Ke4
{+M14/14} 6. Rh5 {-M13/13} h1=Q+ {+M13/13} 7. Rxh1 {-M12/12} Rb1+ {+M12/12}
8. Kd2 {-M11/11} Rxh1 {+M11/11} 9. Kc3 {-M10/10} Rc1+ {+M10/10} 10. Kb4
{-M9/9} Kd5 {+M9/9} 11. Kb3 {-M8/8} Rc4 {+M8/8} 12. Ka3 {-M7/7} Kc5 {+M7/7}
13. Kb3 {-M6/6 0} Kb5 {+M6/6} 14. Ka3 {-M5/5} Rb4 {+M5/5} 15. Ka2 {-M4/4}
Ka4 {+M4/4} 16. Ka1 {-M3/3} Kb3 {+M3/3} 17. Kb1 {-M2/2 0} Rc4 {+M2/2} 18.
Ka1 {-M1/1} Rc1# {+M1/1} 0-1

Jouni
Top
Jouni
Posts: 3652
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: Difficult tablebase/bitbase position

Post by Jouni »

Yes with 5 piece tbs, but sorry, Naum has no problem at all! Of course
castling is problem here. Shredder 6 piece online says: -Kb7 = mate in 4??

Jouni
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Difficult tablebase/bitbase position

Post by bob »

michiguel wrote:I assume that we are looking for a correct mate announcement with engines with 5 pc installed? Obviously, finding the correct move is easy even with no TBs.

Rybka and Naum had 5 pc TBs installed?

Miguel
Jouni wrote:[d]2k5/3P4/8/8/8/8/1r5p/R3K3 b Q -

Yes this is old position. Many current top engines are still helpless e.g. Rybka 3 and Naum 4.I only got complete solution from Chest(!):

1. ...+ Kd8 {+M18/18 0} 2. Ra8+ {-M17/17} Kxd7 {+M17/17} 3. Ra7+ {-M16/16}
Kd6 {+M16/16} 4. Ra6+ {-M15/15} Kd5 {+M15/15} 5. Ra5+ {-M14/14} Ke4
{+M14/14} 6. Rh5 {-M13/13} h1=Q+ {+M13/13} 7. Rxh1 {-M12/12} Rb1+ {+M12/12}
8. Kd2 {-M11/11} Rxh1 {+M11/11} 9. Kc3 {-M10/10} Rc1+ {+M10/10} 10. Kb4
{-M9/9} Kd5 {+M9/9} 11. Kb3 {-M8/8} Rc4 {+M8/8} 12. Ka3 {-M7/7} Kc5 {+M7/7}
13. Kb3 {-M6/6 0} Kb5 {+M6/6} 14. Ka3 {-M5/5} Rb4 {+M5/5} 15. Ka2 {-M4/4}
Ka4 {+M4/4} 16. Ka1 {-M3/3} Kb3 {+M3/3} 17. Kb1 {-M2/2 0} Rc4 {+M2/2} 18.
Ka1 {-M1/1} Rc1# {+M1/1} 0-1

Jouni
Here is one:

this is on my laptop where the egtb's are horribly slow:

Code: Select all

               23->   1:03 -12.94   1. ... Kd8 2. O-O-O Rg2 3. Kb1 Rg1
                                    4. Kc2 Rxd1 5. Kxd1 h1=Q+ 6. Ke2 Qe4+
                                    7. Kd2 Kxd7 8. Kc3 Qe1+ 9. Kd3 Kc6
                                    10. Kd4 Qe2 11. Kc3 Kc5 12. Kb3 Qd1+
                                    13. Kc3 Qe2 <HT> (s=3)
               24     1:04     -1   1. ... Kd8!
               24     1:07     -3   1. ... Kd8!
               24     1:07     -M   1. ... Kd8!
               24     5:58 -Mat18   1. ... Kd8 2. Ra8+ Kxd7 <EGTB> (s=2)
Black(0): ?
              time=6:09  mat=0  n=2136671805  fh=97%  nps=5.8M
My office machine will be about 2x faster, although the NPS will go down since it is a slower CPU. The 8-core box I use will probably get this in under a minute as these nodes have 16gb of ram which is enough to cache all the egtbs and make them fly. Can't get to these at the moment as our cluster switch is in the process of being replaced... I'll try later today if I can.
Top
Jouni
Posts: 3652
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: Difficult tablebase/bitbase position

Post by Jouni »

No need for cluster, Naum on single:

1...Kd8 2.0-0-0 Ra2 3.Kb1 Rg2 4.Kc1 Rg1 5.Kb2 h1Q 6.Rd2
-+ (-8.87) Depth: 10/17 00:00:00 28kN, tb=171
1...Kd8 2.0-0-0 Ra2 3.Kb1 Rg2 4.Kc1 Rg1 5.Kb2 Rxd1 6.Kc2 h1Q
-+ (-12.08) Depth: 11/17 00:00:00 31kN, tb=171
1...Kd8 2.0-0-0 Rg2 3.Rf1 Kxd7 4.Kb1 Kd6
-+ (-#19) Depth: 12/18 00:00:00 37kN, tb=249
1...Kd8 2.0-0-0 Rg2 3.Rf1 Kxd7 4.Kb1 Kd6
-+ (-#19) Depth: 13/13 00:00:00 38kN, tb=284
1...Kd8 2.Ra8+ Kxd7 3.Ra7+ Kd6
-+ (-#18) Depth: 20/27 00:00:01 269kN, tb=4443
1...Kd8 2.Ra8+ Kxd7 3.Ra7+ Kd6
-+ (-#18) Depth: 21/27 00:00:01 335kN, tb=5689

Jouni
Top
Pressie

Re: Difficult tablebase/bitbase position

Post by Pressie »

Jouni wrote:Shredder 6 piece online says: -Kb7 = mate in 4??

Jouni
Shredder is correct.
Top
Dirt
Posts: 2851
Joined: Wed Mar 08, 2006 10:01 pm
Location: Irvine, CA, USA

Re: Difficult tablebase/bitbase position

Post by Dirt »

Pressie wrote:
Jouni wrote:Shredder 6 piece online says: -Kb7 = mate in 4??

Jouni
Shredder is correct.
Shredder has the correct answer but to the wrong problem. The Nalimov tablebases assume castling is never possible, but it is here.
Top
User avatar
michiguel
Posts: 6401
Joined: Thu Mar 09, 2006 8:30 pm
Location: Chicago, Illinois, USA

Re: Difficult tablebase/bitbase position

Post by michiguel »

Jouni wrote:Yes with 5 piece tbs, but sorry, Naum has no problem at all! Of course
castling is problem here. Shredder 6 piece online says: -Kb7 = mate in 4??

Jouni
Gaviota, with 5pc Gaviota tablebases, AMD 2x 2.4 Ghz, 1 core, finds it in 0.4 seconds and resolves it in 1.2 seconds, 9 plies.

Why is this so difficult? is there anything wrong in the TB probing code or are they probing with the wrong position?

Miguel

Code: Select all

         1   1       0.0    +2.07  Kxd7
         5   1:      0.0    +2.07  Kxd7
        21   2       0.0      :-(  Kxd7
        59   2       0.0      :-(  
       111   2       0.0      :-(  Kxd7
       153   2       0.0      :-(  
       217   2       0.0      :-(  Kxd7
       315   2       0.0    +0.14  Kc7 2.O-O-O
       354   2       0.0    +0.15  Kd8 2.O-O-O
       374   2:      0.0    +0.15  Kd8 2.O-O-O
       493   3       0.0    +0.15  Kd8 2.O-O-O Re2
       669   3:      0.0    +0.15  Kd8 2.O-O-O Re2
       966   4       0.0    +0.17  Kd8 2.O-O-O Re2 3.Kb1
      1376   4:      0.0    +0.17  Kd8 2.O-O-O Re2 3.Kb1
      1780   5       0.0    +0.26  Kd8 2.O-O-O Re2 3.Kb1 Rd2
      2541   5:      0.0    +0.26  Kd8 2.O-O-O Re2 3.Kb1 Rd2
      3385   6       0.1      :-)  Kd8
      4040   6       0.1      :-)  Kd8
      5497   6       0.1    +1.61  Kd8 2.O-O-O Re2 3.Rf1 Kxd7 4.Kd1
      5952   6:      0.1    +1.61  Kd8 2.O-O-O Re2 3.Rf1 Kxd7 4.Kd1
      9068   7       0.2    +1.64  Kd8 2.O-O-O Ra2 3.Kb1 Rd2 4.Re1 Kxd7
     12201   7:      0.2    +1.64  Kd8 2.O-O-O Ra2 3.Kb1 Rd2 4.Re1 Kxd7
     14300   8       0.3    +1.63  Kd8 2.O-O-O Ra2 3.Kb1 Rd2 4.Re1 Kxd7
                                   5.Kc1
     17701   8:      0.3    +1.63  Kd8 2.O-O-O Ra2 3.Kb1 Rd2 4.Re1 Kxd7
                                   5.Kc1
     30369   9       0.4      :-)  Kd8
     33258   9       0.4      :-)  Kd8
     37026   9       0.5      :-)  Kd8
     46961   9       0.5      :-)  Kd8
    103868   9       1.2  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    139995   9:      1.5  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    163829  10       1.8  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    227638  10:      2.1  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    290543  11       2.8  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    448888  11:      3.7  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
    658111  12       4.9  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE
   1022416  12:      6.4  +Mat_18  Kd8 2.Ra8+ Kxd7 <=TABLE

GTB CACHE STATS
  probes:              5536
  efficiency:          82.7%
  average searches    205.7
  occupancy:           14.4%
Top
Dann Corbit
Posts: 12792
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Difficult tablebase/bitbase position

Post by Dann Corbit »

I like the smilies and frownies in your pv.

IIRC, Mint was the first program to do this.
Top

Return to “Computer Chess Club: General Topics”