Author: Joachim Rang
Date: 04:26:43 03/04/04
Go up one level in this thread
On March 03, 2004 at 15:23:14, John Merlino wrote: >On March 03, 2004 at 13:48:07, Joachim Rang wrote: > >>On March 03, 2004 at 13:39:40, John Merlino wrote: >> >>>On March 03, 2004 at 10:38:23, Joachim Rang wrote: >>> >>>>On March 03, 2004 at 10:14:50, Uri Blass wrote: >>>> >>>>>I wonder if the solution f6 is correct solution >>>>> >>>>>[D]r3r1k1/6p1/1pqp2Np/2p1nP1P/N1PbPQ2/1P3PR1/P2K4/7R w - - 0 1 >>>>> >>>>>I could not find a convincing line after f6 Nxc4+ but maybe I miss something. >>>>> >>>>>Movei could not solve it at 5 minutes per position on A1000 becasue after few >>>>>minutes it changed its mind from f6 to Nc3 after failing low at depth 10. >>>>> >>>>>Uri >>>> >>>>I think you just discovered an error in this testsuite. Most engines fail low >>>>over and over after: >>>> >>>>1.f6+ Nxc4+ 2. bxc4 Qxa4 3. f7+ Kh7 >>>> >>>>more (engine)-analysis will follow. >>>> >>>>regards Joachim >>> >>>I don't think there is a problem. After forcing those moves, Chessmaster 9000 >>>(SKR personality on a P4-2.4) still shows a good score for White. I think >>>(perhaps inadvertantly) the point is that f6 is the BEST move and that the >>>engine shouldn't fail TOO low on f6 Nxc4+...? Not sure.... >>> >>>Perhaps the point is to see that, after the above moves, White must play 4.Rc1 >>>rather than 4.fxe8=Q which leads to a draw. >>> >>>I also believe, after a lot of analysis and playing around with lines, that f6 >>>does not necessarily lead to a forced win. I DO believe that it is the best >>>move, however. >>> >>>Here is what The King says: >>> >>>Time Depth Score Positions Moves >>>27:10 2/14 1.42 387147633 4.Rc1 Qxa2+ 5.Rc2 Qa5+ 6.Ke2 Qa4 >>> 7.Rc1 Rf8 8.Nxf8+ Rxf8 9.Rg6 Qa2+ >>> 10.Qd2 Qa7 11.Qd3 Qxf7 12.Rb1 Qf4 >>> 13.Rxb6 >>> >>>jm >> >> >>go deeper in the variations. For example after 4.Rc1 Qxa2+ 5.Rc2 Qa5+ 6.Ke2 Qa4 >>7.Rc1 Rf8 8.Nxf8+ Rxf8 9.Rg6 >>9... Qa7! draws. >> >>What score does CM9SKR shows after Nc3? > >My analysis stopped at 9.Rg6, because CM9 has a difficult time determining the >best move here. Depending on the time control, it could play Rg6, Qd2 or Kf1. >All have evals in the neighborhood of +0.5 for White, but the evals were slowly >dropping. Which, as I said, leads me to believe that f6 does NOT necessarily >lead to a forced win, but that it is the best move. > >After 1.Nc3, CM9SKR says Black wins after 1...Nxc4+ 2.bxc4 Bxc3+ 3.Kxc3: > >Time Depth Score Positions Moves >0:00 1/4 -0.04 9967 1...b5 2.cxb5 Bxc3+ 3.Kxc3 Qxb5 > 4.Kb2 >0:00 1/5 0.42 45754 1...b5 2.cxb5 Bxc3+ 3.Kxc3 Qxb5 > 4.Nxe5 Qb4+ 5.Kc2 dxe5 >0:00 1/6 0.42 148074 1...b5 2.f6 Bxc3+ 3.Kxc3 b4+ 4.Kd2 > Rxa2+ 5.Ke3 Qb7 6.fxg7 >0:02 1/7 0.78 434070 1...b5 2.f6 Qc7 3.Nxe5 Bxe5 4.Qxe5 > dxe5 5.Rxg7+ Qxg7 6.fxg7 bxc4 >0:04 1/7 0.69 743221 1...d5 2.cxd5 Qf6 3.Na4 Kh7 4.Nxe5 > Bxe5 >0:10 1/8 0.00 2007590 1...d5 2.Nxd5 Rxa2+ 3.Kd1 Ra1+ > 4.Kc2 Ra2+ 5.Kd1 >0:24 1/9 0.07 5005359 1...d5 2.exd5 Qf6 3.Rc1 Bxc3+ 4.Rxc3 > Nf7 5.a4 Nd6 6.Re3 Qb2+ 7.Kd1 >1:05 1/10 0.00 13307035 1...d5 2.exd5 Qf6 3.Rc1 Nxc4+ 4.bxc4 > Bxc3+ 5.Rxc3 Rxa2+ 6.Rc2 Qa1 7.Rxa2 > Qxa2+ 8.Kc3 Qa1+ 9.Kd3 Qb1+ 10.Kd2 > Qa2+ >3:42 1/11 0.00 46789493 1...d5 2.f6 Bxc3+ 3.Kxc3 d4+ 4.Kc2 > Rxa2+ 5.Kd1 Ra1+ 6.Kc2 Ra2+ >4:24 1/11 -0.05 55885892 1...Nxc4+ 2.bxc4 Bxc3+ 3.Kxc3 Qa4 > 4.Qc1 Qb4+ 5.Kd3 Ra3+ 6.Ke2 Rc3 > 7.Qd1 Qb2+ 8.Qd2 Rc2 9.Rd1 Rxd2+ > 10.Rxd2 >10:45 1/12 -1.13 138913431 1...Nxc4+ 2.bxc4 Bxc3+ 3.Kxc3 Qa4 > 4.Qc1 Qb4+ 5.Kd3 Ra3+ 6.Ke2 Rc3 > 7.Qe1 Qxc4+ 8.Kf2 Rc2+ 9.Ke3 d5 > 10.Rg4 dxe4 11.fxe4 Rxa2 >18:15 1/13 -1.13 238745890 1...Nxc4+ 2.bxc4 Bxc3+ 3.Kxc3 Qa4 > 4.Qc1 Qb4+ 5.Kd3 Ra3+ 6.Ke2 Rc3 > 7.Qe1 Qxc4+ 8.Kf2 Rc2+ 9.Ke3 d5 > 10.Rg4 dxe4 11.fxe4 Rxa2 > >So it looks like 1.f6 is definitely best.... > >jm yes I agree. SEcond best move is probably Rh1-c1 but Shredder gives that move only -0.12 after 12 hours. So f6 is a correct solution although black has still drawing chances after that move. regards Joachim
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.