Author: Uri Blass
Date: 08:56:57 10/02/03
Go up one level in this thread
On October 02, 2003 at 11:10:12, Grzegorz Sidorowicz wrote: >On October 02, 2003 at 10:44:06, Uri Blass wrote: > >>On October 02, 2003 at 10:13:13, Grzegorz Sidorowicz wrote: >> >>>On October 02, 2003 at 10:02:58, Grzegorz Sidorowicz wrote: >>> >>>> >>>>>>My Armageddon 2.250 on PIV 2Ghz without any magic extension >>>>>>can solve it fast and I think it is not bug. >>>>> >>>>>This is the line that it needs to see in order to solve it. >>>>>If 10 plies are enough then you do a lot of extensions. >>>> >>>> >>>>Armageddon extension: >>>>1. check (always), >>>>2. capture on the same square if SEE>=0 and >>>> we are not in check and opponent move was >>>> capture >>>>3. if we have got only one move >>>> >>>>...but if I detect check in QSearch I generate all moves. >>>>More extension I can't remember :-) >>>> >>>>Grzegorz >>> >>> >>>Second point again :-) >>>2. capture on the same square if SEE>=0 and >>> we are not in check and opponent move was >>> capture with SEE>=0 >> >>It is clear that you have a bug >> >>1.Nc6+ Kb7 2.Na5+ Kb6 3.Bf2+ Kxa5 >>4.e8Q Nc7+ 5.Kd6 Nxe8 >>6.Kc5 d2 7.Bd4 d1N >>8.Be5 f2 >>9.Bf4 f1N >>10.Bg5 Nf6 >>11.Bxf6 Nf7 >>12.Bh4 Nf2 >>13.Bxf2 Nd2 >>14.Be1 Be2 >>15.Bxd2# >> >>the line is 29 plies. >>there are only 5 checks and there is no case of recapture or single reply. >> >>you need more than 20 plies to see 1.Nc6+ without bugs. >> >>Maybe the bug is not considering underpromotions(otherwise I cannot exaplain >>your result). >> >>Uri > >I don't know why my program see it. >In my opinion one path can't explain anything. >Sometimes we think: this line is good, but line can be good because >other lines are worse. In this subject I see only one >fact: my program can't solve it without checks in qsearch, >but maybe it is not only one important thing. >In other side programs who can't solve it are better than Armageddon, well >maybe I have got bugs. >We can easy find positions where solution programs can find on >lower depth without extensions. >I think good example is position: >[D]8/k7/3p4/p2P1p2/P2P1P2/8/8/K7 w - - 0 1 >...where helpful is transposition table. Yes but I do not believe that tranposition tables can help you there because the relevant position is not reached in different order of moves. The fact that your program solve what other programs fail to solve suggest that it has a bug. You should check if your program can see all the moves of the line that I posted. Uri
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.