Author: Uri Blass
Date: 22:23:10 07/31/03
Go up one level in this thread
On July 31, 2003 at 21:23:27, Brian Kostick wrote: >On July 31, 2003 at 20:43:43, Eduard Nemeth wrote: > >> >>Hello! >> >>First my original german text: >> >>Junior 8 spielt anders als er analysiert!! >> >>Ich habe jetzt, glaube ich, 5 games gegen Junior 8 gepostet. Die letzten 2 >>Partien habe ich nun analysiert und mir ist es nicht gelungen, sie im analyse >>Modus zu reproduzieren. Mir persönlich erschien es so, dass Junior 8 oft manche >>Moves wie per Zufalls-generator oder als 2. best-move gespielt hat. Auch die >>Suchtiefe ist eine Andere im Analysemodus als beim Spiel. Hat sich Amir Ban da >>etwas raffiniertes dabei gedacht? >> >>(Diese Feststellungen bezeugen auch die Echtheit meiner games) >> >>English (Translator) : >> >>Hello! >> >>Junior 8 plays differently than it analyzed!! >> >>I have now, I believe, 5 games against junior 8 post. The last 2 games I >>analyzed now and me it did not succeed to reproduce it in analysis the mode. Me >>it appeared personal in such a way that junior often played 8 some moves as by >>random number generator or as the 2nd best move. Also the search depth is >>another in the analysis mode than with the play. >> >>Did Amir Ban imagine somewhat refined there thereby? >> >>(these statements testify also the authenticity my games) >> >>Eduard > > >Hello Eduard, > > This has been seen with other ChessBase engines as well. It was a topic of >discussion here sometime ago when several of us were a running test suite and we >noted conflicting results. It may be related to hash tables somehow (I am >guessing). In the case of Junior there are different reasons and it is because of different depthes. If the program jumps from depth 12 to 15 to 17 in analysis mode it may suggest different move relative to the case that it is playing and jumps from depth 14 to 16. 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.