Author: Paul
Date: 14:28:48 06/30/00
Go up one level in this thread
This is what Marty Hirsch had to say about the subject on 12-31-1998: ***************************************************************************** On December 29, 1998 at 16:20:06, Dan Kiski wrote: >The following game is from Fritz 5.16 v MCP8 > >Fritz 5.16 white >MCP8 black > >1. NF3......NF6 >2. D4.......E6 >3. C4.......B6 >4. NC3......BB7 >5. A3.......D5 >6. CD.......ED >7. QA4 +....C6 >8. G3.......NBD7 >9. BG2......BD6 >10. 0-0......0-0 >11. BF4......QE7 >12. BXD6.....QXD6 >13. RFD1.....RFD8 >14. E3.......A5 >15. RAE1.....BA6 > >Fritz is now out of book but chooses NH4, MCP8 is still in book expecting QB3. > >Now the postion after 16. NH4 i have played through countless times at varying >time controls yet MCP8 seems to always choose 16.......NG4. > >It has lost from this position over and over but always chooses 16.....NG4 and >after Fritz's 17. E4 replys with 17......QF6 it might from this point try to >change the odd move yet according to Fritz it was already way in front, can >someone explain how this book learning works and how many losses will it take >before the computer in this case MCP8 varies it's responce, in this case most >likely remove 16........NG4 since this seems to be the start of its problems. >At one point it added 16......NG4 to book after 6 straight losses using it. > >I wait until it is 7 points behind claim the win and then scroll back. Play 16. >NH4 then carry on again. Am I doing something wrong with this method. Dan, Thank you for this information. Machine Learning is difficult to do well and there is clearly room for improvement in this important uncharted area. You will have observed that the Book Learning in M-Chess does NOT work the same as it does in other programs--M Chess rarely kills a book move, and does not alter the priorities of the opening choices based on wins and losses. Instead, it tries to find good moves and add them to the Book. This avoids biasing the Learning/Book Learning to favor only a particular opponent. In the case of the moves shown above something appears to have gone wrong. I would not expect to see the move Ng4 stored. It is possible that this occurred as a result of using the Book Learning when the Score Learning was not enabled. You might want to clear the Learning files and try the experiment again. I ought to provide a dialog to clear either the Learning, Book Learning or both in the next version of M-Chess Professional. For now, you can clear the learning files by deleting, moving or renaming the files learnw.m80 and learnb.m80 in the \MCP8 subdirectory. You can reset the Book Learning by deleting, moving, or renaming the .LBK files in the \MCP8\UBOOK subdirectory. By the way, if you rename these to .UBK files then you can edit them and use them as User Books. In the line quoted above, 16... Ng4 is an ill-fated move and MCP8 SHOULD remove it from the Book after a few games--however, you must play at least eleven more moves for it to do so. That may be the reason this move remains even after your winning games in which you are claiming Victory. This is something I expect to improve in the next version. One more point is that the learning functions do not work in "Analysis" Mode. Learning and Book Learning are operational only when the program is used in normal play mode. Learning works at Infinite Level, but Book Learning does not. I hope you find this information helpful for understanding the MCP8 Learning features. ****************************************************************************** Hope this helps, Paul
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.