News:

Remember if you wish to share a pic or a story, with just a selected few, use the new extended profile with custom buddy list http://www.fibsboard.com/announcements/new-ultimate-profile-a-biopicdocumentsounds-area-that-can-only-be-seen-by-b/

Main Menu

can we trust to XG engine?!

Started by LENNA, February 26, 2014, 10:08:28 AM

Previous topic - Next topic

LENNA

hello everybody.
i have heard that XG is the best in the world ..but i never like its playing style. i dont know why it plays very solid and passive . i think GNU is better because it plays much more risky . you may say xg can beat GNU at a long length competition . if we run a engine match between them we would see xg achieve a good result. ok. i dont deny this . but i want to say something different. i want to say we cant trust its solid playing style.
let me bring you to chess world ..one day fritz 5 was the best in the world ..chess players remember this. but now any chess players uses houdini4 engine ! any good move by fritz5 may has been considered to a bad move by houdini 4 ! i want to say the world of computer software is changing.. some good moves by XG may has been considered a bad move or a doubtful move by some engine at future! so ..in this changing world why should i make myself forced to use a solid playing engine called XG? i'd like to use a dynamic playing engine such as GNU ! if you have a different idea please let me know. waiting to your feedback . (sorry for poor english :-p)

KDP

they are both good programs, which one is better?  i think most people will say XG is the stronger of the two and i personally don't disagree with that.  which one is right for you depends on what you want out of the program.  if you do extensive rollouts then XG is the clear winner.  its also a little more user friendly.  i have both and actually prefer GNU.  this is mainly because i use my macbook almost exclusively now instead of my hp laptop.  i also prefer GNU because it "grades" your performance harder and in my opinion it more accurately reflects my playing strength.  as an example i ran a recent 11 pt match through both and XG graded me "expert" while GNU had me at "advanced".  the error rate numbers were almost identical and the XG rating is an ego booster but the advanced label is probably a more accurate reflection if where i'm at.  also there is the cost factor, while XG isn't going to put most folks into bankruptcy, GNU is still free.

dorbel

The mistake here may be to consider that either bot, or indeed any top class human has a recognisable "style" any more. I think this used to be the case before there were bots available, with some players noted for a high risk tactical approach while others played more positionally. Nowadays a well rounded player (or bot) will adapt his, her or its game according to the position and the score. Sometimes all out attack is appropriate, what Lenna might call "dynamic", while another position might require a more passive approach and of course often something in between.
I have hundreds of matches against GBots on file, effectively Gnu 2-ply, and I analyse all of them with XG. There is very little difference in their ability. XG has an edge, but not by much and no noticeable difference in their approach to the game. This is hardly surprising as XG is of course a development of Gnu in any case!
No doubt a better bot will come along in time, but it won't be much better. Also worth noting that Gnu is in any case better than any human already.
Is it possible that a human "style" could actually play better than either bot? In terms of defeating either in the long run, I think not, but in terms of adapting one's cube play and to a lesser extent one's checker play, one can do a great deal better than just slavishly making the "correct" play. This is also true when playing somebody substantially weaker than oneself, but in either case you do have to have some idea of what the benchmark is in order to be able to vary it!
Bots are tools for learning. Real bg is played between humans. Enjoy the game!