Vas did contribute!

[bob]

I am not aware of _any_ idea from fruit that is in Crafty of today. I certainly tried ...

At least you are fair and give the source of the idea, othes are not!

Code: Select all

 *   22.2    We are now back to using POSIX threads, since all current Linux   *
 *           distributions now use the posix-conforming NTPL implementation    *
 *           which should eliminate the various compatibility issues that      *
 *           caused problems in the past.  This also should make the new       *
 *           smpnice mode work correctly for Linux and Windows since they will *
 *           now both use threads for the SMP search.  Fruit-like scoring      *
 *           (interpolation between mg and eg scores) fully implemented.  AEL  *
 *           pruning (Heinz 2000) also fully implemented (we had razoring and  *
 *           futility, but not extended futility).  "Eval" command has been    *
 *           removed and combined with the "personality" command so that eval  *
 *           parameters can be modified, in addition to some search parameters *
 *           that also belong in the personality data.  Mate threat extension  *
 *           and one legal reply to check extensions have been removed.  Tests *
 *           proved them to be absolutely useless, over 30,000 games for each  *
 *           test showed no gain and sometimes a loss in playing strength with *
 *           them so we followed the "simple is better" and removed them.  The *
 *           fractional ply code was also removed since the only extension we  *
 *           now use is the "give check" extension which is a whole ply.  A    *
 *           significant number of evaluation parameters have been changed,    *
 *           a few even removed as cluster testing helped us find optimal      *
 *           values.  There are a few new terms, with more planned.            *

I've always tried to give credit where credit is due. Seems like the right thing to do. Not everyone agrees of course.

[adieguez]

Ok, I just remember when Fruit was stronger than shredder and other commercials. Maybe it was fruit 2 already, I don't remember well.

<snipped>

And Amyan 3! way into the future..

Gaviota has surpassed Fruit 1.0? maybe am wrong, I remember even the first version of Fruit was very strong, or not?

I do not know what you define as very strong but

Fruit1.0 is 308 elo weaker than fruit2.1

http://www.fruitchess.com/playing-strength.htm

[michiguel]

Ok, I just remember when Fruit was stronger than shredder and other commercials. Maybe it was fruit 2 already, I don't remember well.

<snipped>

And Amyan 3! way into the future..

Gaviota has surpassed Fruit 1.0? maybe am wrong, I remember even the first version of Fruit was very strong, or not?

I do not know what you define as very strong but

Fruit1.0 is 308 elo weaker than fruit2.1

http://www.fruitchess.com/playing-strength.htm

A not so distant test of fast games

Code: Select all

               Engine   rating     points games      %
         Sloppy-0.1.0	2371.3		347	  560	61.96
          Hermann-2.4	2352.5		332.5	560	59.38
      The Baron 1.7.0	2308.3		297.5	560	53.12
       Gaviota 0.68.1	2305.4		1127	2240	50.31
         fruit-1.0-wt	2289.1		282	  560	50.36
          gosu v.0.16	2282.3		276.5	560	49.38
         romichess.sh	2281.0		275.5	560	49.2
              PMv1016	2275.5		271	  560	48.39
       Gaviota 0.66.3	2267.8		1007	2240	44.96
         Xpdnt 080904	2266.8		264	  560	47.14

Fruit 1.5, we are coming for you!

Miguel

[schlucke]

I am not aware of _any_ idea from fruit that is in Crafty of today. I certainly tried ...

At least you are fair and give the source of the idea, othes are not!

Code: Select all

 *   22.2    We are now back to using POSIX threads, since all current Linux   *
 *           distributions now use the posix-conforming NTPL implementation    *
 *           which should eliminate the various compatibility issues that      *
 *           caused problems in the past.  This also should make the new       *
 *           smpnice mode work correctly for Linux and Windows since they will *
 *           now both use threads for the SMP search.  Fruit-like scoring      *
 *           (interpolation between mg and eg scores) fully implemented.  AEL  *
 *           pruning (Heinz 2000) also fully implemented (we had razoring and  *
 *           futility, but not extended futility).  "Eval" command has been    *
 *           removed and combined with the "personality" command so that eval  *
 *           parameters can be modified, in addition to some search parameters *
 *           that also belong in the personality data.  Mate threat extension  *
 *           and one legal reply to check extensions have been removed.  Tests *
 *           proved them to be absolutely useless, over 30,000 games for each  *
 *           test showed no gain and sometimes a loss in playing strength with *
 *           them so we followed the "simple is better" and removed them.  The *
 *           fractional ply code was also removed since the only extension we  *
 *           now use is the "give check" extension which is a whole ply.  A    *
 *           significant number of evaluation parameters have been changed,    *
 *           a few even removed as cluster testing helped us find optimal      *
 *           values.  There are a few new terms, with more planned.            *

I've always tried to give credit where credit is due. Seems like the right thing to do. Not everyone agrees of course.

Yea, it is right. At least you have clear conscience. But that seems to be very unpopular these days ... not only in chess programming!