King of Kings 2.30

[Koundinya Veluri]

Geschrieben von: / Posted by: Koundinya Veluri at 19 August 2003 23:23:40:

Hi,
King of Kings 2.30 is available now on my site. Just a bugfix and PV extensions. It seems to solve many tactical problems at lower plies but also takes longer to search the plies, so I don't really know if it's stronger or not, but it does seem to be a productive change.
Regards,
Koundinya


King of Kings web site

[Steve Maughan]

Geschrieben von: / Posted by: Steve Maughan at 20 August 2003 02:54:13:
Als Antwort auf: / In reply to: King of Kings 2.30 geschrieben von: / posted by: Koundinya Veluri at 19 August 2003 23:23:40:

Hi,

[...] and PV extensions.

What do you mean by PV extensions? Are you entending every time the move played is the same as the PV move?
Regards,
Steve

[Koundinya Veluri]

Geschrieben von: / Posted by: Koundinya Veluri at 20 August 2003 05:05:46:
Als Antwort auf: / In reply to: Re: King of Kings 2.30 geschrieben von: / posted by: Steve Maughan at 20 August 2003 02:54:13:

Hi,

[...] and PV extensions.

What do you mean by PV extensions? Are you entending every time the move played is the same as the PV move?
Regards,
Steve

I guess it's a misnomer. I extend the first move searched by a quarter of a ply. These extensions add up in a variation and so the actual depth in the PV and other important variations would be larger. To save some speed though, I don't extend throughout the tree.
Regards,
Koundinya

[Robert Allgeuer]

Geschrieben von: / Posted by: Robert Allgeuer at 20 August 2003 12:15:47:
Als Antwort auf: / In reply to: Re: King of Kings 2.30 geschrieben von: / posted by: Koundinya Veluri at 20 August 2003 05:05:46:

Hi,

[...] and PV extensions.

What do you mean by PV extensions? Are you entending every time the move played is the same as the PV move?
Regards,
Steve

I guess it's a misnomer. I extend the first move searched by a quarter of a ply. These extensions add up in a variation and so the actual depth in the PV and other important variations would be larger. To save some speed though, I don't extend throughout the tree.
Regards,
Koundinya

Did you play a match between old and new version? The result would be interesting.
Robert

[Koundinya Veluri]

Geschrieben von: / Posted by: Koundinya Veluri at 20 August 2003 18:13:10:
Als Antwort auf: / In reply to: Re: King of Kings 2.30 geschrieben von: / posted by: Robert Allgeuer at 20 August 2003 12:15:47:

Hi,

[...] and PV extensions.

What do you mean by PV extensions? Are you entending every time the move played is the same as the PV move?
Regards,
Steve

I guess it's a misnomer. I extend the first move searched by a quarter of a ply. These extensions add up in a variation and so the actual depth in the PV and other important variations would be larger. To save some speed though, I don't extend throughout the tree.
Regards,
Koundinya

Did you play a match between old and new version? The result would be interesting.
Robert

Yes I did, at 40/5 time control with pondering. The result was +16-15=45. Probably not enough games to measure the difference. The new version is solving tests faster on average. Also, I have some positions that previous versions had a difficult time solving. Some of these were solved at a reasonable time by the new version. Three of them were endgame positions involving high depth and another involved repetition detection at high depth to avoid a move. I guess the extensions really add up in these high depth searches because it was solving these at a relatively low depth. There were several positions though that the new version couldn't solve one or more plys earlier, and because of the extra time it takes to finish a ply, it took longer to solve them. These positions were relatively few so I figured it must be a good change.
Regards,
Koundinya