Chess for Android v5.0: Time Control

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Carlos777
Posts: 1731
Joined: Sun Dec 13, 2009 6:09 pm

Re: Chess for Android v5.0: Time Control

Post by Carlos777 »

Ferdy wrote:
abik wrote:After some delay, I finally released version 5.0 of Chess for Android, available on Google Play or as direct download. New features include:
  • * Time control to define a chess clock for both sides with a choice between the number of moves between time control (possibly the full game), thinking time per time control, and an optional time increment per move.
    * Double engine analysis lines during tournaments, especially useful with pondering on.
    * Re-factored a lot of the engine handling code, with improved cleanup of hanging engines.
This is a relatively major release, so please let me know if you encounter any issues. Also, as usual, suggestions for improvements are welcome.
Feature request again.
Gauntlet type matches.
a vs b, a vs c, a vs d and so on.
Very convenient for testing new engine or new version of engine.
+1
User avatar
abik
Posts: 819
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: Chess for Android v5.0: Time Control

Post by abik »

mclane wrote:And what is with 40/120 20/60 ?
With the clock implementation in place, adding different 2nd and 3nd time control is very straightforward. The most work will be in providing the GUI to enter it :-)

Just curious, is having different time controls really that important?
User avatar
abik
Posts: 819
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: Chess for Android v5.0: minor issues

Post by abik »

supersharp77 wrote:Great work as always Aart! The app is great .....flawless for the most part (except for the shadow engines that show up and won't disappear from the import screen Fruit reloaded and rodent right now) other than that only other thing is some engines will not work regardless of what you do....Firenzina will not play tournaments vs komodo 2.03....APIL plays but fails in tourneys.....same for Rocinante 2.0 Gaviota 1.0 and gaviota 0.86 not working but gaviota 0.85.1 and 0.84 and 0.83 are ok...Textel 1.04 and 1.05 not working for me....SF 3 not working for me but others have no issues but the other (Stockfishes) are......Ifrit crashes quickly and adachess crashes.....Myriddin plays but won't play in tournaments......buzz plays but crashes in tournaments.....I think these are engine and not gui problems?!?........unknown..... Thx Again AAR :)
Thanks! Yes, engines often simply exit for no apparent reason, and it is hard for the GUI to report back why this happens. I typically do some test matches first, and only proceed with the stable ones.

If engine authors are motivated, I am happy to work with them to debug this further.
User avatar
abik
Posts: 819
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: Chess for Android v5.0: Time Control

Post by abik »

Carlos777 wrote:
Ferdy wrote: Feature request again.
Gauntlet type matches.
a vs b, a vs c, a vs d and so on.
Very convenient for testing new engine or new version of engine.
+1
Request noted :-)
Gurcan Uckardes
Posts: 196
Joined: Wed Oct 29, 2014 12:42 am

Re: Chess for Android v5.0: Time Control

Post by Gurcan Uckardes »

Yes. TC is nothing but a command. I think you better implement the same dialog box aa the alarm and users can set whatever they want and the gui just translates it to a uci or xb command.
Weird tc settings can also help scaling between different cpus. Like ccrl for ex.
My blog for Android users: http://chesstroid.blogspot.com
User avatar
abik
Posts: 819
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: Chess for Android v5.0: Time Control

Post by abik »

Some more thorough tests with ponder on revealed another issue that will be fixed in the next release. At least one engine sometimes sends "ponder 0000" (which according to the standard is a null move, which is not the same as no ponder move imho), which causes the GUI to send information that crashed that engine. Safest way around the issue is to verify the ponder move.
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: Chess for Android v5.0: Time Control

Post by Ferdy »

abik wrote:Some more thorough tests with ponder on revealed another issue that will be fixed in the next release. At least one engine sometimes sends "ponder 0000" (which according to the standard is a null move, which is not the same as no ponder move imho), which causes the GUI to send information that crashed that engine. Safest way around the issue is to verify the ponder move.
Some uci engines and their ponder move. I set the Ponder option to on and give a position of mate in 1, so there is nothing to ponder actually. Position given.
[d]8/8/5k2/8/P4b2/1B3p2/8/Kbq5 b - - 7 71
id name Andscacs 0.71
Ponder is supported
bestmove f4e5

id name Arasan 17.4
Ponder is supported
bestmove f4e5

id name Atlas 3.50 x64
Ponder is supported
bestmove f4e5

id name BlackMamba_1.2c x64
Ponder is supported
bestmove f4e5 ponder NULL

id name Bobcat 3.25
Ponder is supported
bestmove f4e5

id name cheng4 0.36c
Ponder is supported
bestmove f4e5

id name Critter 1.6a 64-bit
Ponder is supported
bestmove f4e5

id name Deuterium v14.2.33.276
Ponder is supported
bestmove f4e5

id name Fire 4 x64
Ponder is supported
bestmove f4e5

id name Fruit reloaded 2.1
Ponder is supported
bestmove f4e5

id name Gaviota v1.0
Ponder is supported
bestmove f4e5

id name Glass 2.0 PERSONALITY
Ponder is supported
bestmove f4e5

id name GNU Chess 5.50-64
Ponder is supported
bestmove f4e5

id name Godel 2.3.7
Ponder is supported
bestmove f4e5

id name GreKo 12.0
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Gull 2.2 x64
Ponder is supported
bestmove f4e5

id name Hakkapeliitta 2.0
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Hamsters 0.7.1
Ponder is supported
bestmove f4e5

id name Hannibal 1.4x64
Ponder is supported
bestmove f4e5


id name HIARCS 14 WCSC
Ponder is supported
bestmove f4e5

id name Houdini 4 x64
Ponder is supported
bestmove f4e5 ponder 0000

id name Komodo 6 64-bit
Ponder is supported
bestmove f4e5 ponder 0000

id name MadChess 1.4
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Maverick 0.51 x64
Ponder is supported
bestmove f4e5

id name MinkoChess 1.3 x64
Ponder is supported
bestmove b1c2

id name Murka 3 x64 UCI
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Nebula 2.0
Ponder is supported
bestmove f4e5

id name Nemo SP64o 1.0.1 Beta
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Nirvanachess 1.8
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Pawny 1.0.x64.SSE4.2
Ponder is supported
bestmove f4e5

id name Quazar 0.4 x64
Ponder is supported
bestmove f4e5

id name RedQueen 1.1.4
Ponder is supported
bestmove f4e5

id name Rhetoric 1.4.1 x64
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Rodent 1.5 (build 15)
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Rotor 0.8
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Ruffian 1.0.5
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Rybka 2.3.2a mp
Ponder is supported
bestmove f4e5

id name Senpai 1.0
Ponder is supported
bestmove f4e5

id name SmarThink 1.70
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name spark-1.0
Ponder is supported
bestmove f4e5

id name Stockfish 5 64 SSE4.2
Ponder is supported
bestmove f4e5 ponder (none)

id name Strelka 5.5
Ponder is not supported <<<<<<<<<<<<
bestmove f4e5

id name Twisted Logic 20100131x
Ponder is supported
bestmove f4e5

id name Texel 1.04 64-bit
Ponder is supported
bestmove f4e5

id name TJchess 1.1U-x64
Ponder is supported
bestmove f4e5

id name Toga II 3.0
Ponder is supported
bestmove f4e5

id name Ufim 8.02
Ponder is supported
bestmove f4e5

id name Umko 1.1 x64
Ponder is supported
bestmove f4e5

id name Vajolet2 1.45
Ponder is supported
bestmove f4e5
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: Chess for Android v5.0: Time Control

Post by Ferdy »

abik wrote:Some more thorough tests with ponder on revealed another issue that will be fixed in the next release. At least one engine sometimes sends "ponder 0000" (which according to the standard is a null move, which is not the same as no ponder move imho), which causes the GUI to send information that crashed that engine. Safest way around the issue is to verify the ponder move.
In another case where the side to move has no legal move, lets see what they are sending. Others just crashed.

Code: Select all

id name Andscacs 0.71
Ponder is supported
bestmove a1a1

id name Arasan 17.4
Ponder is supported
bestmove 0000

id name Atlas  3.50  x64 
Ponder is supported
bestmove a1a1

id name BlackMamba_1.2c x64
Ponder is supported
bestmove NULL

id name cheng4 0.36c
Ponder is supported
bestmove 0000

id name Critter 1.6a 64-bit
Ponder is supported
bestmove null

id name Deuterium v14.2.33.276
Ponder is supported
bestmove nomove

id name Gaviota v1.0
Ponder is supported
bestmove 0000

id name GNU Chess 5.50-64
Ponder is supported
bestmove nomove

id name Hamsters 0.7.1
Ponder is supported
bestmove **

id name Hannibal 1.4x64
Ponder is supported

id name Houdini 4 x64
Ponder is supported
bestmove NULL

id name Komodo 6 64-bit 
Ponder is supported
bestmove 0000

id name Maverick 0.51 x64
Ponder is supported
bestmove 0000

id name MinkoChess 1.3 x64
Ponder is supported
bestmove a1a1

id name Murka 3 x64 UCI
Ponder is not supported <<<<<<<<<<<<
bestmove a8a8

id name Nebula 2.0
Ponder is supported
bestmove a1a1

id name Nemo SP64o 1.0.1 Beta 
Ponder is not supported <<<<<<<<<<<<
bestmove a1a1 

id name Nirvanachess 1.8
Ponder is not supported <<<<<<<<<<<<
bestmove `&#58;`&#58;

id name Pawny 1.0.x64.SSE4.2
Ponder is supported
bestmove a1a1 

id name RedQueen 1.1.4
Ponder is supported
bestmove a4a5

id name Rhetoric 1.4.1 x64
Ponder is not supported <<<<<<<<<<<<
bestmove 0000

id name Rodent 1.6 &#40;build 6&#41;
Ponder is not supported <<<<<<<<<<<<
bestmove a1a1

id name Rotor 0.8
Ponder is not supported <<<<<<<<<<<<
bestmove a1a1 

id name Rybka 2.3.2a mp 
Ponder is supported
bestmove a1a1

id name Senpai 1.0
Ponder is supported
bestmove 0000

id name spark-1.0
Ponder is supported
bestmove 0000

id name Stockfish 5 64 SSE4.2
Ponder is supported
bestmove &#40;none&#41; ponder &#40;none&#41;

id name Strelka 5.5
Ponder is not supported <<<<<<<<<<<<
bestmove a1a1

id name Texel 1.04 64-bit
Ponder is supported
bestmove 0000

id name TJchess 1.1U-x64
Ponder is supported
bestmove 

id name Toga II 3.0
Ponder is supported
bestmove a1a1

id name Ufim 8.02
Ponder is supported
bestmove h h

id name Umko 1.1 x64
Ponder is supported
bestmove a1a1

id name Vajolet2 1.45
Ponder is supported
bestmove 0000
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Chess for Android v5.0: Time Control

Post by cdani »

Ferdy wrote: In another case where the side to move has no legal move, lets see what they are sending. Others just crashed.
I suppose most engines like Andscacs simply does not spend time verifying those cases. Anyway I will fix it. Thanks!
User avatar
abik
Posts: 819
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: Chess for Android v5.0: Time Control

Post by abik »

Ferdy wrote:Some uci engines and their ponder move.
Thanks for this extremely useful experiment! Quite some variation still in how to handle *no* ponder move. Never a dull moment in a GUI implementor's life, is there?
Ferdy wrote:In another case where the side to move has no legal move, lets see what they are sending. Others just crashed.
This case was also explored in an earlier thread at this forum. My GUI avoids querying the engine in such situations, but it is good to raise awareness of this again.