Chess for Android v5.0: Time Control

[Ferdy]

Chess for Android 5.0.5 will be made available shortly on my website and Google Play. New features:

• sorted engine list
• Gauntlet and full tournaments
• support for the open exchange format
• GUI time lag option

Results of a sanity tournament under 1 minute for the full game, both sides of the first position in the Nunn opening suite.

Code: Select all

                              Score     S-B    Ko  Te  Ga  Di  Ci  Bi  Ch  
 1 Komodo 8 32-bit          11.0/12    53.0    XX  10  11  11  11  11  11  
 2 Texel 1.05 32-bit         9.0/12    39.0    01  XX  01  01  11  11  11  
 3 Gaviota v1.0              8.0/12    29.0    00  10  XX  10  11  11  11  
 4 DiscoCheck 5.2.1          8.0/12    29.0    00  10  01  XX  11  11  11  
 5 Cinnamon 1.2b             3.0/12     3.0    00  00  00  00  XX  10  11  
 6 BikJump v1.8 (32-bit)     3.0/12     3.0    00  00  00  00  01  XX  11  
 7 Chess for Android         0.0/12     0.0    00  00  00  00  00  00  XX

Thanks for the update, testing is in progress.

[Carlos777]

Although it would be great if we can choose the number of rounds to be played, not only the 10 games by default.

Note that the option to play from both sides of an external book, like the Nunn opening suite, defines the number of rounds. So you could simply make a PGN with n start positions to get 2n rounds. I like to vary the starting positions, so for that reason I never added the feature to change the number of rounds in the random case.

I am aware of that, but sometimes is tiresome to create pgns each time I want to run a 2 or 4 rr tournament. It would be nice if we can choose the number of rounds and a big book, so the engines could play a random opening n times like in any chessgui for PC.

[Gurcan Uckardes]

On Exynos, after engine invitation, i see RR or gauntlet dialog box. On Intel, the box was the same as 5.0.2. I check the version installed. It's 5.0.5.
Could it be something with Intel?
I'll try direct intall from apk on Intel the way i did on Exynos.

[Gurcan Uckardes]

A new issue is that i have two Komodo 8 in the engine list. One appeared after 5.0.5. Other is still pinned on last position in the list.

[abik]

A new issue is that i have two Komodo 8 in the engine list. One appeared after 5.0.5. Other is still pinned on last position in the list.

That is working as intended (as is the case sensitive sorting of names).

Komodo registers itself through the open and the CB format. I could of course filter out duplicates (although it uses a different name, so it has to be done at package level or something like that). However, I decided not to do that as it is unusual for engines to register through both formats.

[Gurcan Uckardes]

That is working as intended (as is the case sensitive sorting of names

Ok. If it's intended, nothing to say. A second appearance in the bottom can easily be ignored. I think using the first, sorted alphabetically is better.

[Gurcan Uckardes]

On Exynos, after engine invitation, i see RR or gauntlet dialog box. On Intel, the box was the same as 5.0.2. I check the version installed. It's 5.0.5.
Could it be something with Intel?
I'll try direct intall from apk on Intel the way i did on Exynos.

Here's my answer to my own moron question: It takes three engines to choose between RR and Gauntlet. If one selects two engines, RR = Gauntlet and there's nothing to ask to the user.
Sometimes i reach Level-10 of stupidness.

[Ferdy]

Chess for Android 5.0.5 will be made available shortly on my website and Google Play. New features:

• sorted engine list
• Gauntlet and full tournaments
• support for the open exchange format
• GUI time lag option

Results of a sanity tournament under 1 minute for the full game, both sides of the first position in the Nunn opening suite.

Code: Select all

                              Score     S-B    Ko  Te  Ga  Di  Ci  Bi  Ch  
 1 Komodo 8 32-bit          11.0/12    53.0    XX  10  11  11  11  11  11  
 2 Texel 1.05 32-bit         9.0/12    39.0    01  XX  01  01  11  11  11  
 3 Gaviota v1.0              8.0/12    29.0    00  10  XX  10  11  11  11  
 4 DiscoCheck 5.2.1          8.0/12    29.0    00  10  01  XX  11  11  11  
 5 Cinnamon 1.2b             3.0/12     3.0    00  00  00  00  XX  10  11  
 6 BikJump v1.8 (32-bit)     3.0/12     3.0    00  00  00  00  01  XX  11  
 7 Chess for Android         0.0/12     0.0    00  00  00  00  00  00  XX

Done the test on gauntlet at TC 1 minute for whole game, using random book at 10 games each, ponder off.

Deuterium v14.3.34.130 vs
Crafty 24.1
Texel 1.04
Cheng 0.38
Gaviota v1.0
Senpai 1.0
Toga II 3.0
Arasan 17.5

In the first tourney, no Gui time lag (GTL) is used, and Deuterium loses due to illegal move "nomove", this is the move I sent when search could not finish and could not produce a move due to very low on time. It loses 4/70 games like this after 137 or more moves, and it is only Deuterium that has problem in the games .

In the second match I use GTL of 20ms and after 70 games, no more illegal move nomove!!, all games are fine.

In both tourneys, there were no tourney interruptions observed.

I will make some modifications to address this problem, I know my TC is too aggressive.

Very good work Aart congratz.

[Ferdy]

Chess for Android 5.0.5 will be made available shortly on my website and Google Play. New features:

• sorted engine list
• Gauntlet and full tournaments
• support for the open exchange format
• GUI time lag option

Results of a sanity tournament under 1 minute for the full game, both sides of the first position in the Nunn opening suite.

Code: Select all

                              Score     S-B    Ko  Te  Ga  Di  Ci  Bi  Ch  
 1 Komodo 8 32-bit          11.0/12    53.0    XX  10  11  11  11  11  11  
 2 Texel 1.05 32-bit         9.0/12    39.0    01  XX  01  01  11  11  11  
 3 Gaviota v1.0              8.0/12    29.0    00  10  XX  10  11  11  11  
 4 DiscoCheck 5.2.1          8.0/12    29.0    00  10  01  XX  11  11  11  
 5 Cinnamon 1.2b             3.0/12     3.0    00  00  00  00  XX  10  11  
 6 BikJump v1.8 (32-bit)     3.0/12     3.0    00  00  00  00  01  XX  11  
 7 Chess for Android         0.0/12     0.0    00  00  00  00  00  00  XX

I have a sample game where one engine is flag due to time but CFA is also showing 0 time remaining for the other side.

[...]
[3/6]128/e2d3 0.05 } 171...Kf3 { [1/4]-52 0.016 } 172.Ke6 { [3/7]120/a4a6 0.039 } 172...Ra6+ { [1/5]-45/e6f5 0.031 white resigned:Time Control flag vs. 0:00:00 ([8/11]116/f3e3) } 0-1

I also observed a game where CFA prematurely adjudicated to draw a game where one side has knight and the other has a bishop. As I understand help mate is still possible in this game. I am not implying that one side may still win by mate, but just for the sake of the rule . An option would be extending this by 20 plies (10 moves) and then adjudicate to draw if no progress in the score is found.

[pgn]
[Event "Chess for Android Gauntlet"]
[Site "United States"]
[Date "2015.02.17"]
[Round "80"]
[White "Rhetoric 1.4.1 Android"]
[Black "Deuterium v2015.1.35.76"]
[Result "1/2-1/2"]
[Device "LenovoA3300-GV"]
[ECO "A15"]
[TimeControl "60"]

1.Nf3 { opening reply 0.0 } 1...Nf6 { opening reply 0.0 } 2.c4 { opening reply 0.0 } 2...g6 { opening reply 0.001 } 3.g3 { opening reply 0.0 } 3...Bg7 { opening reply 0.001 } 4.Bg2 { opening reply 0.0 } 4...O-O { opening reply 0.001 } 5.O-O { [11/23]19/d7d6 1.581 } 5...d5 { [10/25]3/c4d5 2.031 } 6.cxd5 { [11/25]15/d8d5 1.517 } 6...Nxd5 { [12/24]-21/d2d4 1.921 } 7.e4 { [11/24]25/d5b6 1.448 } 7...Nb6 { [11/27]4/b1c3 1.865 } 8.d4 { [11/24]13/c8g4 1.419 } 8...Bg4 { [10/24]16/h2h3 1.799 } 9.d5 { [11/19]13/c7c6 1.396 } 9...N8d7 { [9/26]-4/h2h3 1.737 } 10.h3 { [10/23]46/g4f3 1.356 } 10...Bxf3 { [11/23]8 1.679 } 11.Qxf3 { [10/27]16/e7e6 1.316 } 11...f5 { [10/25]-6/e4f5 1.625 } 12.Nc3 { [10/22]73/d7c5 1.266 } 12...Rf7 { [9/23]-23 1.574 } 13.Re1 { [10/21]70/g7c3 1.256 } 13...Bxc3 { [11/23]-9/f3c3 1.535 } 14.bxc3 { [10/22]132/d7c5 1.225 } 14...fxe4 { [8/18]-57/f3e4 1.471 } 15.Qxe4 { [10/24]119/d7f6 1.186 } 15...Nf6 { [10/24]-108/e4e6 1.422 } 16.Qd4 { [10/20]125/f6d5 1.143 } 16...Nfxd5 { [10/20]-30/c1h6 1.379 } 17.Rb1 { [10/24]136/c7c5 1.137 } 17...c5 { [9/24]-34/d4d2 1.361 } 18.Qxc5 { [10/24]69/a8c8 1.103 } 18...e6 { [8/19]-19/b1b3 1.291 } 19.Qd4 { [10/24]98/a8c8 1.086 } 19...Rc8 { [9/19]-58/c1d2 1.237 } 20.Rb3 { [9/24]102/c8c4 1.034 } 20...Rc4 { [9/23]-42/d4e5 1.199 } 21.Qe5 { [10/22]122/c4c6 1.016 } 21...Rc6 { [10/24]-34/c1h6 1.162 } 22.Bg5 { [9/18]99/d8f8 0.965 } 22...Qc8 { [9/18]-19/g5h6 1.126 } 23.Bh6 { [8/20]114/f7f6 0.946 } 23...Qc7 { [9/20]-11/e5c7 1.089 } 24.Qd4 { [7/19]114/c6d6 0.937 } 24...Rc4 { [8/22]-31/d4d2 1.06 } 25.Rxe6 { [8/18]260/c7b8 0.548 } 25...Nf6 { [9/23]-220/d4d2 1.031 } 26.Qd6 { [8/18]258/c7d6 0.888 } 26...Qxd6 { [11/19]-203 0.974 } 27.Rxd6 { [9/19]285/c4c8 0.301 } 27...Rc8 { [11/18]-224/h6g5 0.942 } 28.Be3 { [10/19]263/c8e8 0.87 } 28...Nc4 { [11/22]-137/d6f6 0.907 } 29.Rxf6 { [10/20]156/f7f6 0.857 } 29...Rxf6 { [11/21]-116/g2d5 0.883 } 30.Bd5+ { [10/21]162/g8f8 0.831 } 30...Kf8 { [12/20]-112/e3h6 0.855 } 31.Bh6+ { [10/12]162/f8e8 0.802 } 31...Ke7 { [11/21]-110/b3b7 0.616 } 32.Rxb7+ { [9/18]193/e7d6 0.76 } 32...Kd6 { [11/21]-128/d5g2 0.602 } 33.Be4 { [9/18]139/d6e6 0.742 } 33...Rc7 { [10/17]-93/b7c7 0.586 } 34.Bf4+ { [9/18]142/f6f4 0.741 } 34...Rxf4 { [12/21]-79/g3f4 0.57 } 35.gxf4 { [10/20]152/c7b7 0.38 } 35...Rxb7 { [12/19]-76/e4b7 0.559 } 36.Bxb7 { [10/17]152/c4b6 0.264 } 36...Nd2 { [11/15]-79/g1g2 0.541 } 37.Bc8 { [10/16]155/d2e4 0.735 } 37...Ne4 { [11/16]-74/c3c4 0.535 } 38.c4 { [11/17]146/d6c5 0.703 } 38...Kc5 { [12/18]-90/c8e6 0.526 } 39.Be6 { [10/17]153/e4c3 0.696 } 39...Nc3 { [10/18]-113/g1g2 0.526 } 40.f5 { [10/17]191/g6f5 0.661 } 40...gxf5 { [11/16]-92/e6f5 0.489 } 41.Bxf5 { [10/16]137/h7h6 0.647 } 41...h6 { [11/15]-72/f5e6 0.482 } 42.f4 { [9/17]114/c3a2 0.614 } 42...Nxa2 { [11/16]-71/f5d7 0.469 } 43.Be6 { [10/18]150/a2c3 0.62 } 43...Nc3 { [11/15]-51/g1f2 0.451 } 44.f5 { [9/14]109/c3e2 0.578 } 44...Ne4 { [12/17]61/g1g2 0.439 } 45.Bd5 { [10/16]99/e4f6 0.609 } 45...Nf6 { [13/17]103/d5f3 0.437 } 46.Bf3 { [11/15]-17/c5c4 0.593 } 46...Kxc4 { [13/17]115/f3e2 0.418 } 47.Kf2 { [10/16]-16/c4d4 0.559 } 47...a5 { [12/18]77/f3d1 0.408 } 48.Bd1 { [10/14]-13/c4d4 0.571 } 48...Kd4 { [12/17]102/d1a4 0.409 } 49.Bc2 { [10/18]-46/d4e5 0.521 } 49...Ke5 { [12/16]101/f2f3 0.385 } 50.Kf3 { [10/15]-117/f6d5 0.524 } 50...Nd5 { [14/20]126/c2b3 0.385 } 51.h4 { [10/17]-160/d5b4 0.504 } 51...Nb4 { [11/15]174/c2b3 0.369 } 52.Bd1 { [10/18]-173/e5f5 0.487 } 52...Kxf5 { [12/17]156/d1b3 0.36 } 53.Ke3 { [10/16]-190/f5e5 0.489 } 53...Ke5 { [10/15]167/d1b3 0.349 } 54.h5 { [10/18]-165/b4d5 0.444 } 54...Nd5+ { [10/16]167/e3d3 0.34 } 55.Kd3 { [12/18]-174/d5b6 0.461 } 55...Nf6 { [10/14]199/d3c4 0.338 } 56.Kc4 { [11/15]-37/f6e4 0.465 } 56...Ne4 { [10/13]167/d1c2 0.324 } 57.Bc2 { [10/15]-31/e4g3 0.429 } 57...Ng3 { [13/18]126/c2d1 0.316 } 58.Bg6 { [11/16]-28/e5f4 0.486 } 58...Kf4 { [14/18]133/c4b5 0.312 } 59.Kb5 { [11/16]-30/f4g5 0.39 } 59...Kg5 { [12/15]131/b5a5 0.303 } 60.Kxa5 { [11/14]-30/g3h5 0.377 } 60...Nxh5 { [12/15]174/g6e8 0.29 } 61.Bf7 { [11/16]-33/h5f4 0.358 } 61...Nf4 { [13/18]239/a5b5 0.295 } 62.Be8 { [11/14]-33/g5f6 0.365 } 62...Kf6 { [13/17]216/e8d7 0.277 } 63.Kb5 { [10/14]-33/f6e5 0.351 } 63...Ke7 { [15/21]211/e8c6 0.269 } 64.Bc6 { [11/15]-32/e7e6 0.343 } 64...h5 { [15/21]197/c6f3 0.266 } 65.Bf3 { [11/15]-35/h5h4 0.346 } 65...h4 { [15/20]195/f3g4 0.263 } 66.Bg4 { [10/13]-37/f4d3 0.33 } 66...Kd6 { [15/19]194/b5c4 0.264 } 67.Kc4 { [9/16]-33/d6e5 0.309 } 67...Ke5 { [13/20]187/c4c5 0.247 } 68.Kc3 { [9/15]-34/e5f6 0.299 } 68...Kd5 { [11/15]205/g4f3 0.239 } 69.Bf5 { [9/15]-31/d5c5 0.304 } 69...Nh5 { [11/15]179/c3d3 0.231 } 70.Bc8 { [10/14]-33/h5f6 0.302 } 70...Ke4 { [12/15]214/c3c4 0.227 } 71.Kc4 { [10/14]-33/h5f6 0.29 } 71...Nf4 { [13/18]217/c4b5 0.223 } 72.Kc3 { [10/14]-33/f4d5 0.292 } 72...Nd5+ { [11/15]219/c3d2 0.217 } 73.Kd2 { [9/14]-31/e4f4 0.284 } 73...Kd4 { [11/16]204/c8e6 0.212 } 74.Ke2 { [10/14]-17/d4e5 0.273 } 74...Ne3 { [9/13]158/e2f3 0.207 } 75.Kf3 { [11/13]0/d4d3 0.263 } 75...Kd3 { [11/14]33/c8h3 0.202 } 76.Ba6+ { [11/17]0/d3d2 0.252 } 76...Kd2 { [12/16]29/a6c8 0.196 } 77.Bc8 { [13/15]0/d2d3 0.259 } 77...Nc2 { [11/13]13/f3g4 0.19 } 78.Kg2 { [10/15]0/c2e1 0.247 } 78...Nd4 { [11/15]18/g2h3 0.19 } 79.Kh3 { [10/13]0/d4f3 0.245 } 79...Ke3 { [11/16]13/h3h4 0.179 } 80.Bg4 { [12/17]0/d4f3 0.242 } 80...Kf4 { [13/15]0/h3h4 0.177 } 81.Kxh4 1/2-1/2
[/pgn]

[abik]

I have a sample game where one engine is flag due to time but CFA is also showing 0 time remaining for the other side.

[...]
[3/6]128/e2d3 0.05 } 171...Kf3 { [1/4]-52 0.016 } 172.Ke6 { [3/7]120/a4a6 0.039 } 172...Ra6+ { [1/5]-45/e6f5 0.031 white resigned:Time Control flag vs. 0:00:00 ([8/11]116/f3e3) } 0-1

The displayed clock has an accuracy of seconds only (to save screen space mainly), unlike the actual built-in clock. So anything less than a second shows up as 00:00:00 during "the last second". The same logic is used to display remaining time in the saved PGN.

So in this case, white ran out of time, and black still had a few ms. left. If really important, I suppose I could display a higher accuracy in the final PGN report.

As for the insufficient material rule, again I am not sure how important it is to play on in these cases (most tournament managers actually demand that the GUI steps in for more cases rather than playing on). The few built-in rules in Chess for Android on claiming a draw seem rather non-controversial