RAPIDROID RELOADED: June-2015 // Komodo 9 on top !

[Gurcan Uckardes]

Original post and comments: http://chesstroid.blogspot.com/2015/06/ ... odo-9.html

Code: Select all

BAYES RATINGS AFTER 6346 GAMES PLAYED BY 115 PROGRAMS
Rnk Name                     O/S T  Elo   +   - gam sco oppo dra
001 Komodo 9                 A32 4 3239  57  54 120 78% 3041 36%
002 Stockfish 6              A32 4 3223  55  53 120 76% 3043 41%
003 Critter 1.6a             A32 4 3110  50  49 120 58% 3056 51%
004 Sting SF 4.8.4 JA        A32 4 3078  53  52 120 58% 3021 37%
005 Firenzina 2.4.1 xTreme   A32 4 3061  49  49 120 50% 3062 51%
006 Texel 1.05               A32 4 3039  53  53 120 55% 2998 33%
007 BeeKay                   A32 4 3019  51  51 120 47% 3038 45%
008 BlackMamba 2.0           A32 4 3018  50  50 120 48% 3031 53%
009 DeepSaros ver.2.3f       A32 1 2997  54  53 120 53% 2957 37%
010 Senpai 1.0               A32 4 2973  56  55 120 58% 2919 27%
011 RobboLito 0.085e4l       A32 1 2929  54  54 122 47% 2956 33%
012 Hiarcs 13.71             IOS 2 2877 172 157  12 71% 2755 25%
013 cheng4 0.38              A32 4 2876  53  53 120 49% 2891 33%
014 Gaviota v1.0-d           A32 4 2849  53  53 122 48% 2863 33%
015 Grapefruit 1.0           A32 4 2810  51  51 124 51% 2801 39%
016 DiscoCheck 5.2.1         A32 1 2801  53  53 126 48% 2814 37%
017 Arasan 17.5              A32 4 2787  52  53 122 43% 2844 37%
018 Cyclone 3.4              A32 4 2779  51  51 124 52% 2770 37%
019 Deep Saros 0.9           A32 4 2751  52  52 124 49% 2751 32%
020 Doch32 1.3.4 JA          A32 1 2749  53  53 116 47% 2774 34%
021 Toga II 3.0              A32 1 2742  52  52 122 48% 2760 32%
022 Bobcat 6.4b              A32 4 2727  53  53 124 50% 2726 26%
023 Deuterium v14.3.34.130   A32 1 2709  50  50 130 59% 2640 42%
024 Fruit reloaded 2.1       A32 1 2696  52  52 114 50% 2696 43%
025 GNU Chess 5.60-32        A32 1 2687  53  52 124 58% 2627 31%
026 Chess Pro 2013.09        IOS 1 2674 269 269   4 50% 2691  0%
027 Murka 3 ARM UCI          A32 1 2655  50  49 132 53% 2633 35%
028 Hakkapeliitta dev 120115 A32 1 2648  52  52 130 53% 2621 24%
029 Strelka 5                A32 1 2629  52  51 130 56% 2586 25%
030 TheMadPrune 1.7.04       A32 4 2622  55  56 112 47% 2643 26%
031 Scorpio_2.7.7.JA_xb      A32 1 2615  54  54 112 46% 2648 34%
032 Rodent 1.7 build 1       A32 1 2606  51  51 130 48% 2621 29%
033 IvanHoe 9.46b            A32 4 2603  50  50 132 50% 2602 34%
034 Crafty_24.1.JA_xb        A32 1 2594  54  53 130 55% 2549 19%
035 Tucano_5.00.JA_xb        A32 1 2574  58  58 102 45% 2612 31%
036 Bison 15.1               A32 1 2536  52  52 130 52% 2524 27%
037 Rhetoric 1.4.1           A32 1 2516  50  50 132 50% 2523 34%
038 Gull 1.2 JA              A32 1 2509  51  51 130 52% 2489 32%
039 RedQueen 1.1.97          A32 4 2507  52  52 132 47% 2524 22%
040 CNVCS 1.2.0              A32 1 2505 226 243   4 38% 2570 25%
041 Alfil 12.10              A32 1 2467  51  52 132 46% 2491 26%
042 Chess Genius 4.0.00      IOS 1 2466 249 304   4 13% 2691 25%
043 Cheese 1.7               A32 1 2455  50  51 134 43% 2507 31%
044 Daydreamer 1.75 JA       A32 1 2441  52  52 128 50% 2446 27%
045 GarboChess 3             A32 1 2427  53  53 128 48% 2446 23%
046 Chess Genius 2.6.4       A32 1 2407 223 238   4 38% 2454 75%
047 Glaurung Mainz           A32 1 2401  63  65  88 42% 2458 25%
048 Rotor 0.8                A32 1 2395  52  53 128 39% 2479 29%
049 BBChess 1.3b JA          A32 4 2379  53  53 126 50% 2378 23%
050 Danasah_5.07.JA_xb       A32 1 2377  54  54 122 54% 2347 23%
051 DanasahZ_0.4.JA_xb       A32 1 2377  52  52 124 46% 2404 28%
052 Dirty_030411.JA_xb       A32 1 2353  54  54 118 46% 2388 27%
053 Sloppy_0.23.JA_xb        A32 1 2345  52  52 120 54% 2316 33%
054 Phalanx_XXIV.JA_xb       A32 1 2327  54  54 124 49% 2337 17%
055 GreKo_12.5.JA_xb         A32 1 2315  53  53 122 57% 2266 26%
056 BetsabeII_1.47.JA_xb     A32 1 2298  57  57 120 48% 2314 15%
057 Pepito v1.59             A32 1 2289  51  51 126 49% 2291 33%
058 Maverick 1.0 JA          A32 1 2276  53  53 126 50% 2275 21%
059 Pawny_1.0.JA_uci2xb      A32 1 2263  54  54 126 55% 2230 19%
060 Diablo 0.5.1b JA         A32 1 2229  52  52 126 47% 2255 25%
061 Typhoon_1.0.r358.JA_xb   A32 1 2225  54  54 120 56% 2174 26%
062 Ifrit_m1.8.JA_uci2xb     A32 1 2196  53  54 126 47% 2212 25%
063 Olithink_5.3.2.JA_xb     A32 1 2195  56  55 122 57% 2137 18%
064 Amy_0.8.JA_xb            A32 1 2189  54  54 126 47% 2213 22%
065 TJchess 1.1U             A32 1 2168  54  54 126 53% 2146 22%
066 Myrddin_0.86.JA_xb       A32 1 2154  59  57 124 60% 2059 18%
067 MangoPaola_1.1.JA_xb     A32 1 2149  56  56 118 50% 2143 22%
068 Natwarlal_0.14.JA_xb     A32 1 2130  55  55 118 53% 2106 24%
069 Sungorus 1.4 JA          A32 1 2082  55  57 122 39% 2167 18%
070 KmtChess_1.21.JA_xb      A32 1 2044  55  56 120 48% 2062 25%
071 Scidlet_2.61b2.JA_xb     A32 1 2036  58  57 124 57% 1974 10%
072 Rattate_Nosferatu.JA_xb  A32 1 2030  56  56 118 50% 2030 21%
073 Resp_0.19.JA_xb          A32 1 2024  55  55 126 51% 2008 17%
074 NGplay_9.86.JA_xb        A32 1 2021  56  56 118 51% 2011 22%
075 Zurichess Basel          A32 1 1984  93  93  54 57% 1913 22%
076 DanChess_1.04.JA_xb      A32 1 1948  56  56 122 50% 1949 16%
077 Witz_Alpha21.JA_xb       A32 1 1887  56  56 116 49% 1903 22%
078 Kurt 0.9.2.2 JA          A32 1 1876  55  55 126 50% 1883 17%
079 Robocide 281214 JA       A32 1 1855  54  55 128 41% 1921 18%
080 Woodpecker_2.11.JA_xb    A32 1 1849  58  57 118 55% 1813 14%
081 Clubfoot_r2293.JA        A32 1 1829  57  57 122 48% 1849 12%
082 Knightcap_3.7F.JA_xb     A32 1 1824  57  57 118 51% 1819 16%
083 BikJump v1.8             A32 1 1799  55  55 124 53% 1780 19%
084 Leonidas_r83.JA_xb       A32 1 1792  57  57 120 49% 1799 12%
085 Cilian_4.14.JA_xb        A32 1 1786  53  52 126 56% 1742 30%
086 Sjeng_1.12.JA_xb         A32 1 1780  56  57 118 44% 1831 18%
087 AdroitChess0.4 JA        A32 1 1778  57  57 122 50% 1773 16%
088 ZCT-0.3.2500             A32 1 1774  56  57 124 43% 1830 12%
089 Samchess_JA_xb           A32 1 1765  57  57 116 44% 1819 18%
090 Ecce rev. 508            A32 1 1758  55  55 124 49% 1773 17%
091 Faile_1.44.JA_xb         A32 1 1757  56  55 108 56% 1717 33%
092 Surprise_4.3.b13.JA_xb   A32 1 1721  79  78  66 55% 1677  9%
093 Colchess_8.0.JA_xb       A32 1 1691  58  58 108 52% 1676 22%
094 Smash 1.03 JA            A32 1 1647  57  57 120 48% 1662 14%
095 Bismark_1.1              A32 1 1625  53  53 122 48% 1641 32%
096 Claudia v. 0.5           A32 1 1575  76  76  66 52% 1565 21%
097 Zzzzzz_3.5.1.JA_xb       A32 1 1571  58  58 100 54% 1513 36%
098 Jester_0.84.JA_xb        A32 1 1569  66  65  98 56% 1495  8%
099 Rocinante 2.0 JA         A32 1 1568  57  57 122 51% 1533 20%
100 Hoichess_0.12.1.JA_xb    A32 1 1554  57  58 120 52% 1518 20%
101 Kitteneitor_060513.JA_xb A32 1 1549  58  59  98 46% 1555 40%
102 Gunborg_1.0.JA_uci2xb    A32 1 1529  58  58 122 57% 1432 21%
103 Tscp_1.8.1.AB_xb         A32 1 1512  62  63 102 43% 1547 16%
104 Colossus 4.0 100X        C64 1 1496 274 227   8 88% 1092 25%
105 Chenard_2014.07.11.JA_xb A32 1 1494  68  69  96 48% 1483 10%
106 Pulse 1.5-cpp            A32 1 1457  60  60 124 60% 1313 26%
107 Chess for Android        A32 1 1331  67  68 114 47% 1315 14%
108 VIRUTOR CHESS 1.1.1      A32 1 1322  67  68 118 55% 1243  8%
109 K2 v.075                 A32 1 1282  80  81  80 44% 1328  9%
110 Trappy_Beowulf_2.0.JA_xb A32 1 1131  69  70 112 49% 1133  8%
111 Xadreco_5.7.JA_xb        A32 1  987  84  92  72 23% 1254  7%
112 Byak 8.10.14.JA          A32 1  984  81  88  76 22% 1254 12%
113 Colossus 4.0             C64 1  981 258 160   8 38% 1092  0%
114 Joanna_2002.JA_xb        A32 1  837 116  28  74 19% 1224  8%
115 OliveChess 0.2.7         A32 1  604 255-201  78  3% 1266  4%

Rapidroid test platform:
* GT-N7100 1.7 Ghz quad, 256MB hash: All Android progs
* GT-N5105 1.7 Ghz quad, 256MB hash: All Android progs
* Codegen Novatab tablet 1.4 Ghz quad, 256MB hash: All Android progs
* Polypad 1010IPS tablet 1.61 Ghz dual, 128MB hash: Single thread Android progs
* HTC Diam 528Mhz, 16MB hash: Windows Mobile
* i7 M620 2.67 Ghz dual + Arena 3.5 + 2GB hash: Windows 64
* iPhone5S A7 1.3Ghz dual: iOS programs
* DosBox 1.74: DOS progs
* WinVICE 2.24: Commodore-64 progs
* Messtiny UCI adapters or CB-Emu2014: Mephisto progs
* Openings: 20 ply from Adam Hair or 16 ply from TCEC, no Q exchange, +0.15 to +0.40 by Stockfish/Komodo @ depth 20 min, played twice each color.
* Repeating openings and twin games not allowed between two programs
* Tablebases and pondering off
* Time control: 10 to 30 sec/move or 600+0 to 1800+5 or closest known by both programs

[Ferdy]

Original post and comments: http://chesstroid.blogspot.com/2015/06/ ... odo-9.html

Good work.
BTW is there any reason why there are two DeepSaros/Deep Saros?

Doch32 1.3.4 JA and BeeKay, does Komodo come from these engines? Perhaps to save some test time, just use Komodo, unless of course if Doch and Beekay are still under development.

My only concern is you already have too much load to take these engines.

Nice to see Hiarcs, I have this in ipad.

[pedrox]

Hello Gurcan,

Thanks for the list. I want to tell a few things on DanaSah engine.

1. The difference between DanaSah 5.07 and DanaSahZ 0.4

DanaSah 5.07 is possibly the strongest version of my engine in the PC list, it is approximately 20 Elo points above the rest.

DanaSahZ 0.4 is a version of DanaSah (part of DanaSah 5.00) that is designed to play with reduced force levels and has Elo regulation.

In your tournament you play with DanaSahZ 0.4 without force reduction, which is why in the list are very close, but as I said before I think the 5.07 version should be about 20 strong points. I think it is not necessary to test DanaSahZ 0.4 and you save one engine.

Another thing is that you want to try reducing force in DanaSahZ 0.4. To this must the engine has 2 parameters that can be changed, one is "easylevel" and another is "elo". These two parameters are set in the configuration file or passed to the engine when it starts.

For example you can create a text file (and install it and import from Chess for Android, you have to do this with the engine and the text file).

./danaz04 elo 1200

or

./danaz04 easylevel 3

In the first case the engine play with an ELO of 1200 and in the second case would play with a random elo between 1000 and 1199. More information in the readme file or web page.


2. I have seen that there are some engines on your list who can play well with a fixed time per move and when they are set to play time with increase, then move instantly. this also happens with danasah.

I think this is not a engine bug. DanaSah played correctly in all GUI except Chess for Android.I think Aart must have an error when use time with increase for xboard protocol, he must do something different to other GUI.

[Gurcan Uckardes]

Two Saros' in the list may be logical. I've read they are based on different sources.
Same applies to two Danasah. Versions Z and 5.07 are assumed to be different, though I didn't insist to verify.
Regarding Beekay and Doch, yes they are ancient Komodos. I've already taken out Komodo 3 to reduce noise in the list but maybe it's not enough.
Keeping Doch may be defendable but I guess Beekay can be removed. It's very similar to Komodo CCT and Komodo 5.1. The gap behind Komodo9 is not large enough to justify its presence.
By the way, Fruit derivatives in the list are also subject to similar questions.

[Gurcan Uckardes]

Hello Gurcan,

Thanks for the list. I want to tell a few things on DanaSah engine.

1. The difference between DanaSah 5.07 and DanaSahZ 0.4

DanaSah 5.07 is possibly the strongest version of my engine in the PC list, it is approximately 20 Elo points above the rest.

DanaSahZ 0.4 is a version of DanaSah (part of DanaSah 5.00) that is designed to play with reduced force levels and has Elo regulation.

In your tournament you play with DanaSahZ 0.4 without force reduction, which is why in the list are very close, but as I said before I think the 5.07 version should be about 20 strong points. I think it is not necessary to test DanaSahZ 0.4 and you save one engine.

Another thing is that you want to try reducing force in DanaSahZ 0.4. To this must the engine has 2 parameters that can be changed, one is "easylevel" and another is "elo". These two parameters are set in the configuration file or passed to the engine when it starts.

For example you can create a text file (and install it and import from Chess for Android, you have to do this with the engine and the text file).

./danaz04 elo 1200

or

./danaz04 easylevel 3

In the first case the engine play with an ELO of 1200 and in the second case would play with a random elo between 1000 and 1199. More information in the readme file or web page.


2. I have seen that there are some engines on your list who can play well with a fixed time per move and when they are set to play time with increase, then move instantly. this also happens with danasah.

I think this is not a engine bug. DanaSah played correctly in all GUI except Chess for Android.I think Aart must have an error when use time with increase for xboard protocol, he must do something different to other GUI.

Thanks Pedro,
Yes. If DanasahZ is nothing but v5.00 with trimmable strength, I better keep 5.07 only.
I will also check about time usage issues and replace all games if necessary.
Please tell me more about the other engines which play fixed time well and Fischer clock incorrectly.
To be more precise, Rapidroid gives priority to Fischer clock.
All engines are classified in 3 categories:
1) FTTC: Fischer and fixed time both run fine.
2) TC: Only Fischer clock
3) FT: Only fixed time.
Then pairings are calculated according to the categories and each division is played in two steps. Well 2nd and 3rd categories can never play each other by definition. That's why Zurichess for example has to play fewer games.
Category 1 is passe-partout here. They can play with any engine.

[Gurcan Uckardes]

I checked my time usage categories: Both Danasah versions are set to FIXED TIME only.
With Fischer clock Danasah 5.07 is marked as using extra time and Danasah Z as instant move.
Therefore, everything was already set correctly for these engines and no worries


All engines which don't comply to Fischer clock are tested and detected before introduction to Rapidroid. If they play fine with fixed time they are marked category 3: FT

Again, no worries.

[pedrox]

Hi Gurcan,

If you play as a human against DanaSahZ 0.4 engine on Chess for Android and you uses Fischer time control, you can see that the engine does not move instantaneously. However when the engine is playing with another engine , then if it does, which it makes no sense.

It is for this reason that I believe that Chess for Android should have some small error in the xboard protocol and when the confrontation is against another engine. I should point out this problem to Art , but I find difficult to express me in English and in recent months I had no time to chess computer. If other engines also move instantly and they have the xboard protocol , maybe the error is not the engine , maybe it is the GUI.

When you do a test to see if the engine pass Fischer control time, what time control you uses?

I have to admit that I have not tried Fischer time control, because it was not implemented just the latest versión, but here a list with engines (xboard) which fail to Fischer control time on Chess for Android. It is almost 75 % of engines with this problem, for me this is a bit strange.

Code: Select all

Engine Pro Arch Fixed Fischer 
Amy 4.8 xb arm5 Pass Pass 
Beowulf 2.04a.JA xb arm5 Fail Fail 
Betsabe 1.47.JA xb arm5 Pass Fail 
Bruja 1.9.1.JA xb arm5 Pass* Fail 
Buzz 11-27-08.JA xb arm5 Fail Fail 
Buzz OS 1.0.JA xb arm5 Fail Fail 
Cassandre 0.26.JA xb arm5 Fail Fail 
Chenard 2014.07.11.JA xb arm5 Pass Pass 
Chess One 2.01.JA xb arm5 Fail Fail 
Chess Rikus 1.466.JA xb arm5 Fail Fail 
Cilian 4.14.JA xb arm5 Pass Pass 
Colchess 8.0.JA xb arm5 Pass* Fail 
Crafty 24.1.JA xb arm5 Pass Pass 
Dabbaba 7.49.JA xb arm5 Fail Fail 
Danasah 5.07.JA xb arm5 Pass Fail 
DanasahZ 0.4 xb arm5 Pass Fail 
DanChess 1.04.JA xb arm5 Pass Fail 
Darky 0.5d xb arm5 Fail Fail 
Dirty 030411.JA xb arm5 Pass Fail 
Exchess 7.51.beta.JA xb arm5 Fail Fail 
Faile 1.4.4.JA xb arm5 Pass Pass 
FairyMax 48L.AB xb arm5 Fail Pass 
Ges 1.34.JA xb arm5 Fail Pass 
Gfc 0.8.r5.JA xb arm5 Fail Fail 
Gk 0.90.JA xb arm5 Fail Pass 
GreKo 12.5.JA xb arm5 Pass Fail 
Gullydeckel 2.15pl10.JA xb arm5 Fail Fail 
Jester 0.84.JA xb arm5 Pass Fail 
Jfresh 0.1a.JA xb arm5 Fail Fail 
Kitteneitor 060513.JA xb arm5 Pass Pass 
Kiwi 0.6d.JA xb arm5 Fail Fail 
Kmtchess 1.21.JA xb arm5 Pass Fail 
KnightCap 3.7F.JA xb arm5 Pass Pass 
Leonidas r83.JA xb arm5 Pass Fail 
Minace 1.0.JA xb arm5 Fail Fail 
Mizar 3.0.JA xb arm5 Fail Fail 
Morphy Chess 1.05.JA xb arm5 Fail Fail 
MsKcp 1.45.JA xb arm5 Fail Fail 
Myrddin 0.86.JA xb arm5 Pass Fail 
Napoleon 1.50.JA xb arm5 Fail Fail 
Natwarlal 0.14.JA xb arm5 Pass Fail 
Olithink 5.3.2.JA xb arm5 Pass Fail 
Phalanx XXIV.JA xb arm5 Pass Fail 
Purple Haze 2.10.JA xb arm5 Fail Fail 
RattateChess Nosferatu.JA xb arm5 Pass Fail 
Resp 0.19.JA xb arm5 Pass Pass 
Romichess p3k.JA xb arm5 Fail Pass 
Scidlet 2.6.1b2.JA xb arm5 Pass Fail 
Scorpio 2.7.7.JA xb arm5 Pass Fail 
Sjaak II beta5.JA xb arm5 Pass Pass 
Sjeng 1.12.JA xb arm5 Pass Fail 
Sloppy 0.23 xb arm5 Pass Fail 
Surprise 4.3.b13.JA xb arm5 Fail Pass 
Toledo Nanochess 2010.JA xb arm5 Fail Pass 
TSCP 1.8.1.AB xb arm5 Pass Fail 
Tucano 1.04.AB xb arm5 Pass Fail 
Typhoon 1.0.r358.JA xb arm5 Pass Fail 
µMax 4.8W xb arm5 Fail Pass 
Wing 1.3.JA xb arm5 Fail Fail 
Witz Alpha21.JA xb arm5 Pass Pass 
Woodpecker 2.11.JA xb arm5 Pass Fail 
Xadreco 5.7.JA xb arm5 Fail Pass 
Xpdnt 091007.JA xb arm5 Fail Fail 
Zzzzzz 3.5.1 xb arm5 Pass Fail 

[Ferdy]

I checked my time usage categories: Both Danasah versions are set to FIXED TIME only.
With Fischer clock Danasah 5.07 is marked as using extra time and Danasah Z as instant move.
Therefore, everything was already set correctly for these engines and no worries


All engines which don't comply to Fischer clock are tested and detected before introduction to Rapidroid. If they play fine with fixed time they are marked category 3: FT

Again, no worries.

With Fischer clock Danasah 5.07 is marked as using extra time

You mean this is time forfeit?

I have a winboard engine compiled for android, I will try it in latest CFA to check if it has problem in Fischer TC.

[Gurcan Uckardes]

Yes. CfA may have issues with Fischer in xb. Aart knows already about this but i remember he could not solve. XB engines manage their clocks on their own and there's no interaction with the GUI.

I always use 100ms additional time per move to reduce time forfeits but even Crafty had some true forfeit losses, even that happened very rarely.

I agree with you about different behaviours between human-engine and engine-engine games. I've seen this during my manual games between Colossus C64 and xb engines: Xadreco plays not bad in Rapidroid with Fischer, average 11 secs with 900+2 but it played 5 sec fixed followed by instant moves in manual games. These games were cancelled.

In addition, i see some engines pass the test but they still may fail in the long run because termination issues impact tournaments. Some engines terminate without reason without a chance to adjudication, implying useless replays. I exclude them. Jazz is one example.

There's a post in my blog and a summary file detailing all my findings. That summary seems outdated by now and i must update it asap.

Hi Gurcan,

If you play as a human against DanaSahZ 0.4 engine on Chess for Android and you uses Fischer time control, you can see that the engine does not move instantaneously. However when the engine is playing with another engine , then if it does, which it makes no sense.

It is for this reason that I believe that Chess for Android should have some small error in the xboard protocol and when the confrontation is against another engine. I should point out this problem to Art , but I find difficult to express me in English and in recent months I had no time to chess computer. If other engines also move instantly and they have the xboard protocol , maybe the error is not the engine , maybe it is the GUI.

When you do a test to see if the engine pass Fischer control time, what time control you uses?

I have to admit that I have not tried Fischer time control, because it was not implemented just the latest versión, but here a list with engines (xboard) which fail to Fischer control time on Chess for Android. It is almost 75 % of engines with this problem, for me this is a bit strange.

Code: Select all

Engine Pro Arch Fixed Fischer
Amy 4.8 xb arm5 Pass Pass
Beowulf 2.04a.JA xb arm5 Fail Fail
Betsabe 1.47.JA xb arm5 Pass Fail
Bruja 1.9.1.JA xb arm5 Pass* Fail
Buzz 11-27-08.JA xb arm5 Fail Fail
Buzz OS 1.0.JA xb arm5 Fail Fail 
Cassandre 0.26.JA xb arm5 Fail Fail
Chenard 2014.07.11.JA xb arm5 Pass Pass
Chess One 2.01.JA xb arm5 Fail Fail 
Chess Rikus 1.466.JA xb arm5 Fail Fail 
Cilian 4.14.JA xb arm5 Pass Pass 
Colchess 8.0.JA xb arm5 Pass* Fail 
Crafty 24.1.JA xb arm5 Pass Pass 
Dabbaba 7.49.JA xb arm5 Fail Fail 
Danasah 5.07.JA xb arm5 Pass Fail 
DanasahZ 0.4 xb arm5 Pass Fail 
DanChess 1.04.JA xb arm5 Pass Fail 
Darky 0.5d xb arm5 Fail Fail 
Dirty 030411.JA xb arm5 Pass Fail 
Exchess 7.51.beta.JA xb arm5 Fail Fail 
Faile 1.4.4.JA xb arm5 Pass Pass 
FairyMax 48L.AB xb arm5 Fail Pass 
Ges 1.34.JA xb arm5 Fail Pass 
Gfc 0.8.r5.JA xb arm5 Fail Fail 
Gk 0.90.JA xb arm5 Fail Pass 
GreKo 12.5.JA xb arm5 Pass Fail 
Gullydeckel 2.15pl10.JA xb arm5 Fail Fail 
Jester 0.84.JA xb arm5 Pass Fail 
Jfresh 0.1a.JA xb arm5 Fail Fail 
Kitteneitor 060513.JA xb arm5 Pass Pass 
Kiwi 0.6d.JA xb arm5 Fail Fail 
Kmtchess 1.21.JA xb arm5 Pass Fail 
KnightCap 3.7F.JA xb arm5 Pass Pass 
Leonidas r83.JA xb arm5 Pass Fail 
Minace 1.0.JA xb arm5 Fail Fail 
Mizar 3.0.JA xb arm5 Fail Fail 
Morphy Chess 1.05.JA xb arm5 Fail Fail 
MsKcp 1.45.JA xb arm5 Fail Fail 
Myrddin 0.86.JA xb arm5 Pass Fail 
Napoleon 1.50.JA xb arm5 Fail Fail 
Natwarlal 0.14.JA xb arm5 Pass Fail 
Olithink 5.3.2.JA xb arm5 Pass Fail 
Phalanx XXIV.JA xb arm5 Pass Fail 
Purple Haze 2.10.JA xb arm5 Fail Fail 
RattateChess Nosferatu.JA xb arm5 Pass Fail 
Resp 0.19.JA xb arm5 Pass Pass 
Romichess p3k.JA xb arm5 Fail Pass 
Scidlet 2.6.1b2.JA xb arm5 Pass Fail 
Scorpio 2.7.7.JA xb arm5 Pass Fail 
Sjaak II beta5.JA xb arm5 Pass Pass 
Sjeng 1.12.JA xb arm5 Pass Fail 
Sloppy 0.23 xb arm5 Pass Fail 
Surprise 4.3.b13.JA xb arm5 Fail Pass 
Toledo Nanochess 2010.JA xb arm5 Fail Pass 
TSCP 1.8.1.AB xb arm5 Pass Fail 
Tucano 1.04.AB xb arm5 Pass Fail 
Typhoon 1.0.r358.JA xb arm5 Pass Fail 
µMax 4.8W xb arm5 Fail Pass 
Wing 1.3.JA xb arm5 Fail Fail 
Witz Alpha21.JA xb arm5 Pass Pass 
Woodpecker 2.11.JA xb arm5 Pass Fail 
Xadreco 5.7.JA xb arm5 Fail Pass 
Xpdnt 091007.JA xb arm5 Fail Fail 
Zzzzzz 3.5.1 xb arm5 Pass Fail 

[Ferdy]

I did the test with Deuterium WB android, and indeed there is a problem when num moves >= 50 or close to this, using Fischer TC 1m+1s inc. There are no problem running same engines in a windows PC with same TC.

I create a script to look at the time log of the android games and make a comparison with uci engine, the WB engine is taking more time per move. All games are time forfeits.

Code: Select all

player                 : Deuterium v14.1.1.2 uci 32bit android
num games              : 10
num moves with book    : 38
num moves with comment : 504
num moves with time    : 503
sum moves              : 542
sum depth              : 5531
sum time               : 775.968s
ave depth              : 11.0
ave time               : 1.543s

player                 : Deuterium-1_1-wb-x32-android
num games              : 10
num moves with book    : 38
num moves with comment : 508
num moves with time    : 508
sum moves              : 546
sum depth              : 5310
sum time               : 1120.627s
ave depth              : 10.5
ave time               : 2.206s

I run another test WB engine vs CFA engine TC 1m+1s inc, all games were won by WB engine all games are less than 50 moves, CFA engine is mated.