The GVG-AI Competition

Training Set 2-Player Games 4 (Test Set CIG 2016)

Rank	Username	Country	Points	Avg. Glicko Score
1	adrienctx	Belgium	162	954.69
2	MaastCTS2	Netherlands	143	912.79
3	NextLevel	Poland	125	839.07
4	Number27	Germany	123	823.82
5	sampleOLMCTS Sample	United Kingdom	120	776.32
6	ToVo2	Slovenia	109	800.04
7	random Sample	United Kingdom	71	438.17
8	CatLinux	United Kingdom	71	720.08
9	NUDTQi	China	70	722.83
10	webpigeon	United Kingdom	16	185.99

Egg Hunt:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Egg Hunt	adrienctx	Belgium	25	1252.43	65% (885 / 0)	54.69	762
Egg Hunt	Number27	Germany	18	1234.03	65% (973 / 5)	24.76	845
Egg Hunt	MaastCTS2	Netherlands	15	1225.56	49% (396 / 0)	42.83	263
Egg Hunt	ToVo2	Slovenia	12	1164.86	31% (96 / 0)	33.71	322
Egg Hunt	sampleOLMCTS	United Kingdom	10	1155.37	40% (867 / 1)	30.43	916
Egg Hunt	NUDTQi	China	8	1114.02	40% (10 / 0)	52.5	777
Egg Hunt	CatLinux	United Kingdom	6	1105.26	52% (869 / 12)	24.46	932
Egg Hunt	NextLevel	Poland	4	1046.4	33% (872 / 0)	28.11	842
Egg Hunt	random	United Kingdom	2	809.38	15% (843 / 0)	19.19	947
Egg Hunt	webpigeon	United Kingdom	1	82.25	2% (1052 / 705)	-667.1	250

Fatty:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Fatty	MaastCTS2	Netherlands	25	1181.72	61% (396 / 0)	11.4	562
Fatty	adrienctx	Belgium	18	1168.46	61% (880 / 0)	11.94	605
Fatty	ToVo2	Slovenia	15	1150.18	49% (96 / 0)	9.33	559
Fatty	Number27	Germany	12	1134.38	66% (961 / 3)	10.83	474
Fatty	sampleOLMCTS	United Kingdom	10	1094.7	54% (865 / 0)	9.72	778
Fatty	CatLinux	United Kingdom	8	948.89	54% (867 / 0)	9.8	774
Fatty	NUDTQi	China	6	930.21	30% (10 / 0)	7.8	905
Fatty	random	United Kingdom	4	786.22	26% (846 / 0)	7.7	693
Fatty	webpigeon	United Kingdom	2	744.07	35% (1054 / 119)	-105.49	699
Fatty	NextLevel	Poland	1	721.93	37% (866 / 1)	5.93	842

I Saw Santa:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
I Saw Santa	adrienctx	Belgium	25	1359.16	69% (884 / 0)	0.69	1236
I Saw Santa	NextLevel	Poland	18	1129.26	53% (871 / 1)	-0.62	1253
I Saw Santa	sampleOLMCTS	United Kingdom	15	1064.37	42% (868 / 0)	0.42	1393
I Saw Santa	ToVo2	Slovenia	12	1057.42	39% (96 / 0)	0.39	1100
I Saw Santa	Number27	Germany	10	1038.17	46% (974 / 0)	0.46	1366
I Saw Santa	random	United Kingdom	8	1035.19	41% (841 / 0)	0.41	1426
I Saw Santa	CatLinux	United Kingdom	6	1018.37	43% (867 / 1)	-0.72	1389
I Saw Santa	MaastCTS2	Netherlands	4	995.68	51% (396 / 1)	-2.02	1204
I Saw Santa	NUDTQi	China	2	962.42	30% (10 / 1)	-99.7	1238
I Saw Santa	webpigeon	United Kingdom	1	847.6	37% (1049 / 98)	-93.05	1288

Mimic:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Mimic	NextLevel	Poland	25	978.81	46% (318 / 0)	0.46	903
Mimic	sampleOLMCTS	United Kingdom	18	964.44	53% (713 / 1)	-0.87	976
Mimic	random	United Kingdom	15	957.67	55% (715 / 0)	0.55	983
Mimic	CatLinux	United Kingdom	12	956.04	55% (704 / 0)	0.55	974
Mimic	ToVo2	Slovenia	10	943.27	40% (96 / 0)	0.4	782
Mimic	Number27	Germany	8	917.56	56% (816 / 0)	0.56	984
Mimic	adrienctx	Belgium	6	908.57	36% (715 / 0)	0.36	977
Mimic	MaastCTS2	Netherlands	4	884.22	35% (326 / 0)	0.35	943
Mimic	webpigeon	United Kingdom	2	845.82	48% (1002 / 103)	-102.31	899

Reflection:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Reflection	MaastCTS2	Netherlands	25	779.58	16% (396 / 0)	0.16	1045
Reflection	NextLevel	Poland	18	771.99	21% (887 / 0)	0.21	1089
Reflection	CatLinux	United Kingdom	15	737.1	24% (866 / 24)	-27.47	1118
Reflection	ToVo2	Slovenia	12	709.84	9% (96 / 0)	0.09	959
Reflection	sampleOLMCTS	United Kingdom	10	695.01	20% (867 / 0)	0.2	1144
Reflection	Number27	Germany	8	688.86	18% (988 / 17)	-17.03	1060
Reflection	adrienctx	Belgium	6	676.15	14% (892 / 3)	-3.22	1066
Reflection	NUDTQi	China	4	665.2	0% (10 / 0)	0	1115
Reflection	random	United Kingdom	2	12.52	3% (843 / 0)	0.03	267
Reflection	webpigeon	United Kingdom	1	0	3% (1048 / 854)	-814.86	364

Train Ride:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Train Ride	adrienctx	Belgium	25	1485.33	69% (877 / 0)	92.75	375
Train Ride	Number27	Germany	18	1385.58	61% (970 / 0)	39.54	259
Train Ride	NUDTQi	China	15	1374.1	50% (10 / 0)	55.4	314
Train Ride	MaastCTS2	Netherlands	12	1353.63	64% (396 / 0)	40.76	306
Train Ride	ToVo2	Slovenia	10	1252.1	35% (96 / 0)	33.29	227
Train Ride	sampleOLMCTS	United Kingdom	8	1189.49	61% (857 / 0)	37.8	242
Train Ride	NextLevel	Poland	6	1178.27	53% (821 / 1)	41.66	274
Train Ride	CatLinux	United Kingdom	4	1140.89	57% (865 / 1)	53.47	227
Train Ride	webpigeon	United Kingdom	2	0	9% (1054 / 406)	-368.16	204
Train Ride	random	United Kingdom	1	0	7% (844 / 0)	15.48	119

Treasure Map:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Treasure Map	random	United Kingdom	25	0	0% (801 / 0)	0	1342
Treasure Map	NUDTQi	China	18	0	0% (4 / 0)	0	1347
Treasure Map	adrienctx	Belgium	15	0	0% (780 / 17)	-21.79	1369
Treasure Map	NextLevel	Poland	12	0	0% (793 / 63)	-79.45	1115
Treasure Map	ToVo2	Slovenia	10	0	0% (96 / 14)	-145.83	1056
Treasure Map	sampleOLMCTS	United Kingdom	8	0	0% (793 / 136)	-171.5	1279
Treasure Map	MaastCTS2	Netherlands	6	0	0% (338 / 61)	-180.47	917
Treasure Map	Number27	Germany	4	0	0% (882 / 171)	-193.88	1023
Treasure Map	CatLinux	United Kingdom	2	0	0% (791 / 207)	-261.69	1176
Treasure Map	webpigeon	United Kingdom	1	0	0% (1016 / 868)	-854.33	395

Upgrade-X:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Upgrade-X	sampleOLMCTS	United Kingdom	25	576.16	56% (795 / 2)	132.87	943
Upgrade-X	NextLevel	Poland	18	571.99	43% (782 / 0)	130.71	914
Upgrade-X	ToVo2	Slovenia	15	381.11	29% (96 / 0)	105.75	682
Upgrade-X	MaastCTS2	Netherlands	12	364.16	33% (328 / 2)	95.66	714
Upgrade-X	NUDTQi	China	10	305.6	30% (10 / 1)	36.9	1056
Upgrade-X	Number27	Germany	8	254.44	23% (883 / 9)	114.3	905
Upgrade-X	adrienctx	Belgium	6	225.09	20% (795 / 0)	122.21	988
Upgrade-X	random	United Kingdom	4	59.11	7% (786 / 0)	105.94	355
Upgrade-X	CatLinux	United Kingdom	2	0	25% (797 / 229)	-188.44	749
Upgrade-X	webpigeon	United Kingdom	1	0	0% (996 / 996)	-1000	1

War Zone:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
War Zone	Number27	Germany	25	1585.2	69% (969 / 5)	87.24	365
War Zone	adrienctx	Belgium	18	1470.86	54% (877 / 1)	45.05	677
War Zone	MaastCTS2	Netherlands	15	1393.57	40% (396 / 1)	51.08	520
War Zone	ToVo2	Slovenia	12	1341.62	31% (95 / 0)	21.64	592
War Zone	CatLinux	United Kingdom	10	1294.28	48% (868 / 2)	31.63	838
War Zone	NextLevel	Poland	8	1292.73	41% (889 / 0)	32.54	748
War Zone	sampleOLMCTS	United Kingdom	6	1211.78	36% (867 / 0)	30.7	794
War Zone	NUDTQi	China	4	1153.94	30% (10 / 0)	21.6	920
War Zone	random	United Kingdom	2	721.62	9% (845 / 0)	-7.46	532
War Zone	webpigeon	United Kingdom	1	0	2% (1050 / 923)	-879.25	177

Wheel Me:

Game	Username	Country	Points	Glicko Score	Winner % (N / D)	Game Score Avg.	Timesteps Avg.
Wheel Me	MaastCTS2	Netherlands	25	8.43	17% (396 / 1)	-2.35	1300
Wheel Me	adrienctx	Belgium	18	0	20% (882 / 0)	0.2	1537
Wheel Me	NextLevel	Poland	15	0	18% (872 / 0)	0.18	1446
Wheel Me	Number27	Germany	12	0	14% (966 / 1)	-0.89	1578
Wheel Me	sampleOLMCTS	United Kingdom	10	0	13% (868 / 0)	0.13	1649
Wheel Me	random	United Kingdom	8	0	12% (845 / 0)	0.12	1657
Wheel Me	CatLinux	United Kingdom	6	0	12% (869 / 3)	-3.33	1664
Wheel Me	webpigeon	United Kingdom	4	0	12% (1053 / 111)	-105.3	1296
Wheel Me	NUDTQi	China	2	0	10% (10 / 0)	0.1	1837
Wheel Me	ToVo2	Slovenia	1	0	5% (96 / 0)	0.05	1470

Rank	Username	Country	Description	G-1	G-2	G-3	G-4	G-5	G-6	G-7	G-8	G-9	G-10	Total
1	adrienctx	Belgium	Controller description: adrienctx simple MCTS Description	25	18	25	6	6	25	15	6	18	18	162
2	MaastCTS2	Netherlands	Controller description: MaastCTS2 Two-player variant of the MaastCTS2 agent. A number of enhancements from the single-player variant have been disabled. Description	15	25	4	4	25	12	6	12	15	25	143
3	NextLevel	Poland	Controller description: NextLevel OLMCTS equipped with a state heuristic based on features of objects, evaluated during preliminary learning. Description	4	1	18	25	18	6	12	18	8	15	125
4	Number27	Germany	Controller description: Number27 Using a GA for local movement and a value map to direct the player across the level. The value map is created by evaluating each object type which results in a certain influence across the map. Notable events, the time spent in one area or non deterministic movements influence the players behaviour or how it chooses its optimal action. Description	18	12	10	8	8	18	4	8	25	12	123
5	sampleOLMCTS Sample	United Kingdom	Controller description: sampleOLMCTS Sample OLMCTS Description	10	10	15	18	10	8	8	25	6	10	120
6	ToVo2	Slovenia	Controller description: ToVo2 A Sarsa-UCT algorithm, using TD backups and space-local Q-value normalization. Builds a tree, does not compute transpositions. Most primitive opponent model: random player, assumed as part of the environment � keeping value estimates only for own agent. No expert knowledge, no handcrafted strategies, no heuristics. Description	12	15	12	10	12	10	10	15	12	1	109
7	random Sample	United Kingdom	Controller description: random Sample random Description	2	4	8	15	2	1	25	4	2	8	71
8	CatLinux	United Kingdom	Controller description: CatLinux RHGA with circular buffer Description	6	8	6	12	15	4	2	2	10	6	71
9	NUDTQi	China	Controller description: NUDTQi The method is based on 1+1 RHEA, there are following changes: save the history optimal information to shift on next tick; the search depth is initially 1, after finishing the least iterations then increase gradually; the mutation probability is changeable according to the max search depth Description	8	6	2	1	4	15	18	10	4	2	70
10	webpigeon	United Kingdom	Controller description: webpigeon OLMCTS with simple agent prediction Description	1	2	1	2	1	2	1	1	1	4	16

* Note: These rankings only include controllers executed after 01/05/2016

* For details on how rankings are calculated, check the GVGAI Wiki.