Datasets:

Bluefir
/

escampe-dataset

score float64 -1 1	white_paladins int64 31 60.7B	black_paladins int64 2.39k 66.6B	white_unicorn int64 0 35	required_band int64 0 3	white_to_move bool 2 classes	black_unicorn int64 0 35
-0.00152	84,112	35,605,446,656	0	1	false	32
-0.00005	100,512	1,514,143,744	0	2	false	32
0.00005	100,512	1,514,143,744	0	0	false	20
-0.00134	100,512	1,447,034,880	6	2	true	2
0.00039	68,000	1,447,034,880	6	3	false	2
0.0017	34,890	1,384,644,608	2	0	false	32
-0.00006	34,890	1,653,080,064	2	2	true	32
0.00096	133,194	1,653,080,064	2	3	false	32
-0.0027	133,194	1,619,542,016	2	1	true	32
1	198,722	1,652,572,160	2	1	false	32
1	722,946	1,653,604,352	2	1	false	32
1	11,076,096	18,288,218,112	2	0	true	32
-1	76,087,808	18,288,218,112	2	1	false	32
-1	76,087,808	18,288,218,112	0	1	false	32
-0.00173	8,397,408	17,734,057,984	2	3	false	32
-0.00071	8,397,920	621,297,664	2	3	false	32
-0.00049	8,528,960	621,297,664	2	0	true	32
-0.00241	4,334,656	621,297,664	2	3	false	32
-0.00117	6,430,784	621,297,664	2	2	false	32
0.00146	6,422,593	621,297,664	2	1	false	25
0.00057	6,422,593	688,406,528	2	3	true	25
0.0025	2,228,305	688,406,528	2	1	false	25
-1	10,485,841	1,745,895,424	2	3	true	25
1	2,099,841	28,118,614,016	14	0	false	24
1	2,099,841	28,118,614,016	14	0	false	25
1	2,097,826	22,716,350,464	23	2	true	20
-1	49,314	22,716,350,464	23	2	false	20
-0.00076	66,130	36,138,123,264	15	2	false	32
-0.00071	525,592	34,997,276,672	6	2	true	32
0.00047	525,592	35,014,049,792	6	1	true	32
-0.00278	8,716,312	35,538,337,792	6	1	false	32
-0.00323	8,732,704	35,472,277,504	6	2	false	32
1	8,732,704	35,472,277,504	6	3	true	34
1	8,732,704	35,472,277,504	7	0	true	34
-1	537,215,008	35,472,277,504	7	2	false	34
1	537,216,000	35,472,277,504	7	3	false	34
-0.00297	557,974,528	44,091,572,224	1	1	false	31
1	155,321,344	48,319,430,656	1	3	false	31
-1	155,321,344	48,318,382,208	1	1	true	31
-1	138,806,272	48,318,382,208	1	1	true	19
0.00092	4,194,977	22,145,941,504	4	1	true	31
-0.00388	4,194,721	22,145,941,504	4	3	false	31
1	4,194,721	22,011,724,288	4	1	true	31
1	4,202,848	5,100,290,560	4	0	true	31
-0.00077	2,105,696	5,100,290,560	4	2	false	31
-0.00126	2,105,696	838,877,696	4	3	true	31
-0.00146	18,882,848	838,878,208	4	3	true	31
-1	18,874,660	838,878,208	4	2	false	31
1	18,939,940	17,482,384,384	4	1	true	31
1	1,075,904,548	17,482,384,384	4	3	false	31
1	1,080,033,316	17,448,846,336	4	3	false	31
1	9,665,773,604	4,572,315,904	3	2	false	31
-1	9,665,773,728	278,135,040	9	1	false	31
1	9,665,773,728	269,877,504	9	3	true	31
0.00029	51,265	52,781,121,536	2	3	false	28
0.0024	43,074	52,613,365,888	2	2	false	28
-0.00448	43,074	18,262,016,128	2	1	true	28
0.00238	8,778	4,305,469,568	2	0	false	28
0.00228	8,778	4,305,485,952	14	1	false	28
1	10,818	4,305,485,952	14	2	false	22
-1	10,562	4,571,824,256	20	3	false	4
-1	10,568	4,571,824,256	20	3	false	4
1	10,568	4,571,824,256	20	0	true	4
0.00033	1,384	9,143,664,640	2	0	true	32
-0.00021	2,098,528	9,143,664,640	2	2	false	32
1	67,125,424	17,217,781,760	2	0	false	32
0.00078	559,140	8,942,387,200	21	2	false	30
-1	559,140	8,942,387,200	21	0	true	30
-1	296,996	8,942,387,200	21	0	true	30
1	2,101,267	1,413,480,704	2	0	false	27
0.00256	8,405,648	34,932,264,960	2	2	false	27
-0.00221	9,437,840	35,198,603,264	2	3	false	27
1	9,438,344	37,312,544,768	2	0	true	26
0.00077	276,841,088	37,179,359,232	1	2	false	32
-0.00086	276,841,088	37,179,359,232	1	3	true	25
0.00025	276,841,088	37,179,359,232	1	0	false	25
0.00044	276,841,216	37,045,207,040	1	3	false	25
-1	276,857,096	34,915,549,184	13	3	false	19
1	276,857,096	34,915,549,184	13	0	true	19
1	276,858,888	34,913,468,416	13	1	true	19
-1	268,601,352	34,913,468,416	13	3	false	19
1	268,601,352	34,913,468,416	13	0	true	19
1	2,164,808	17,884,774,400	14	3	false	20
1	2,164,808	17,817,665,536	14	1	true	20
1	6,293,576	17,817,665,536	14	3	false	20
1	6,293,576	9,227,730,944	14	1	true	20
-0.00091	71,169	17,317,494,784	4	3	true	32
-0.00173	8,458,753	17,317,494,784	4	1	false	32
0.00008	8,458,754	17,317,494,784	4	2	false	32
-0.00069	8,458,754	18,390,974,464	4	0	false	32
-0.00099	8,458,754	18,390,450,176	4	1	true	32
-0.00127	4,264,450	19,464,192,000	4	2	true	32
-0.00114	4,264,456	19,464,192,000	4	3	false	32
-1	138,448,904	614,989,824	10	3	false	9
1	138,448,904	614,989,824	10	0	true	9
-0.00194	8,389,328	36,138,123,264	14	0	true	32
1	8,397,392	36,138,123,264	14	3	false	32
0.00218	8,397,384	44,828,721,152	14	1	true	32
-0.00127	8,405,576	40,533,753,856	10	3	false	27
0.00502	8,405,576	40,533,753,856	10	1	true	19

End of preview. Expand in Data Studio

Escampe Game State Evaluation Dataset

Dataset Summary

This dataset contains millions of evaluated game states for the board game Escampe. It is designed to train Deep Neural Networks (such as the BandDPER architecture) to accurately predict the heuristic value of any given board state.

Escampe is a deterministic, perfect-information 6x6 board game played with Unicorns and Paladins. The dataset was procedurally generated using an Alpha-Beta search engine in Java that plays games of Escampe against itself, taking random opening positions and semi-random moves to guarantee high diversity in board states. Every position in this dataset has been deeply evaluated by the minimax engine using various heuristic configurations (Default, SPSA, Bayes, and Simplex) to assign a continuous score.

Dataset Structure

Data Fields

Each JSON entry corresponds to a specific board state from a game in progress. The fields include:

white_paladins (int): A 64-bit integer bitboard mask representing the positions of White's Paladins.
black_paladins (int): A 64-bit integer bitboard mask representing the positions of Black's Paladins.
white_unicorn (int): The 0-35 index representing the square of White's Unicorn.
black_unicorn (int): The 0-35 index representing the square of Black's Unicorn.
required_band (int): An integer (1, 2, or 3) representing the required liseret/band the current player must start their move on. A 0 means no band constraint (free move).
white_to_move (bool): true if it is White's turn to move, false otherwise.
score (float): The target evaluation score normalized to the range [-1.0, 1.0]. The score is relative to the current player to move. A score of +1.0 indicates a guaranteed win (mate), while -1.0 indicates a guaranteed loss.

Data Instances

{
  "white_paladins": 8913048,
  "black_paladins": 40298872832,
  "white_unicorn": 14,
  "black_unicorn": 34,
  "required_band": 2,
  "white_to_move": true,
  "score": -1.0
}

Dataset Creation

The dataset was generated using a high-performance, multi-threaded Java minimax engine utilizing Alpha-Beta pruning.

Diversity: Random openings are selected from a predefined pool of Escampe openings. Games proceed using a mix of best-moves and random moves (70% exploration) to ensure the network is exposed to sub-optimal, chaotic, and end-game positions.
Normalization: The raw heuristic scores (which scale up to a theoretical mate score of 100,000) are clamped and normalized to [-1.0, 1.0] before being written to the JSON file.

Considerations for Using the Data

The score field is strictly evaluated from the perspective of the player to move. If you are training a model to output an absolute score (e.g., White's advantage), you must invert the score when white_to_move is false.
Escampe relies heavily on spatial constraints (the required band) and blocking logic. It is highly recommended to translate the bitboards into 6x6 spatial tensors to expose the orthogonal paths and band geometries to your neural network.

Citation

If you use this dataset, please consider linking back to this repository and crediting the Escampe AI project.

Downloads last month: 37