starting
sequencelengths
3
8
target
int64
1
999
closest
int64
1
999
expression
stringlengths
1
54
delta
int64
0
9
score
int64
1
10
size
int64
3
8
[ 29, 34, 37 ]
26
26
((29 + 34) - 37)
0
10
3
[ 71, 82, 87, 89 ]
92
93
(((89 - 87) * 82) - 71)
1
9
4
[ 51, 63, 92, 14, 22, 100, 93 ]
374
374
((((100 - 63) * 14) - 93) - 51)
0
10
7
[ 11, 11, 58, 98, 97, 25, 43 ]
970
970
(((25 + 43) - 58) * 97)
0
10
7
[ 7, 91, 59, 16 ]
258
262
(((16 * 7) + 91) + 59)
4
6
4
[ 22, 21, 90, 8, 9, 100, 72, 42 ]
534
534
((21 * 22) + 72)
0
10
8
[ 78, 61, 4, 49, 36, 38, 33 ]
796
796
((((4 + 49) - 33) * 38) + 36)
0
10
7
[ 58, 50, 77 ]
34
31
(50 - (77 - 58))
3
7
3
[ 6, 57, 26, 96 ]
41
41
(57 - (96 / 6))
0
10
4
[ 79, 39, 41, 41, 49 ]
240
233
(((49 - 41) * 39) - 79)
7
3
5
[ 86, 13, 22, 73, 49 ]
935
936
((((49 + 22) * 13) - 73) + 86)
1
9
5
[ 44, 84, 24, 53, 59, 53 ]
656
655
((((53 - (59 - 24)) * 44) - 84) - 53)
1
9
6
[ 25, 39, 74, 2, 99 ]
938
936
((74 - (2 * 25)) * 39)
2
8
5
[ 49, 72, 38 ]
27
23
(72 - 49)
4
6
3
[ 14, 36, 21, 4 ]
271
274
(((21 - 4) * 14) + 36)
3
7
4
[ 82, 38, 26 ]
898
906
((38 * 26) - 82)
8
2
3
[ 93, 45, 2, 26, 54 ]
208
208
((((26 + 54) * 2) - 45) + 93)
0
10
5
[ 28, 44, 85, 61, 96, 2, 23 ]
510
510
((23 - (61 - 44)) * 85)
0
10
7
[ 64, 54, 63, 70, 16 ]
695
694
(((63 - 54) * 70) + 64)
1
9
5
[ 80, 25, 35 ]
963
955
((25 * 35) + 80)
8
2
3
[ 54, 25, 41 ]
116
120
((41 + 25) + 54)
4
6
3
[ 81, 26, 64, 2, 33, 19, 27 ]
911
911
((((81 - 33) * 19) + 26) - 27)
0
10
7
[ 40, 63, 66, 11, 1, 38, 18, 7 ]
61
61
(63 - (40 - 38))
0
10
8
[ 50, 29, 15, 89, 15 ]
736
736
(((50 * 15) - 29) + 15)
0
10
5
[ 92, 23, 47, 64, 19, 20, 100 ]
725
725
(((20 + 23) * 19) - 92)
0
10
7
[ 43, 41, 94 ]
12
10
((94 - 43) - 41)
2
8
3
[ 30, 7, 35, 81, 14, 93, 28, 27 ]
76
76
(81 - (35 - 30))
0
10
8
[ 67, 8, 76, 35, 58, 90, 86, 8 ]
814
814
(((90 * 8) + 8) + 86)
0
10
8
[ 94, 56, 92, 32, 50, 27, 3 ]
346
346
((((92 * 3) + 94) - 56) + 32)
0
10
7
[ 59, 91, 66, 84 ]
457
462
((91 - 84) * 66)
5
5
4
[ 88, 48, 38, 38, 66 ]
295
292
(((48 - 38) * 38) - 88)
3
7
5
[ 86, 63, 25, 47 ]
83
85
((63 + 47) - 25)
2
8
4
[ 79, 40, 30, 61, 51, 48 ]
317
317
((((51 - 40) * 30) - 61) + 48)
0
10
6
[ 32, 76, 71, 17 ]
711
704
(((76 + 17) - 71) * 32)
7
3
4
[ 65, 75, 6, 62 ]
614
614
(((75 - 65) * 62) - 6)
0
10
4
[ 65, 82, 90, 37, 13 ]
798
800
(((13 * 65) + 37) - 82)
2
8
5
[ 3, 62, 25 ]
202
211
((3 * 62) + 25)
9
1
3
[ 27, 14, 74, 9 ]
92
92
((74 - 9) + 27)
0
10
4
[ 17, 86, 58, 78, 76, 68 ]
903
904
((((78 - 68) * 76) + 58) + 86)
1
9
6
[ 36, 45, 52, 35 ]
660
657
(((52 - 35) * 36) + 45)
3
7
4
[ 92, 7, 8, 33, 21 ]
804
804
((((92 + 7) * 8) - 21) + 33)
0
10
5
[ 24, 37, 16 ]
576
568
((16 * 37) - 24)
8
2
3
[ 20, 10, 85, 55, 84, 33, 32 ]
774
774
((((33 * 20) + 84) - 55) + 85)
0
10
7
[ 24, 70, 13 ]
762
770
((24 - 13) * 70)
8
2
3
[ 65, 4, 56, 55 ]
265
264
(((65 - 55) + 56) * 4)
1
9
4
[ 79, 72, 53, 52, 93, 58, 43 ]
205
205
(((58 - 52) * 43) - 53)
0
10
7
[ 20, 5, 54, 67, 44, 98 ]
477
477
(((98 * 5) - 67) + 54)
0
10
6
[ 50, 64, 29, 58 ]
461
464
(((64 - 50) * 29) + 58)
3
7
4
[ 73, 42, 2, 11, 23, 64, 74, 12 ]
369
369
(((11 * 23) + 42) + 74)
0
10
8
[ 100, 33, 85, 16 ]
546
543
(((16 * 33) + 100) - 85)
3
7
4
[ 40, 36, 75, 67, 72 ]
641
640
((((75 - 67) * 72) / 36) * 40)
1
9
5
[ 70, 94, 96, 30 ]
9
10
(70 - ((96 - 94) * 30))
1
9
4
[ 36, 36, 48, 79, 35, 8 ]
408
408
(((((79 - 35) + 36) / 8) * 36) + 48)
0
10
6
[ 47, 5, 69, 15, 8, 15, 20 ]
244
244
(((15 + 20) * 5) + 69)
0
10
7
[ 14, 42, 52 ]
76
80
((52 - 14) + 42)
4
6
3
[ 79, 40, 25, 67, 81, 64, 59, 40 ]
870
870
((((40 - 25) * 59) - 79) + 64)
0
10
8
[ 39, 58, 21 ]
11
18
(39 - 21)
7
3
3
[ 11, 28, 43, 98 ]
916
913
(((28 + 98) - 43) * 11)
3
7
4
[ 80, 87, 44, 94, 45, 65, 94 ]
730
730
(((((94 - 80) * 44) + 65) + 94) - 45)
0
10
7
[ 7, 2, 31, 66, 3, 74, 77 ]
276
276
(((77 - 31) * 2) * 3)
0
10
7
[ 32, 67, 15, 82, 57, 39, 52, 15 ]
802
802
(((32 - 15) * 52) - 82)
0
10
8
[ 96, 83, 79, 88 ]
93
92
((83 + 88) - 79)
1
9
4
[ 70, 60, 92 ]
222
222
((60 + 92) + 70)
0
10
3
[ 74, 16, 38, 22, 73, 84, 48 ]
95
95
(22 + 73)
0
10
7
[ 88, 49, 95, 26 ]
184
183
(95 + 88)
1
9
4
[ 87, 10, 10, 96, 58, 85, 43 ]
59
59
((10 / 10) + 58)
0
10
7
[ 36, 10, 71, 68, 13, 65 ]
343
343
(((68 - (71 - 36)) * 10) + 13)
0
10
6
[ 69, 49, 34, 64 ]
622
616
(((69 - 49) * 34) - 64)
6
4
4
[ 44, 20, 79, 12, 22 ]
487
487
((((20 * 22) - 44) + 79) + 12)
0
10
5
[ 46, 33, 76, 11 ]
207
214
(((33 / 11) * 46) + 76)
7
3
4
[ 70, 52, 27 ]
51
52
52
1
9
3
[ 66, 18, 33, 43, 59, 15, 74 ]
149
149
(((74 - (66 / 33)) + 59) + 18)
0
10
7
[ 29, 24, 2, 28, 72, 65 ]
548
548
(((29 - (24 / 2)) * 28) + 72)
0
10
6
[ 49, 92, 76 ]
779
784
((92 - 76) * 49)
5
5
3
[ 74, 17, 30 ]
519
510
(17 * 30)
9
1
3
[ 87, 4, 2, 56, 41 ]
757
756
((((2 * 87) - 41) + 56) * 4)
1
9
5
[ 95, 19, 93, 98, 52 ]
246
246
((((98 - 95) * 93) - 52) + 19)
0
10
5
[ 90, 21, 56 ]
65
69
(90 - 21)
4
6
3
[ 10, 78, 7, 20, 35 ]
334
333
(((7 * 35) + 10) + 78)
1
9
5
[ 77, 69, 1, 4 ]
292
284
(((4 - 1) * 69) + 77)
8
2
4
[ 39, 28, 11, 100, 33 ]
258
257
((((39 - 33) * 28) + 100) - 11)
1
9
5
[ 61, 94, 86, 51, 62, 71 ]
997
998
(((((86 - 71) * 61) - 62) + 51) + 94)
1
9
6
[ 5, 72, 50, 96 ]
519
514
(((72 + 50) * 5) - 96)
5
5
4
[ 70, 92, 25, 43, 49, 48, 65 ]
842
842
(((25 - (49 - 43)) * 48) - 70)
0
10
7
[ 71, 44, 98, 30, 30 ]
239
243
(((71 + 98) + 44) + 30)
4
6
5
[ 50, 12, 41, 67, 59 ]
886
888
(((59 - 41) * 50) - 12)
2
8
5
[ 8, 94, 62, 5, 36, 7, 42, 87 ]
55
55
(62 - 7)
0
10
8
[ 78, 98, 19, 91, 10 ]
579
578
((((78 + 19) - 91) * 98) - 10)
1
9
5
[ 75, 4, 90, 9 ]
185
178
(((90 + 75) + 9) + 4)
7
3
4
[ 63, 54, 42, 61, 38 ]
175
174
((((61 + 63) + 54) + 38) - 42)
1
9
5
[ 86, 59, 54, 61, 72, 80, 34, 6 ]
114
114
(80 + 34)
0
10
8
[ 17, 9, 94, 77, 30 ]
829
829
((9 * 94) - 17)
0
10
5
[ 11, 79, 81, 78, 75, 24, 32, 58 ]
441
441
((((75 - 58) * 32) - 79) - 24)
0
10
8
[ 50, 5, 96, 73, 23 ]
88
88
((((5 * 23) - 50) + 96) - 73)
0
10
5
[ 81, 7, 71, 91, 66 ]
15
15
(81 - 66)
0
10
5
[ 2, 48, 44, 2, 56, 93, 28 ]
58
58
(2 + 56)
0
10
7
[ 93, 67, 78, 85 ]
844
842
(((78 - 67) * 85) - 93)
2
8
4
[ 27, 37, 31 ]
302
310
((37 - 27) * 31)
8
2
3
[ 80, 12, 20, 5, 83, 25, 24, 58 ]
739
739
((((80 - 20) * 12) + 24) - 5)
0
10
8
[ 86, 23, 78, 85, 36, 64, 68, 20 ]
884
884
((36 - 23) * 68)
0
10
8

Countdown Numbers Game Dataset

This dataset contains configurations and solutions for variations of the Countdown numbers game. Each example comprises a sequence of numbers, a target number, the computed solution (closest value), the arithmetic expression that achieves that value, the difference between the target and the computed value, and the final Countdown score.

HuggingFace Download Links

Dataset Variant Dataset Name Download
Random countdown-numbers-3-8 🤗 HuggingFace
Random Solvable countdown-numbers-3-8-nz 🤗 HuggingFace
Coundown Game Rules countdown-numbers-6-gr 🤗 HuggingFace

Dataset Overview

Each data point in the dataset includes:

  • Numbers:
    A sequence of $n$ integers $s_1, s_2, \ldots, s_n$ where $s_i \in {1, 2, \ldots, 100}$ for all $i \in {1, 2, \ldots, n}$, and $n \in {3, 4, \ldots, 8}$. (Note: In the traditional Countdown game, the numbers are subject to more specific restrictions.)

  • Target:
    An integer $t \in {1, 2, \ldots, 999}$. (For context, the standard Countdown game usually features targets from 101 and above.)

  • Closest:
    The value computed by a solver $r \in {1, 2, \ldots, 999}$ that is closest to the target number.

  • Expression:
    The arithmetic expression used to compute the closest value. For instance, $((2 + 48) \times 5) \div 10$

  • Delta:
    The absolute difference between the target and the closest value, i.e. $|t - r|$.

  • Score:
    The Countdown score calculated as $\max(0, 10 - |t - r|)$. This score reflects how close the computed value is to the target.


Dataset Variants

This dataset is provided in three variants:

  1. Random:
    Configurations and solutions generated by uniformly sampling and solving one million game instances, without additional restrictions.

  2. Random Solvable (Score > 0):
    Configurations are generated by uniformly sampling numbers and then rejecting any sample that results in an unsolvable instance (i.e., a score of 0). This variant ensures that each instance has a solution that yields a positive score.

  3. Countdown:
    Configurations generated by sampling 6 numbers in the style of the British TV show Countdown.

Score Distributions

The following histograms show the distribution of scores for each dataset variant:

Random Variant

Random Solvable (Score > 0) Variant

Countdown Game Rules


Generation Process

The dataset was created by:

  • Uniformly sampling numbers within the specified ranges.
  • Solving each sampled instance to determine the closest value, the corresponding expression, the difference from the target, and the score.
  • For the Random Solvable (Score > 0) variant, rejection sampling was applied: instances that did not yield a positive score were discarded.

The train and test splits were created by randomly partitioning the instances into 80% training and 20% testing, using a stratified split based on the score and number of starting values.

Split Score/Size Distributions

The final distributions of scores and numbers are shown in the following histograms:

Random Variant

Random Solvable (Score > 0) Variant

Countdown Game Rules


How to Use the Dataset

You can load and use this dataset with the Hugging Face datasets library. For example:

from datasets import load_dataset

dataset = load_dataset("alexjackson17/countdown-numbers-6-gr")

# Example: Access the first entry in the training split
example = dataset["train"][0]
print("Numbers: ", example["starting"])
print("Target: ", example["target"])
print("Closest: ", example["closest"])
print("Expression: ", example["expression"])
print("Difference: ", example["delta"])
print("Score: ", example["score"])

Citation

If you use this dataset in your research or projects, please cite it as follows:

@misc{jackson2025countdown,
  title = {Countdown Numbers Game Dataset},
  author = {Alex Jackson},
  year = {2025},
  note = {Released under the MIT License},
}

Funding Attribution

This work was supported by UK Research and Innovation [grant number EP/S023356/1], in the UKRI Centre for Doctoral Training in Safe and Trusted Artificial Intelligence (www.safeandtrustedai.org).


License

This dataset is released under the MIT License. See the LICENSE file for more information.

For questions, feedback, or further information, please contact Alex Jackson.

Downloads last month
11