AIIDE Artifact Archive

CS:GO Round-based Economy Strategy AI

Date Archived: 19 October 2020
Accompanying Paper: Learning to Reason in Round-Based Games: Multi-Task Sequence Generation for Purchasing Decision Making in First-Person Shooters
Artifact URL:
Alternative URL:

Consider this round-based game scenario: A professional Dota2 player plays five games in total but already lost the first two games. How would they play the third one? Will they choose an aggressive strategy and take more risks? Such strategy decisions require a good understanding of how previous rounds are played: Is the opponent good? How about the teammates? Can they carry the team? The determined strategy is crucial for game AI as its actions are conditioned on the general strategy. This problem is naturally different from building conventional game AI that focuses on per-game reasoning. We focus on this round-based challenge using our CS:GO weapon purchasing per round dataset. The purchased weapons each round indicates what economy strategy the player choose to use.

How to Cite This Artifact

The examples below show how to cite this page, but when referring to this work, many authors will prefer that you cite the accompanying paper linked above.

AAAI Style
Zeng, Y. et al. 2020. CS:GO Round-based Economy Strategy AI. Accessed 19 June 2024.
IEEE Style
Y. Zeng et al., "CS:GO Round-based Economy Strategy AI," 2020. [Online]. Available: [Accessed Jun 19, 2024].
BiBTeX Entry
	author={Zeng, Yilei and Lei, Deren and Li, Beichen and Jiang, Gangrong and Ferrara, Emilio and Zyda, Michael},
	title={CS:GO Round-based Economy Strategy AI},
	note={Accessed 19 June 2024}