Predicting Russian-Ukrainian War with GPT-3 and the Predictioneer's Game

• In Event: Poster Session: Politics of Post-Communist Countries & Politics and History
  In Poster Session: Poster Session: Politics of Post-Communist Countries

Abstract

Predicting the future developments of the Russian-Ukrainian War is a complex task, as it involves multiple factors that are difficult to anticipate. In this paper, I propose a two-staged approach that utilizes a large language model, such as GPT-3, as an information retrieval tool and combines it with strategic calculations based on the "Predictioneer's Game" model. The first stage of the proposed approach involves using GPT-3 to generate data on the Russian-Ukrainian War by “asking” relevant questions about the war. The second stage involves using the "Predictioneer's Game" model to make strategic calculations based on the data generated in the first stage. The model's ability to analyze complex situations and make predictions based on a wide range of factors can provide additional insights into the future developments of the Russian-Ukrainian crisis. In addition, this paper also engages the data from expert interviews that is also used in the "Predictioneer's Game" model. The paper compares both approaches and draws important conclusions about the future of the war. The results of this research can provide valuable insights for policymakers and scholars studying the Russian-Ukrainian crisis and help them make more informed decisions.

Author

• Kirill Kalinin, Stanford University