Skip to main content

Trinity College Dublin, The University of Dublin

Menu Search


Trinity College Dublin By using this website you consent to the use of cookies in accordance with the Trinity cookie policy. For more information on cookies see our cookie policy.

      
Profile Photo

Professor Thomas Chadefaux

Professor in Political Science (Political Science)
NO.2 CLARE STREET
      
Profile Photo

Professor Thomas Chadefaux

Professor in Political Science (Political Science)
NO.2 CLARE STREET

 chadefat@tcd.ie

 3531896 3484


Thomas Chadefaux holds a Ph.D. in political science from the University of Michigan, and an M.A. from the Graduate Institute of International Studies in Geneva. Prior to becoming an assistant Professor of Political Science at Trinity College Dublin, he has been a visiting assistant professor in the department of political science at the University of Rochester and a postdoctoral researcher at ETH Zurich.
  BARGAINING   BARGAINING POWER   BARGAINING THEORY   civil war   CONFLICT   Econometrics/Forecasting   FORECASTING   Game Theory   international conflict   Interstate war
Project Title
 Patterns of Conflict Emergence
From
To
Summary
Are there recurring patterns in the escalation and emergence of wars? The idea that history may repeat itself is old. But recent advances overcoming methodological and data barriers present an opportunity to identify these recurrences empirically and to examine whether these patterns can be classified to improve forecasts and inform theories of conflict. I propose to combine new methods-using the shape of the sequence of events rather than its raw values-and novel data on conflict from finance, diplomatic cables, and newspapers, to extract typical pre-war motifs. Just as DNA sequencing has been critical to medical diagnoses, PaCE aims to diagnose international politics by uncovering the relevant patterns in the area of conflict. Our goals are to:
(i) Identify patterns in the pre-conflict actions using data on conflict events-from the onset of WWI to Hamas's rocket launches-and in their perceptions using data from financial markets (the "crowd's" perception), news articles (the "experts"), and diplomatic documents (the policy-makers). This will allow us to evaluate the patterns of escalation over different timescales-from the decade to the minute. The similarity between temporal sequences will be measured using algorithms which allow for flexible matching, such as Dynamic Time Warping.
(ii) Evaluate the utility of these patterns to improve forecasts of conflict with both historical and live out-of-sample predictions. Our results, using shape-based classification methods, will be made public and evaluated in real time. Moreover, using new measures of complexity to distinguish regular, chaotic, and random behavior, I will measure possible fundamental limits to the predictability of conflict events.
(iii) Summarize the core features of dangerous patterns into motifs-recurring patterns-that can help build new theories of conflict emergence and escalation. PaCE will build a repository of shapes-a grammar of patterns-to be used as the building blocks of new theories.
Funding Agency
European Research Council (ERC)
Programme
ERC Consolidator Grant
Project Title
 The steps to war: Developing an automated pattern recognition system for conflict
From
2019
To
2023
Summary
Funding Agency
Trinity College Dublin
Programme
Provost's PhD Project Award
Project Title
 The Patterns of Conflict Escalation: Why leaders, markets, and analysts fail to anticipate wars
From
2018
To
2019
Summary
Funding Agency
Enterprise Ireland
Programme
H2020
Project Title
 Explaining the variation in refugee flows: the effect of rebel group dynamics in civil conflicts
From
2018
To
2022
Summary
Funding Agency
Irish Research Council
Programme
Postgraduate Scholarship Award
Project Title
 Explaining variation in the diffusion of civil wars: the role of short-term triggers.
From
2017
To
2021
Summary
Funding Agency
Irish Research Council
Programme
Postgraduate Scholarship Award

Details Date
. Editorial Board Member, Data Science
Journal reviewer
I have reviewed article manuscripts for the following journals (in alphabetical order):
Advances in Complex Systems, American Journal of Political Science, American Political Science Review, British Journal of Political Science, Conflict Management and Peace Science, Empirical Economics, International Interactions, International Studies Quarterly, Journal of Artificial Societies and Social Simulation, Journal of Conflict Resolution, Journal of Economic Interaction and Coordination, Journal of Global Security Studies, Journal of Information Technology \& Politics, Journal of Mathematical Sociology, Journal of Peace Research, Journal of Politics, Journal of the Royal Society Interface, PLoS One, Political Science Research and Methods. 		2008-
External examiner, PhD viva, University of St Andrews Jan. 2020
External examiner, hiring committee, UCD 2018
Language Skill Reading Skill Writing Skill Speaking
English Fluent Fluent Fluent
French Fluent Fluent Fluent
German Fluent Medium Medium
Japanese Basic Basic Basic
Details Date From Date To
American Political Science Association 2004
American Economic Association 2009
Midwest Political Science Association 2005
Chien Lu, Thomas Chadefaux, Structured Pixels: Satellite Imagery as the Cause in Causal Effect Estimation, Hawaii International Conference on System Sciences, Maui, Hawaii, January 6-9, 2026, 2026, Conference Paper, PUBLISHED  TARA - Full Text
Håvard Hegre and ... and Cornelius Fritz ..., The 2023/24 VIEWS Prediction Challenge: Predicting the Number of Fatalities in Armed Conflict, with Uncertainty, Journal of Peace Research, 2025, Journal Article, IN_PRESS  DOI
Thomas Chadefaux and Thomas Schincariol, Temporal Patterns in Migration Flows Evidence from South Sudan, Journal of Forecasting, 44, (2), 2025, p575 - 588, Journal Article, PUBLISHED  DOI
T Schincariol, H Frank, T Chadefaux, Temporal Patterns in Conflict Prediction: An improved Shape-Based Approach, Journal of Peace Research, 2025, Journal Article, ACCEPTED
Yani Xue, Thomas Schincariol, Thomas Chadefaux, Derek Groen, Using machine learning to forecast conflict events for use in forced migration models, Scientific Reports, 2025, Journal Article, PUBLISHED  DOI  URL
Thomas Schincariol, Hannah Frank, Thomas Chadefaux, Accounting for variability in conflict dynamics: A pattern-based predictive model, Journal of Peace Research, 2025, Journal Article, PUBLISHED  DOI
Thomas Chadefaux and Jian Cao, Dynamic Synthetic Controls, Political Analysis, 2024, Notes: [doi:10.1017/pan.2024.14], Journal Article, PUBLISHED  DOI
Turkoglu Oguzhan, Chadefaux Thomas, The effect of terrorist attacks on attitudes and its duration, Political Science Research and Methods, 2023, Journal Article, PUBLISHED  TARA - Full Text  DOI
Thomas Chadefaux, An automated pattern recognition system for conflict, Journal of Computational Science, 2023, Journal Article, PUBLISHED  DOI
Boussalis, C., Chadefaux, T., Decadri, S., Salvi, A., Public and Private Information in International Crises: Diplomatic Correspondence and Conflict Anticipation, International Studies Quarterly, 2022, Journal Article, PUBLISHED  TARA - Full Text  DOI
  

Page 1 of 3
Thomas Chadefaux and Humanitarian Forecast Working Group, Harnessing AI: How to develop and integrate automated prediction systems for humanitarian anticipatory action, CEPR Policy Insight, 2024, Journal Article, PUBLISHED
Jian Cao and Thomas Chadefaux, 'DSC: Dynamic Synthetic Controls', 2024, -, Software, PUBLISHED
Thomas Chadefaux, The Triggers of War: Disentangling the Spark from the Powder Keg., 2014, Working Paper, SUBMITTED

  


Award Date
Best paper award, American Political Science Review 2018
Best Paper award, Oxford, "Internet, Politics, Policy" conference Sept. 2012
Best visualisation award, Journal of Peace Research 2014
Best visualisation award 2014
There have been more than 200 wars since the start of the 20th century, leading to about 35 million battle deaths. This recurrence of wars despite their tremendous economic, social and institutional costs, may suggest that we are doomed to repeat the errors of the past. Unfortunately, we still know little about the predictability of conflict.

In particular, are forecasting failures due to limitations of our models, data, or assumptions? Or are there simply aspects of conflicts that will always remain fundamentally unpredictable? Are there systematic patterns in the escalation and emergence of conflict, and can these patterns be clustered and classified in meaningful ways that help us improve future forecast?

My current research generally focuses on the causes of interstate conflict and on their prediction. In particular, I rely on large amounts of fine-grained spatial and temporal data (e.g., newspapers, satellite images, financial markets) to reveal early warning signals for war. Of special interest is whether decision-makers correctly anticipate the risks of war, or rather `sleepwalk' into conflict. Estimating their perceptions of risk is therefore central to my work, which relies on financial data, text analysis of news reports and of diplomatic cables to understand when and how leaders, journalists and the public anticipate wars. Methodologically, my work relies mostly on statistical methods and machine learning approaches.

This work leads to three main types of outputs. First, empirical contributions to the question of observers' perceptions of risk. If we can identify the causes of underestimation of risk, we can suggest corrections to learn from history. Second, forecasts of conflicts--here I seek to build on and improve existing forecasts as part of an international network, and on that basis to advise governments (e.g., German Department of Foreign affairs) and international organisations (e..g, EU). Finally, the work leads to theoretical contributions using game theory to better understand the causes of war. This work has been published in the field's top journals.