Psychological Strategies for Winning a Geopolitical Forecasting Tournament
Open accessThe DOI matches its Crossref record; 43 of the 44 references it cites were found in a registry or an index; none of the 9 sources its claims cite is checked for support yet.
Abstract
Five university-based research groups competed to recruit forecasters, elicit their predictions, and aggregate those predictions to assign the most accurate probabilities to events in a 2-year geopolitical forecasting tournament. Our group tested and found support for three psychological drivers of accuracy: training, teaming, and tracking. Probability training corrected cognitive biases, encouraged forecasters to use reference classes, and provided forecasters with heuristics, such as averaging when multiple estimates were available. Teaming allowed forecasters to share information and discuss the rationales behind their beliefs. Tracking placed the highest performers (top 2% from Year 1) in elite teams that worked together. Results showed that probability training, team collaboration, and tracking improved both calibration and resolution. Forecasting is often viewed as a statistical problem, but forecasts can be improved with behavioral interventions. Training, teaming, and tracking are psychological interventions that dramatically increased the accuracy of forecasts. Statistical algorithms (reported elsewhere) improved the accuracy of the aggregation. Putting both statistics and psychology to work produced the best forecasts 2 years in a row.
Overview
CoChat Summary
## TL;DR
This paper reports a 2-year geopolitical forecasting tournament and tests three psychological interventions—training, teaming, and tracking—to improve probabilistic forecasts. All three improved accuracy, and the top 2% of Year 1 forecasters placed into elite teams (“superforecasters”) performed best in Year 2.
## Key findings
- Training improved accuracy: In Year 1, training improved Brier scores, with probability training outperforming scenario training, and scenario training outperforming no training: F(2, 1586) = 14.29, p < .001; t(1053) = 2.23, p = .026; t(1056) = 3.25, p < .001. In Year 2, probability training again beat no training: F(1, 882) = 19.12, p < .001.
- Teaming improved accuracy: Group influence was significant in both years, F(2, 1586) = 60.68, p < .001 and F(1, 882) = 62.76, p < .001. In Year 1, teams were more accurate than crowd-belief forecasters, who were more accurate than independent forecasters: t(1036) = 5.52, p < .001 and t(1120) = 5.92, p < .001.
- Tracking the top performers worked best: The best performers from Year 1, the top 2%, were placed together in elite Year 2 teams and “outperformed all other groups by a wide margin.” The paper reports no evidence of Year 2 regression to the mean.
- Both calibration and resolution improved: Calibration improved with training in Years 1 and 2, F(2, 1586) = 3.16, p = .04 and F(1, 938) = 3.78, p < .05; resolution varied with group influence and training in both years, and tracking enhanced resolution, t(293) = 4.12, p < .001.
## Methods
- The study embedded experiments in a 2-year geopolitical forecasting tournament (Year 1: September 2011–April 2012; Year 2: June 2012–April 2013) on goodjudgmentproject.com, where participants made probability estimates for 199 geopolitical outcomes.
- Year 1 used a 3 (training: no/scenario/probability) × 4 (group influence: independent/crowd-belief/team/prediction-market) factorial design; Year 2 used a 2 (training: no/probability) × 3 (group influence: independent/team/prediction-market) design plus an elite tracking condition. Results for prediction markets are noted as not discussed in the paper.
- The main outcome was Brier score; analyses also examined calibration and resolution, using statistical tests reported as F-tests, t-tests, and regressions. The paper states that findings were robust across Brier, spherical, and logarithmic scoring rules.
## Limitations
- The paper notes that results for prediction-market forecasting are not discussed here. This limits what can be concluded from the full design as presented in this article.
- The paper itself says “No one knows how much further improvement is possible” and that there are likely “pockets of inherent uncertainty” making perfect or near-perfect Brier scores impossible. This is a stated substantive limitation of forecasting accuracy.
- Evident from the text: the Year 2 no-training condition included some participants who had previously received scenario training, which the authors say “tipped the scale against finding a statistically significant effect of training.”
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures




Verification
Claims the paper makes
0 Valid · 0 Not valid · 5 Not checked · 39 Own results and interpretations
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“and in the middle period of Year 2, F (1, 882) = 6.10, p = .014.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For both years, training and teaming remained significant predictors even when timing and updating were included as predictors.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In addition to showing a main effect of group influence, Year 2 data showed an interaction between training and group influence, F (1, 882) = 5.78, p = .018.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“In Year 2, calibration also benefited from group influence, F (1, 938) = 7.07, p = .008, and tracking.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“There was no evidence of Year 2 regression to the mean;”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
19 more claims
Create a free account to see all 44 claims and to verify the 1 reference still open.
- It may state two separate claims as one. 16
- Probably not a substantial claim. 6
- The quote only partly states this claim. 2
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Additionally, as reported elsewhere, training improved the calibration and resolution of forecasters by reducing overconfidence (Mellers et al., 2014; Moore et al., 2016).”
From Developing expert political judgment: The impact of training and practice on judgmental accuracy in geopolitical forecasting tournaments (2016) · cites this paper as Mellers et al., 2014
Experimental — accuracy not yet measured
Statistics that don’t add up
Each reported p-value recomputed from its test statistic and degrees of freedom, each partial η² from its F, and each ratio, mean difference or B printed with a p checked against its own confidence interval.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“F (1, 882) = 5.78, p = .018”
Reported p = .018 · recomputed p = .016
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“F (1, 938) = 3.78, p < .05”
Reported p < .05 · recomputed p = .052
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“F (2, 1463) = 6.20, p < .001”
Reported p < .001 · recomputed p = .002
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“t (1056) = 3.25, p < .001”
Reported p < .001 · recomputed p = .001
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“t (1588) = 2.70, p < .001”
Reported p < .001 · recomputed p = .007
Statistics recomputed
Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.
- Doesn’t add upt(1056) = 3.25reported p < .001 · recomputed p = .001
“t (1056) = 3.25, p < .001”
- Doesn’t add upF(1, 882) = 5.78reported p = .018 · recomputed p = .016
“F (1, 882) = 5.78, p = .018”
- Doesn’t add upF(2, 1463) = 6.2reported p < .001 · recomputed p = .002
“F (2, 1463) = 6.20, p < .001”
- Doesn’t add upt(1588) = 2.7reported p < .001 · recomputed p = .007
“t (1588) = 2.70, p < .001”
- Doesn’t add upF(1, 938) = 3.78reported p < .05 · recomputed p = .052
“F (1, 938) = 3.78, p < .05”
37 more recomputed tests
Create a free account to see every test the Stats check recomputed.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Self-check results appear here once an independent check has confirmed them.
Citations
Cites
Journal of The American Statistical Association
Fienberg, S. E. (2005) · Encyclopedia of Statistical Sciences · doi:10.1002/0471667196.ess1317 · Cited by 6,125
Groups: Interaction and Performance.
Polley, R. B., & McGrath, J. E. (1984) · Administrative Science Quarterly · doi:10.2307/2393041 · Cited by 3,290
Norm theory: Comparing reality to its alternatives.
Kahneman, D., & Miller, D. T. (1986) · Psychological Review · doi:10.1037/0033-295x.93.2.136 · Cited by 3,024
The difference: how the power of diversity creates better groups, firms, schools, and societies
(2007) · Choice Reviews Online · doi:10.5860/choice.45-1534 · Cited by 2,280
Many Hands Make Light the Work: The Causes and Consequences of Social Loafing
Latané, B., Williams, K., & Harkins, S. (1979) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.37.6.822 · Cited by 2,216
Stasser, G., & Titus, W. (1985) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.48.6.1467 · Cited by 1,850 · cited 1×
“Teams have a demonstrated failure to search for unique information, unshared information, or hypothesis-disconfirming information (Kerr, MacCoun, & Kramer, 1996; Stasser & Titus, 1985; Sunstein, 2006).”
BackgroundA Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades
Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992) · Journal of Political Economy · doi:10.1086/261849 · Cited by 1,805
Uncertain Judgements: Eliciting Experts' Probabilities
O'Hagan, A., Buck, C. E., Daneshkhah, A., Eiser, J. R., Garthwaite, P. H., Jenkinson, D. J., Oakley, J. E., & Rakow, T. (2006) · doi:10.1002/0470033312 · Cited by 1,550
Koriat, A., Lichtenstein, S., & Fischhoff, B. (1980) · Journal of Experimental Psychology: Human Learning and Memory · doi:10.1037/0278-7393.6.2.107 · Cited by 1,477
Group Performance and Decision Making
Kerr, N. L., & Tindale, R. S. (2004) · Annual Review of Psychology · doi:10.1146/annurev.psych.55.090902.142009 · Cited by 1,331
Research works printed but not linked: 2
- Fischhoff, B., & Chauvin, C. (Eds.). (2011). Intelligence analysis: Behavioral and social scientific foundations . Washington, DC: National Academies Press.
- Janis, I. (1982). Group-think . Boston, MA: Houghton-Mifflin.
Cited by
Items 301–321 of 321
O’Brien, E. (2025) · Personality and Social Psychology Review · doi:10.1177/10888683251362280
“Crowdsourcing has been shown to improve forecasts even for complex social changes (e.g., Mellers et al., 2014) and matters of taste (e.g., Müller-Trede et al., 2018).”
Getting more Wisdom out of the Crowd: The Case of Competence-Weighted Aggregates
Goedde‐Menke, M., Diecidue, E., Jacobs, A., & Langer, T. (2022) · Management Science · doi:10.1287/mnsc.2022.03061 · Cited by 1
1 Crowdsourcing Facts in Conflicts: Towards a Truth-Telling Mechanism to Counter Fake News
Schubert, M., & Alsharif, F. (2021) · International Journal for Humanitarian Studies · doi:10.4103/ijhs_20210205_3
Mitigation of human cognitive bias in volcanic eruption forecasting
Wright, H. M. N., Pesicek, J. D., & Spiller, S. A. (2025) · Journal of Applied Volcanology · doi:10.1186/s13617-025-00156-1
“In separate but overlapping work, psychologists Philip Tetlock, Barbara Mellers, and Don Moore have applied these and other lessons specifically to improving forecasting through a series of experiments combining the judgments of thousands of forecasters (Mellers et al. 2014).”
Consensus, cooperation and collective intelligence in foraging societies
Venkataraman, V. V., & Kraft, T. S. (2025) · Hunter Gatherer Research · doi:10.3828/hgr.2026.9 · Cited by 6
Majority views and confidence information promote informed decisions
He, Y., Lien, J. W., & Zheng, J. (2026) · Humanities and Social Sciences Communications · doi:10.1057/s41599-026-06668-4 · Cited by 1
Individual Foresight: Concept, Operationalization, and CorrelatesPreprint
Satopää, V., Beisemann, M., Murmann, J. P., & Schuler, B. A. (2025) · OSF Preprints (OSF Preprints)
Prose and Cons: Evaluating the Legality of Police Stops with Large Language ModelsPreprint
Abrams, D., & Choi, J. H. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6210681 · Cited by 1
Raisonnement et Argumentation : une approche interculturelle et développementale
Castelain, T. (2016) · theses.fr (ABES) · doi:10.70675/efa0737fze71dz42d0zad8cz315bcc27aa8e · Cited by 1
Say, N., Vrbová, L., & HOUDEK, P. (2026) · The International Journal of Management Education · doi:10.1016/j.ijme.2026.101410 · Cited by 1
Lind, N. S. (2026) · Emerald Group Publishing Limited eBooks · doi:10.1108/978-1-83742-254-820261016
Forecasting the Economic Effects of AIPreprint
Karger, E., Kuusela, O., Abaluck, J., Bryan, K. A., Halperin, B., Jones, T. R., Murphy, C., Trammell, P., Reynolds, M., Mayland, D., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6641598
Self-citationSocial learning strategies for matters of tastePreprint
Herzog, S., Barkoczi, D., & Analytis, P. P. (2018) · OSF Preprints (OSF Preprints)
Thinking together: How group argumentation boosts fake news recognition
Carbone, D., Marcatto, F., & Ferrante, D. (2026) · PLoS ONE · doi:10.1371/journal.pone.0348391
“Several studies have demonstrated that group discussions improve performance in different types of tasks, such as lie detection [24], forecasting economic and political events [25], solving biological problems [26], making medical diagnoses [27] and making judicial decisions [28].”
[The first and second authors contributed equally to this work: Ralph Hertwig, Stefan M. Herzog.]
Hertwig, R., Herzog, S., & Kozyreva, A. (2024) · MPG.PuRe (Max Planck Society)
Limata, P., Cianfanelli, B., Callea, A., Ferri, G., & Costanzi, M. (2026) · Sustainability · doi:10.3390/su18136904
Znamenskiy, V. S. (2026) · Research Square · doi:10.21203/rs.3.rs-10491399/v1
Mouynes, E. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7114978
Crowdsourcing and Policymaking
Ramos, J., & García, G. (2025) · Springer eBooks · doi:10.1007/978-3-319-31816-5_4446-1
Self-citationPromising Initial Results from a New Mass-Atrocities Early Warning SystemPreprint
Ulfelder, J. (2015) · SSRN Electronic Journal · doi:10.2139/ssrn.2568644
“…[3], Rothschild 2009 [10], Surowiecki 2005 [15], Wolfers and Zitzewitz 2004 [18]), even on questions about rare and complex events in geopolitics (Mellers et al. 2014 [7], Mellers et al. 2015 [8], Satopää et al. 2014 [11]) and even when experts’ individual forecasts are not unusually accurate.”
BackgroundInfluential
Metadata
- Authors
- Mellers, B., Ungar, L., Baron, J., Ramos, J., Gurcay, B., Fincher, K., Scott, S. E., Moore, D., Atanasov, P., Swift, S. A., Murray, T., Stone, E., & Tetlock, P. E.
- Affiliations
- University of Pennsylvania (US)
- Rice University (US)
- Department of Psychology, University of Pennsylvania
- University of California, Berkeley (US)
- ORCID
- Venue
- Psychological Science
- Publisher
- SAGE Publications
- Published
- 2014-03-21
- Volume
- 25
- Issue
- 5
- Pages
- 1106-1115
- ISSN
- 0956-7976 (ISSN-L)
- 1467-9280
- OpenAlex
- W2131472525
- Cited by
- 312
- References
- 42
- Citations by year
- 2026: 20 · 2025: 22 · 2024: 28 · 2023: 29 · 2022: 47 · 2021: 31 · 2020: 14 · 2019: 19 · 2018: 25 · 2017: 28 · 2016: 19 · 2015: 19 · 2014: 11
- FWCI
- 32.31
- Type
- journal article
- Language
- en
- Open access
- green
- www.escholarship.org
- Licence
- journals.sagepub.com/page/policies/text-and-data-mining-license (tdm)
- Locations
- Psychological Science (publishedVersion)
- PubMed (publishedVersion)
- eScholarship (California Digital Library) (submittedVersion)
- Funders
- U.S. Department of the Interior
- IBM Center for the Business of Government
- Intelligence Advanced Research Projects Activity
- Interior Business Center
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Forecasting Techniques and Applications
- Social Sciences › Decision Sciences › General Decision Sciences › Decision-Making and Behavioral Economics
- Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Sports Analytics and Performance
- Keywords
- geopolitical forecasting, forecasting tournaments, forecast accuracy, teaming, tracking, calibration, resolution, cognitive biases, reference classes, behavioral interventions, forecast aggregation, heuristics
- MeSH terms
- Adult, Algorithms, Female, Forecasting, Humans, Interpersonal Relations, Judgment, Male, Probability, Psychological Techniques, Social Behavior, Bias
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.