First Field Support for the Correction Freeze Problem: Evidence for Its Step Trigger and Inherited Freeze

This report presents the first field evidence in support of predictions that belong to the Correction Freeze Problem (CFP) alone. Using 933,448 probability forecasts on 491 geopolitical questions from the Good Judgment Project tournament (2011 to 2015), a setting of voluntary, intermittent monitoring that matches the conditions CFP describes, we tested two of its unique predictions against rules fixed before the analysis. Both passed every criterion. First, forecasters revised when the crowd consensus had just moved, as CFP’s step trigger requires, and not when their own beliefs had drifted far from it: those near a consensus that had just moved revised 7.6 times as often as those far from one that had stood still, and the step model outperformed the gap model by 251 log-likelihood units, a margin whose lower 95% bound is still 199. Second, team members who leaned on a stale team reference responded little to change, while those leaning on a fresh one responded most, the pattern of the inherited freeze that follows from Route 3; this second result is encouraging but weaker, since fresh-reference forecasters were also more active overall. Neither prediction is made by the established frameworks CFP is set against: signal detection, integrating observers, satisficing and rational-inattention models all trigger correction on the accumulated gap, and cue-reliance and cascade accounts predict that reliance dulls responsiveness whether or not the reference is current. The analyses are exploratory; the crowd consensus stands in for the true state, and the support holds for sparse, voluntary monitoring, the domain CFP was written for; the same rule fails in attentive laboratory tasks, and corrections after a long freeze close only part of the gap, which might call for revising CFP’s full-correction clause.

Authors

Institutions

Publication Details

Journal
Knowledge Commons Works (Michigan State University)
Published
2026-11-15
DOI
https://doi.org/10.17613/pbzz4-s7t17
Primary Topic
Decision-Making and Behavioral Economics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

First Field Support for the Correction Freeze Problem: Evidence for Its Step Trigger and Inherited Freeze

Minh-Hoang Nguyen
Knowledge Commons Works (Michigan State University)
Decision-Making and Behavioral Economics
article

First Field Support for the Correction Freeze Problem: Evidence for Its Step Trigger and Inherited Freeze

Minh-Hoang Nguyen
article en

Abstract

This report presents the first field evidence in support of predictions that belong to the Correction Freeze Problem (CFP) alone. Using 933,448 probability forecasts on 491 geopolitical questions from the Good Judgment Project tournament (2011 to 2015), a setting of voluntary, intermittent monitoring that matches the conditions CFP describes, we tested two of its unique predictions against rules fixed before the analysis. Both passed every criterion. First, forecasters revised when the crowd consensus had just moved, as CFP’s step trigger requires, and not when their own beliefs had drifted far from it: those near a consensus that had just moved revised 7.6 times as often as those far from one that had stood still, and the step model outperformed the gap model by 251 log-likelihood units, a margin whose lower 95% bound is still 199. Second, team members who leaned on a stale team reference responded little to change, while those leaning on a fresh one responded most, the pattern of the inherited freeze that follows from Route 3; this second result is encouraging but weaker, since fresh-reference forecasters were also more active overall. Neither prediction is made by the established frameworks CFP is set against: signal detection, integrating observers, satisficing and rational-inattention models all trigger correction on the accumulated gap, and cue-reliance and cascade accounts predict that reliance dulls responsiveness whether or not the reference is current. The analyses are exploratory; the crowd consensus stands in for the true state, and the support holds for sparse, voluntary monitoring, the domain CFP was written for; the same rule fails in attentive laboratory tasks, and corrections after a long freeze close only part of the gap, which might call for revising CFP’s full-correction clause.

Knowledge Commons Works (Michigan State University)
Phenikaa University (VN)
Decision-Making and Behavioral Economics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon