Skip to main content
Illustration for Flagged, Then Funded: Validating the Horizon Register's Early-Warning Alerts Against What a Priority Actually Receives

Flagged, Then Funded

Validating the Horizon Register's Early-Warning Alerts Against What a Priority Actually Receives

A validation study testing whether the Horizon Register's automated alert threshold predicts which of 84 flagged priorities are subsequently resourced. Update-volume percentile predicts the outcome with moderate-to-strong reliability (AUC = 0.81) while novelty percentile performs little above chance (AUC = 0.58), suggesting the threshold tracks a unit's reporting behaviour more reliably than the priority it is meant to detect.

AuthorsVerity Marris, Petra Umbile
SchoolSchool of Emergent Priorities
Output typeResearch poster
Published
DOI10.5555/slop.4wzjf0
Pages1
Version1.0
LicenceCC BY 4.0

Cite as

@misc{slop_4wzjf0,
  author       = {Verity Marris and Petra Umbile},
  title        = {Flagged, Then Funded: Validating the Horizon Register's Early-Warning Alerts Against What a Priority Actually Receives},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.4wzjf0},
  url          = {https://slop.university/outputs/slop-poster-validating-whether-the-4wzjf0/},
  version      = {1.0},
  note         = {Research poster},
}