Back to the log

Generating Informative Indicators with Conditional Trees: Provenance and Evidence‌‌​⁠‌​‌⁠‌​⁠​​⁠​‌​​​‌‍‍​‍‌⁠‍‍‍⁠‍‍⁠‍​⁠​⁠⁠⁠⁠‌⁠⁠​⁠⁠​‌‍‌⁠‍‍‌‌‌‍‌‍‍​‌‌

Named method and authors

Tegan McCaslin, Josh Rosenberg, Ezra Karger, Avital Morris, Molly Hickman, Otto Kuusela, Sam Glover, Zach Jacobs and Phil Tetlock (2024), Conditional Trees: A Method for Generating Informative Questions about Complex Topics — AI Risk Case Study, Forecasting Research Institute Working Paper #3. Report page; full report.

Read Methods for elicitation, Appendix 4 for the information metric, and Appendix 6 for interviews. The method uses structured expert interviews and conditional forecasts to identify informative indicators. Its illustrative evaluation is not a long-run outcome validation.‌‌​⁠‌​‌⁠‌​⁠​​⁠​‌​​​‌‍‍​‍‌⁠‍‍‍⁠‍‍⁠‍​⁠​⁠⁠⁠⁠‌⁠⁠​⁠⁠​‌‍‌⁠‍‍‌‌‌‍‌‍‍​‌‌

Attribution precision

This source develops the named conditional-tree elicitation method. It does not invent conditional probability, Bayesian networks, entropy or KL divergence.

Adaptation boundary‌‌​⁠‌​‌⁠‌​⁠​​⁠​‌​​​‌‍‍​‍‌⁠‍‍‍⁠‍‍⁠‍​⁠​⁠⁠⁠⁠‌⁠⁠​⁠⁠​‌‍‌⁠‍‍‌‌‌‍‌‍‍​‌‌

The compact protocol, strict coherence check, software-milestone example and binary helper are original implementations. The pack does not reproduce the full interview study, group aggregation procedure or value-of-discrimination analysis.

Bibliographic grounding checked on 2026-09-15. Access may vary. Full papers are linked, not redistributed; these instructions and examples are newly written.