exam-forecast

You forecast exam topics by loading past exams and a local class context file to compute stable versus variable pattern signals. The Skill maps these historical exam patterns to your current syllabus, generating a dated study notes file with subject weighting and a next-steps decision tree.

Analyzing past exam patterns with a small sample size flags thin samples and distinguishes stable versus variable pattern signals. The system applies confidence discipline by framing outputs as weighting heuristics rather than absolute predictions, ensuring you handle limited historical data appropriately.

Yes, you can use pasted text instead of PDF files for exam prediction. The workflow accepts user-provided past exams across variations in formats including PDFs, pasted text, or file paths, mapping the observed patterns to your current syllabus for finals preparation.

Confidence signaling works by distinguishing stable versus variable pattern signals from historical exams and flagging thin samples. It frames the generated draft study notes as weighting heuristics rather than definitive predictions, helping you allocate study time across topics with appropriate caution.

Before mapping past exam patterns to your current syllabus, you need to provide past exams in formats like PDF, pasted text, or file paths, and ensure a local CLAUDE.md file is loaded to establish the class context. This setup allows the system to compute pattern signals and output a dated study file.

The limitation of using pattern analysis for law school exam prediction is that outputs are framed as weighting heuristics rather than exact predictions. The system flags thin samples and distinguishes stable versus variable patterns to prevent overreliance on limited historical data.