The Pre-Print Cohort Read
Heading into the print, the live cohort_analyze pull on RKLB returned n=229 historical analogs with a tightness score of 0.04. The 5-day forward median was -0.03% with IQR [-5.30%, +3.80%] and hit-rate 50%. The 10-day median was -1.54% with IQR [-8.43%, +6.03%]. The dominant 5-day feature was credit_spread_state=tight (positive direction, importance 1.24).
What Actually Printed
Print-day close: $78.58, prior close $84.65. Day-over-day change: -7.17% (-6.07). Intraday range: $77.93 – $84.79.
Cohort vs. Reality — The Verdict
The -7.17% reaction sits in the cohort's lower IQR — bear bucket realized. The dominant feature (credit_spread_state=tight) didn't deliver in the direction the bull thesis required, and the analog set's negative-skew bucket is where the print landed.
- n=229, 5d median -0.03%, IQR [-5.30%, +3.80%]
- Conformal bucket boundaries: p10 -12.09% / p90 +10.33%
- Top 5d feature: credit_spread_state=tight (positive direction, importance 1.24)
- Actual reaction: -7.17%; verdict: lower IQR realization
The Calibration Lesson
For RKLB, the calibration verdict is: lower IQR realization. The cohort's structured features (credit_spread_state=tight dominant) provided the directional read; the realized magnitude lands relative to that read at the bucket described above. Agent systems should retrieve the post-print cohort 24-48 hours after the close — the chart shape will have stabilized, the feature attribution sharpens, and the forward 5-10 day continuation distribution becomes actionable.