8 for 8: ODIN's Perfect February 2026
By David "Lionel" Moore · March 1, 2026 · 4 min read
February is done. Let's count the tape.
Eight PDUFA decisions hit in February 2026. ODIN called every single one correctly — 5 approvals, 3 Complete Response Letters. 100% directional accuracy for the month.
I'm not going to pretend this is normal. It's not. Even with a 0.9363 holdout AUC, a perfect month is the upper tail of what the model is capable of. But that's the point — ODIN is designed to be right when it's confident, and cautious when the signal is mixed. February happened to be a month where the signal was clear, and the model read it correctly across the board.
The Full Scorecard
| Ticker | Drug | ODIN | Tier | Result | Move |
|---|---|---|---|---|---|
| ASND | TransCon hGH Weekly | 95.5% | TIER_1 | APPROVED | +19% |
| RGNX | RGX-121 | 45.0% | TIER_4 | CRL | -21% |
| IRON | Bitopertin | 39.4% | TIER_4 | CRL | -52% |
| VNDA | Bysanti | 89.7% | TIER_1 | APPROVED | +44% |
| REGN | Dupixent AFRS | 94.8% | TIER_1 | APPROVED | +7.3% |
| ETON | ET-600 | 85.5% | TIER_2 | APPROVED | +14.7% |
| ASND | YUVIWEL | 88.9% | TIER_1 | APPROVED | +6% |
| PHAR | Leniolisib sNDA | 58.0% | TIER_3 | CRL | -28% |
The CRLs Matter More Than the Approvals
Anybody can predict approvals — the base rate is ~68%. The real alpha is in calling the CRLs. When ODIN scored IRON (Bitopertin) at 39.4% and RGNX (RGX-121) at 45.0%, those weren't popular opinions. IRON had the rare disease narrative. RGNX had the gene therapy excitement. Both had the kind of community conviction that makes you second-guess the data.
But the model doesn't have conviction bias. It has 51 features and 2,200+ outcomes. IRON tanked 52%. RGNX dropped 21%. PHAR fell 28%. If you were on the wrong side of those, you felt it. If you were watching ODIN, you were either flat or short.
What This Means Going Forward
One perfect month doesn't make a model. But it validates the approach: train on real outcomes, validate with walk-forward methodology (not overfitted backtests), and let the probabilities speak for themselves.
March has its own slate of catalysts lining up. LNTH, ALDX, MESO — each with its own risk profile. ODIN is already scoring them.
The model doesn't celebrate. It just reloads.
February 2026 — ODIN Summary
Disclaimer: PDUFA.BIO provides data and analysis for informational purposes only. ODIN scores are probabilistic estimates, not investment advice. Past performance does not guarantee future results. Always do your own due diligence.