PROJECT III — 5 DOWN. 15 REMAIN. OOS-05 IS VERIFIED. Another observation has been frozen into the Project III Out-of-Sample record. This time: Borrow Fee: 1.3907% Fee: −0.1333 percentage points SOUN at alignment: $4.6600 SOUN +60 minutes: $4.7850 +60 Return: +2.6824% Every physical evidence hash matched the manifest. No prediction was added afterward. No threshold was changed. No observation was removed. Something is noticeable — but we're not calling it a signal yet. The first five frozen OOS observations now read: OOS-01: −0.0882 → −0.88% OOS-02: −0.0589 → +0.67% OOS-03: −0.1270 → +0.26% OOS-04: +0.1478 → 0.00% OOS-05: −0.1333 → +2.68% OOS-05 produced the largest positive +60-minute move of the OOS test so far. But there's also an important wrinkle: Candidate 1's discovery relationship expected higher/more-positive Fee to correspond with stronger subsequent performance. OOS-05 went the other direction — a substantially negative Fee accompanied the strongest positive +60 move so far. That's exactly why we aren't celebrating individual observations. Five observations cannot decide the hypothesis. We're looking for whether the relationship survives across 20 completely new observations, not whether we can tell a good story about five. So nothing changes. Same variable. Same +60-minute horizon. Same direction. Same rules. Same evidence standard. At OOS-20, the data gets to speak. If Candidate 1 survives, we document what survived. If it doesn't? It dies. PROJECT III — 5/20 VERIFIED 15 remain. READ BETWEEN THE LINES. The machine is on. Frozen before the outcome. $SOUN $GME $AMC