PROJECT III — 40% THROUGH THE TEST. THE SIGNAL IS STILL ON TRIAL. Friday closes with two more observations locked into the record. First, the SIP window for OOS-07 cleared. Its borrow observation had already been frozen: Fee: −0.0020 pp When the SIP evidence became available, Project III measured the already-locked +60 window: +60 return: −0.4468% The SIP page, combined evidence, derived result, and manifest record were physically verified. OOS-07: VERIFIED ✓ Then came today's observation. OOS-08: VERIFIED ✓ Fee: −0.1323 pp +60 return: +0.4384% Again, the borrow evidence was captured and physically verified before the +60 outcome was recorded. Then the SIP evidence arrived. It matched. Another observation entered the record exactly as it happened. 8/20 COMPLETE. 12 REMAIN. 40% THROUGH THE TEST. SO… ARE WE SEEING ANYTHING? We're seeing something important—but it isn't a victory announcement. The first 8 out-of-sample observations are mixed. Some behave consistently with the relationship discovered during Phase II. Others don't. And that's exactly why Project III doesn't stop when the results look good or rewrite itself when they don't. Candidate 1 has not earned confirmation yet. It gets all 20 observations. No cherry-picking. No changing the +60 window. No convenient new filters. No deleting observations. No moving the goalposts. If the relationship survives, we'll document exactly what survived. If it dies, the record will show exactly where it died. Either result teaches the machine something. For now: 8 DOWN. 12 TO GO. 40% COMPLETE. The weekend doesn't count. The machine waits for the next valid market session. Monday → potential OOS-09. $SOUN remains the sandbox. $GME & $AMC are the mission. The market is the end game. READ BETWEEN THE LINES. The signal still has to earn its existence.