Encevra Research Round 52 (MCB-52): The Second Reader or the Second Call? Matched Call Baselines for Agreement Routing, Validation and Recovery

Round 52 of the Theory of Collaborative Intelligence program asks whether a gain from a second or third model call comes from a different reader looking or only from another call being made. Every composition (agreement routing, validation, preservation with independent recovery, majority and unanimity) is compared with baselines that use the same number of calls: the same reader resampled, the same reader validating its own read, and one call with room to argue each verdict. Web page: https://research.encevra.com/round-52 Materials: all 36 Round 49 packets replayed verbatim with the Round 49 prompt (a direct replication of the Round 49 near misses), 16 authored raw table claims with thresholds within 2 percent of the correct total, and 24 real SEC merger agreement exhibits, none used before, each with one termination fee claim. Structurally, over 81 states of gold and three verdicts, unanimity, recovery and majority are unsafe in 4, 20 and 28 states, strictly nested. Live, Witness D reproduced its Round 49 verdicts on 36 of 36 items, including a wrong FAIL on all 9 near misses; two resamples and a self validation gave the same FAIL on 9 of 9, and self validation changed none of 75 reads. At two calls, routing Witnesses C and D released 6 unsafe verdicts against 5 for Witness C with its own resample. Witness C as validator moved 2 of 7 wrong reads to HOLD and turned one correct HOLD into PASS. Recovery released 5 unsafe verdicts against 0 for three reader unanimity. One deliberate Witness D call released 2 unsafe verdicts on 69 usable items. Tables and SEC claims produced no decisive error from any reader. The record contains the manuscript and a reproducibility archive with the fixed run plan, exact prompts, verbatim transcripts, Round 49 reference calls, materials, SEC windows with accessions and links, deterministic checks, replay analysis, verifier, completion disclosure and limitations. Twenty one of 797 experimental outputs were unusable and are disclosed. Calls are matched by count, not by reader capability. No universal theory is claimed. The program continues through Round 55.