Historian: July 15 gp120 sampler trace

gp120 trace: branch/node alignment moves, heights, rescale, prune/regraft. Count accepted / rejected / no-change separately, best likelihood alone misses most of what's happening.

timestamped trace: Some log records MCMC starting July 15, 2025 at 18:03:10. Input gp120.fa has ten sequences; guide alignment and UPGMA initialization produce a 19-node tree. The log announces 100 samples per node and 1,900 total samples.

gp120.fa STOCKHOLM output: Newick tree and aligned sequences.

model header: one substitution mixture component, expected rate approximately one, insertion/deletion rates both 0.01, and expected insertion/deletion lengths about 2.94118. Preserve this configuration separately from the later SCOP benchmark settings.

trace events: branch-alignment, node-alignment, node-height, rescale and prune/regraft proposals; entries distinguish accepted, rejected and bypassed/no-change moves. Some log-acceptance quantities are positive, so their label should not be interpreted as an already capped log probability without checking implementation semantics.

review questions: compute proposal/acceptance/no-change counts by move class, not only best likelihood. Does likelihood improvement accompany actual topology/alignment exploration? The visible trace is a debugging record, not proof of posterior convergence or a complete evaluation.