sots-re/findings/control-flow
alex 9fa1ee2600 Z: the first instrumented battle in this campaign costs zero RNG words
A Von Neumann encounter at Gallandro on turn 54 gave the workload the finding
said did not exist. Auto-resolved, with P10 committed before the click.

P10 predicted a non-zero tail cost and was wrong: res_no_battle flipped to 0
for the first time in 55 turns, the fleet was destroyed, and the generator
moved by zero. The bracket residual stayed 0, so combat proper drew nothing
either -- all 22 words were inside ProcessTurn, exactly as on a peaceful turn.

That is the strong form of lane J's static reading, and it means a
reimplementation can model a turn's RNG while modelling nothing about combat.
One auto-resolved encounter against an NPC is not combat in general, and 10.2
says so at length.
2026-09-08 10:14:53 -04:00
..
combat-done-tail.md Z: the unnamed counter at S+0x8 is ModCount, and my players flag was my own error 2026-09-08 10:04:21 -04:00
combat-resolver.md lane J: the combat resolver, read from the instruction stream 2026-09-08 09:37:02 -04:00
inlined-draws.md lane I: the complete inlined-draw inventory, and the seven RNG entry points 2026-09-08 10:03:33 -04:00
tail-rng-ledger.md Z: the first instrumented battle in this campaign costs zero RNG words 2026-09-08 10:14:53 -04:00
turn-driver.md lane K: StrategyServer::OnAllCombatDone_Tail mapped, 36 phases from the instruction stream 2026-09-08 08:47:08 -04:00
turn-spine.md tech effects (196-entry TechId table, OnTechResearched), 8 formula gaps answered, std::string settled 2026-09-07 22:05:49 -04:00