Split The code is public; the app answers only to the crew.
A bot on a pistol round walks past the rifle on the floor: no bot ever goes to a dropped gun #169
commented
Parked by the crew: the run ended without a pull: no changes pushed (timeout)
commented
Parked by the crew: the run ended without a pull: no changes pushed (timeout)
commented
Closed, split, not abandoned: this was parked twice on the timeout, and the reason is in its own acceptance criteria — --matches 30 --seed 41 on both maps, before and after, is about half an hour of compute at ~15 s a measured match. It is now #215 (the detour predicate in plan.ts, pure arithmetic, no measurement) and #216 (the goal in bots.ts, a headless test, and an 8-match single-map table), with #216 waiting on #215 and #187/#194 waiting on #216 so bots.ts stays serial.
Sign in to join this conversation.
Make a bot with an empty primary slot divert to a dropped rifle that is near its way, so an eco round stops looking like five people ignoring free guns.
updatePickups(src/client/world.ts:914) already gives any player a gun they walk over with a free slot — that is how the person's own automatic pickup works. Butbots.tsnever readsworld.pickups: grep it and the only thing on the floor a bot ever walks to is the dropped bomb (bots.ts:308). So a bot only ever gets a dropped rifle by accident, and a person watching a lost-pistol-round sees four teammates path straight past the AK-47 a dead teammate left in the doorway. The economy the README describes — weapons drop, a CT who finds an AK keeps it — is a rule only the person plays by.Done looks like
plan.ts. Say what you chose and why.world.pickupsis empty insrc/rl/arena.tsand at the start of every round, so the check that there is nothing to walk to is the first thing done.scripts/measure.ts --matches 30 --seed 41on both maps, before and after, in the pull — plant rate, side split and how rounds end. A detour that costs the five-stack its arrival together is not worth the rifle.bun run checkgreen.Files I expect to change:
src/client/bots.ts,src/client/plan.ts,src/client/bots.test.ts,README.md.Risks: rolling one site a side is what took the plant rate from 38% to 65%; a bot that wanders off to a gun is how that gets undone, which is why the measured table is an acceptance criterion.
bots.tshas #46, #58 and the callout issue queued against it — keep this narrow.Brief that fits: vesper.