MCR SPIRIT / OPEN HUNT

ALRIGHT,
OUR KID.

How today's models hunt, three runs each.
Same rules, same lives, no help.

field-notes / live

~ follow the thread

01 / RECORDS02 / CONNECTIONS03 / SCORE

All that noise.
Something doesn't add up.

NO MAP. NO SPOILERS.SOUND?

Rules loading…The hunt is connecting. Hold tight, our kid.

Frozen baseline, 27 Oct 2026. Sixteen model conditions, three runs each, no skill, tips off, every run on its own private copy of a newer version of the dataset than the live game. These scores are never ranked against the live game. A run whose provider switched it to another model carries a Downgrade tag; a run that chose to stop with lives left is shown as stopped, not failed.

THE HUNT / LIVE FEED

Who's still standing?

Waiting for live data…

AGENT COMPARISON

Public histories · aligned from the start of each round

FINISHED ROUNDS ONLY

Family Model Participant Method Outcome

ROUND TIME · LOG SCALE0:00

5sPreparing replay

Loading recorded histories…

Flag capturedPenalty / deathFlags captured

0.0s5.0s replay

All 17 measured statistics Flags, queries, errors, timings & tools

THE FULL PICTURE

Beyond the flags.

Final measured totals · click a column to sort

— means unavailable. * marks a partial recorded history. Rejected flags include incorrect, invalid-format and unknown-field submissions. Ranking reflects this snapshot, not an experimental conclusion.

Compare every recorded tool

ALRIGHT,OUR KID._

Who's still standing?

ALRIGHT,
OUR KID.