OBSERVATION 01
Fluorescence fell by 62%.
VEHICLEV-17−62%
OMC / 01
A SCIENTIFIC REASONING GAME
THREE PLAUSIBLEANSWERS.ONE EXPERIMENTTHAT MATTERS.
Every test costs time, samples, and certainty. You do not win by collecting more data. You win by choosing the control that makes the wrong mechanisms impossible.
1 playable case · 10–12 minutes · No sign-in
Y / SIGNAL
X / TIME
LIVE SIGNAL
H1
Direct inhibitionCATALYSIS CHANGED
ALIVEH2
Enzyme lossAMOUNT CHANGED
ALIVEH3
Optical interferenceREADOUT CHANGED
ALIVEEXPERIMENTAL BUDGET100 UNITS LEFT
TIMING CONTROL · COST 15Add compound after the reactionPREVIEW SPLIT ↗HOVER THE EXPERIMENT TO PREVIEW ITS PREDICTION SPLIT
02 THE METHOD
MORE DATA CAN STILL
LEAVE YOU WRONG.
The game rewards experiments that separate hypotheses—not experiments that merely reproduce the same observation.
01◎
OBSERVE
Read the signal without pretending it already explains the mechanism.
02⌁
PREDICT
Commit to what each hypothesis should produce before spending the budget.
03×
ELIMINATE
Use the result to kill explanations that can no longer survive.
03 AFTER THE VERDICT
YOUR REASONING
LEAVES A FINGERPRINT.
At the end of each case, see where you chased confirmation, repeated low-information tests, calibrated uncertainty, or found the decisive control.
FALSIFICATION EFFICIENCY CALIBRATION
REASONING FINGERPRINT / SAMPLETHE FALSIFIER
91/100FALSIFICATION FIRST84%
REDUNDANCY RATE12%
EVIDENCE EFFICIENCY2.8b/¢
CALIBRATION GAP04pts
THE NEXT MOVE IS YOURS
DON'T ASK AI
FOR THE ANSWER.
Ask what evidence would prove it wrong.
OPEN THE FIRST CASE