How it works, and how we know

There is a better way in now.Diagnose runs both of these instruments at once and tells you what the answer changes. These two panels still work, and the numbers are the same.

Point the instruments at your data

The same gated machinery behind every finding here, on whatever you paste in. Verdicts arrive with their honesty warnings attached — flat likelihoods warn instead of quoting, borderline reads say so, and underpowered tests tell you they are. Python users: pip install strataq gives the identical calls offline (strataq.toolkit).

▶ try it · is my time series irreversibly driven?
live float64 solver

Paste a scalar series (numbers separated by spaces, commas or newlines — prices, flows, anything ordered in time; ≥ 300 points for real power). It is phase-embedded and tested against a reversible null with matched persistence — the instrument that found the day-ahead market's diurnal loop (F-0009).

▶ try it · is my system reciprocal? (χ matrix)
live float64 solver

Enter your measured 2×2 cross-response matrix — χ[i][j] = how agent i's action moves when agent j's incentives shift (e.g. cost pass-through between two firms). The defaults are the actual Dominick's pass-through estimates (F-0011). Add standard errors for an uncertainty-aware verdict.

standard errors (optional):