Orderflow Bias Engine

Built a small orderflow bias engine to measure true price efficiency-not signals, not direction.
TradingView doesn’t give true tick-by-tick bid/ask.
Most tools either ignore that or fake footprint logic.
I didn’t!
Instead of absolute precision, I focused on relative effort vs result.
Instead of pretending otherwise, I leaned into relative effort vs result.
Lower TF volume delta is aggregated and mapped candle-by-candle against price displacement.
Initiative vs absorption isn’t inferred - it’s classified.
The core question isn’t who traded more -
it’s how much net aggression was required to move price one unit.
High effort, poor displacement - absorption / inventory build
Low effort, clean displacement - thinning liquidity / expansion
Features are intentionally minimal:
Candle-level classification (initiative buy/sell, buyer/seller absorption)
Rolling control score (who’s pressing vs who’s absorbing)
Bias state, not signal state
Efficiency read to detect compression vs expansion
It’s auction efficiency under imperfect data.
Structure defines location.
This just tells me whether price is being dragged or allowed to move.
If that distinction matters to you, you already know how to read it.