How a score gets to +38.4, with nothing left out.
Every number on FlowDesk comes from this page. If something here changes, it’s posted to the changelog first, with before-and-after results.
The composite
| Part | Definition | AAPL now | × weight | points |
|---|---|---|---|---|
| V · volume bias | (calls − puts) ÷ all contracts traded today, ×100 640,555 calls · 291,340 puts | +37.47 | 0.62 | +23.23 |
| OI · open-interest bias | (call OI − put OI) ÷ all open interest, ×100 94,768 call OI · 40,621 put OI | +39.99 | 0.38 | +15.20 |
| Composite | +38.43 | |||
Evidence is not conviction
Confidence (the evidence meter)
How much there is to stand on: contract count, liquidity and how stable the read has been. Five cells, one per 20 points.
Quality
How clean the data is: fresh chain, both sides quoted, no stale fallback price. Below 50, the verdict bar is hatched.
The verdict words
| Composite | Verdict | Shown as |
|---|---|---|
| ≥ +25 | Strongly bullish | |
| +10 … +25 | Mildly bullish | |
| −10 … +10 | Neutral | |
| −25 … −10 | Mildly bearish | |
| ≤ −25 | Strongly bearish |
Momentum and change
Δ1d is today’s composite minus the previous session’s (tape_history). The dashed orange tick on every bar is yesterday. A row turns orange in your book when |Δ| ≥ 5.
What a hit is
A call is checked at 1h, 4h, 1d and 3d. Hit: price moved the way the composite pointed. Flat: moved less than 0.05% either way, shown as flat, never as a win. Everything else is a miss. Horizons with nothing checked are shown as empty.
What this does not do
It doesn’t forecast earnings, treat a headline as truth, or know your position size. The tape transformer is research: it has never been used to score or rank anything.