ZeroGEX · Scorecard
SPY · Fri, Nov 20, 2026
SPY · 2026-11-20 — 2 Playbook calls. Best: Tape Flow Bias +0.15%. Worst: Squeeze Setup −0.49%. Regime: transition.
Best signal (60m)
Tape Flow Bias
+0.15%
Worst signal (60m)
Squeeze Setup
−0.49%
Closing regime
transition
MSI close
0.5
All Playbook calls · 2
Oldest first. Times are Eastern (UTC−5).
All signals · 60-minute forward return
- Tape Flow Bias+0.15%8 flips · 4 scored · 2W 2L
- Market Pressure+0.14%6 flips · 3 scored · 2W 1L
- Vol Expansion+0.01%102 flips · 84 scored · 41W 43L
- Dealer Delta Pressure−0.01%3 flips · 3 scored · 1W 2L
- Zero Dte Position Imbalance−0.04%10 flips · 10 scored · 5W 5L
- Gex Gradient−0.06%24 flips · 20 scored · 10W 10L
- Vanna Charm Flow−0.10%1 flips · 1 scored · 0W 1L
- Squeeze Setup−0.49%8 flips · 2 scored · 1W 1L
| Signal | Flips | Scored | Wins | Losses | Avg fwd return |
|---|---|---|---|---|---|
| Tape Flow Bias | 8 | 4 | 2 | 2 | +0.15% |
| Market Pressure | 6 | 3 | 2 | 1 | +0.14% |
| Vol Expansion | 102 | 84 | 41 | 43 | +0.01% |
| Dealer Delta Pressure | 3 | 3 | 1 | 2 | −0.01% |
| Zero Dte Position Imbalance | 10 | 10 | 5 | 5 | −0.04% |
| Gex Gradient | 24 | 20 | 10 | 10 | −0.06% |
| Vanna Charm Flow | 1 | 1 | 0 | 1 | −0.10% |
| Squeeze Setup | 8 | 2 | 1 | 1 | −0.49% |
About this scorecard
Daily aggregate of the ZeroGEX engine’s output for SPY. “Playbook calls” counts every non-STAND_DOWN Action Card persisted that day, midnight to midnight Eastern, so it includes calls issued before the open; each one has its own /cards/<id> permalink. “Best/Worst signal” picks the signal whose direction-flip events that day produced the highest/lowest average 60-minute forward return on SPY, with a 2-flip minimum so a single outlier doesn’t crown a signal of the day. “Scored” is how many of a signal’s flips could be graded: the forward price always comes from the same regular session, so a flip inside the last 60 minutes has nothing to grade against and is counted but not scored. A signal that only fires near the bell can read “not scorable” for a whole session - that is an absent measurement, not a flat one. The receipt is immutable once written - the engine cannot retroactively edit a published scorecard.