Tape calibration
Average next-period price return by prior composite bucket. Not advice — sample sizes matter.
Samples
344306
Strong bear
Mild bear
Neutral
Mild bull
Strong bull
Strong bull (≥25)
+0.032%
n=172279
Mild bull (10–25)
+0.001%
n=60239
Neutral
+0.002%
n=57377
Mild bear (−25–−10)
+0.004%
n=23993
Strong bear (≤−25)
+0.008%
n=30418
Uses consecutive points in tape_history (≈5‑min or daily depending on cron cadence).
Extended profile — N days forward × money-flow skew
Same idea, two real differences: a real 3-day forward window instead of just the next tape_history row, and a second axis (call/put volume skew at that same moment) alongside composite — "when the tape looked like X and flow looked like Y, what did price do 3 days later".
Samples
161188
| Composite \ Flow | Put-heavy | Balanced | Call-heavy |
|---|---|---|---|
| Strong bear | -0.24% n=9,038 | -0.11% n=8,737 | — |
| Mild bear | +2.95% n=2,709 | +0.52% n=8,637 | — |
| Neutral | +0.04% n=818 | -0.65% n=28,663 | -1.95% n=796 |
| Mild bull | — | +0.43% n=19,927 | +1.29% n=6,416 |
| Strong bull | — | -0.19% n=49,841 | +0.04% n=25,606 |
Each cell's forward-return average and sample count. Empty (—) cells mean that composite/flow combination hasn't occurred enough in the window to have a 3-day-later data point yet.