metal · dukascopy · prices in USD/oz · horizons 4h / 1d / 3d / 5d · session analysis in America/New_York
From a random entry at a bar's close, how often the target arrived before the stop, within the horizon — going long, and going short. The number to read is the gap between the two. Its null is exactly zero under sign symmetry — if a path and its mirror image are equally likely once drift is removed — and that holds at any horizon. A hit rate on its own is mostly arithmetic. Note what the null does not say: a skewed series with no drift can break it without either direction being profitable. See methodology.
Entries are every bar, unconditionally. There is no entry rule and no signal: this is the unconditional first-passage structure of the price path, not a backtest. Consecutive entries overlap heavily, so every cell reports an effective sample size alongside its row count — see methodology.
118,574 bars · 2021-07-27T19:00:00+00:00 to 2026-08-02T23:45:00+00:00
Over this window the price drifted +18.7% a year. Each cell below shows how much of its long/short gap that drift accounts for; only the remainder is tested.
7,751 non-overlapping entries · 10,993 coverage-span equivalents · median 16 bars per window
At this horizon: 0.25% ≈ 0.25σ · 0.5% ≈ 0.51σ · 1% ≈ 1.01σ · 2% ≈ 2.01σ · 3% ≈ 3.01σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 44.4% | 49.7% | 5.9% | -3.6 pp | -5.0 to -2.4 pp | 17600 |
| +0.25% / −0.5% | 57.2% | 28.1% | 14.8% | -3.8 pp | -5.4 to -2.5 pp | 12541 |
| +0.25% / −1% | 63.1% | 11.0% | 25.9% | -3.2 pp | -4.3 to -2.1 pp | 6724 |
| +0.25% / −2% | 64.8% | 2.5% | 32.7% | -1.1 pp | -1.6 to -0.6 pp | 4360 |
| +0.25% / −3% | 65.0% | 0.8% | 34.2% | -0.5 pp | -0.7 to -0.2 pp | 4505 |
| +0.5% / −0.25% | 24.7% | 60.6% | 14.7% | -3.7 pp | -5.2 to -2.4 pp | 12197 |
| +0.5% / −0.5% | 33.1% | 36.7% | 30.2% | -5.7 pp | -7.7 to -3.9 pp | 9484 |
| +0.5% / −1% | 37.9% | 15.0% | 47.1% | -6.0 pp | -8.0 to -4.1 pp | 4800 |
| +0.5% / −2% | 39.5% | 3.6% | 57.0% | -2.5 pp | -3.6 to -1.5 pp | 2303 |
| +0.5% / −3% | 39.7% | 1.2% | 59.1% | -1.2 pp | -1.8 to -0.7 pp | 2097 |
| +1% / −0.25% | 8.8% | 65.5% | 25.7% | -3.1 pp | -4.1 to -2.0 pp | 8547 |
| +1% / −0.5% | 12.3% | 41.0% | 46.7% | -6.0 pp | -8.0 to -4.0 pp | 5637 |
| +1% / −1% | 14.7% | 17.4% | 67.9% | -9.4 pp | -12.6 to -6.4 pp | 3880 |
| +1% / −2% | 15.7% | 4.3% | 80.0% | -6.5 pp | -9.4 to -3.7 pp | 1206 |
| +1% / −3% | 15.8% | 1.5% | 82.6% | -4.1 pp | -5.9 to -2.3 pp | 806 |
| +2% / −0.25% | 1.9% | 66.5% | 31.7% | -1.1 pp | -1.6 to -0.6 pp | 6307 |
| +2% / −0.5% | 2.7% | 42.1% | 55.2% | -2.5 pp | -3.6 to -1.5 pp | 3032 |
| +2% / −1% | 3.4% | 18.1% | 78.5% | -6.5 pp | -9.3 to -3.6 pp | 1634 |
| +2% / −2% | 3.7% | 4.7% | 91.6% | -11.8 pp | -18.3 to -5.2 pp | 885 |
| +2% / −3% | 3.8% | 1.7% | 94.4% | -12.9 pp | -19.0 to -6.8 pp | 446 |
| +3% / −0.25% | 0.5% | 66.6% | 32.9% | -0.5 pp | -0.7 to -0.2 pp | 7349 |
| +3% / −0.5% | 0.8% | 42.2% | 57.0% | -1.2 pp | -1.8 to -0.6 pp | 3270 |
| +3% / −1% | 1.0% | 18.3% | 80.8% | -4.0 pp | -5.8 to -2.3 pp | 1366 |
| +3% / −2% | 1.1% | 4.7% | 94.2% | -12.9 pp | -19.0 to -6.7 pp | 534 |
| +3% / −3% | 1.1% | 1.8% | 97.1% | -22.6 pp | -33.6 to -13.2 pp | 368 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 44.4% | 46.7% | 47.4% | 5.9% | 47.2% | 49.7% | 94% | 94% | -2.4 pp | -1.2 pp | -3.6 pp | halves agree | 2.95% | 2.95% | 117,662 |
| +0.25% / −0.5% | 57.2% | 59.5% | 25.9% | 14.7% | 67.1% | 69.7% | 85% | 85% | -2.6 pp | -1.2 pp | -3.8 pp | halves agree | 1.22% | 1.15% | 117,662 |
| +0.25% / −1% | 63.1% | 65.2% | 9.1% | 25.7% | 85.2% | 87.7% | 74% | 74% | -2.5 pp | -0.6 pp | -3.2 pp | halves agree | 0.38% | 0.30% | 117,662 |
| +0.25% / −2% | 64.8% | 66.4% | 1.9% | 31.7% | 96.2% | 97.2% | 67% | 68% | -1.0 pp | -0.2 pp | -1.1 pp | halves agree | 0.05% | 0.03% | 117,662 |
| +0.25% / −3% | 65.0% | 66.6% | 0.6% | 32.9% | 98.8% | 99.2% | 66% | 67% | -0.4 pp | -0.1 pp | -0.5 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +0.5% / −0.25% | 24.7% | 26.9% | 58.4% | 14.8% | 29.0% | 31.5% | 85% | 85% | -2.5 pp | -1.2 pp | -3.7 pp | halves agree | 1.15% | 1.22% | 117,662 |
| +0.5% / −0.5% | 33.1% | 36.0% | 33.8% | 30.2% | 47.4% | 51.6% | 70% | 70% | -4.2 pp | -1.4 pp | -5.7 pp | halves agree | 0.68% | 0.68% | 117,662 |
| +0.5% / −1% | 37.9% | 40.8% | 12.5% | 46.7% | 71.7% | 76.6% | 53% | 53% | -4.9 pp | -1.1 pp | -6.0 pp | halves agree | 0.22% | 0.22% | 117,662 |
| +0.5% / −2% | 39.5% | 42.1% | 2.7% | 55.2% | 91.7% | 93.9% | 43% | 45% | -2.2 pp | -0.3 pp | -2.5 pp | halves agree | 0.05% | 0.03% | 117,662 |
| +0.5% / −3% | 39.7% | 42.2% | 0.8% | 57.0% | 97.0% | 98.2% | 41% | 43% | -1.1 pp | -0.1 pp | -1.2 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +1% / −0.25% | 8.8% | 10.6% | 63.5% | 25.9% | 11.9% | 14.3% | 74% | 74% | -2.4 pp | -0.6 pp | -3.1 pp | halves agree | 0.30% | 0.38% | 117,662 |
| +1% / −0.5% | 12.3% | 14.8% | 38.1% | 47.1% | 23.0% | 27.9% | 53% | 53% | -4.9 pp | -1.0 pp | -6.0 pp | halves agree | 0.22% | 0.22% | 117,662 |
| +1% / −1% | 14.7% | 17.3% | 14.8% | 67.9% | 45.8% | 53.9% | 32% | 32% | -8.1 pp | -1.3 pp | -9.4 pp | halves agree | 0.09% | 0.09% | 117,662 |
| +1% / −2% | 15.7% | 18.1% | 3.4% | 78.5% | 78.4% | 84.2% | 20% | 22% | -5.8 pp | -0.7 pp | -6.5 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +1% / −3% | 15.8% | 18.2% | 1.0% | 80.8% | 91.1% | 94.9% | 17% | 19% | -3.8 pp | -0.3 pp | -4.1 pp | halves agree | 0.01% | 0.00% | 117,662 |
| +2% / −0.25% | 1.9% | 2.5% | 64.9% | 32.7% | 2.7% | 3.7% | 68% | 67% | -0.9 pp | -0.2 pp | -1.1 pp | halves agree | 0.03% | 0.05% | 117,662 |
| +2% / −0.5% | 2.7% | 3.5% | 39.5% | 57.0% | 6.0% | 8.2% | 45% | 43% | -2.1 pp | -0.3 pp | -2.5 pp | halves agree | 0.03% | 0.05% | 117,662 |
| +2% / −1% | 3.4% | 4.3% | 15.7% | 80.0% | 15.8% | 21.5% | 22% | 20% | -5.7 pp | -0.7 pp | -6.5 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +2% / −2% | 3.7% | 4.7% | 3.7% | 91.6% | 44.6% | 55.4% | 8% | 8% | -10.8 pp | -1.0 pp | -11.8 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +2% / −3% | 3.8% | 4.7% | 1.1% | 94.2% | 68.7% | 80.9% | 6% | 6% | -12.2 pp | -0.7 pp | -12.9 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +3% / −0.25% | 0.5% | 0.8% | 65.0% | 34.2% | 0.8% | 1.2% | 67% | 66% | -0.4 pp | -0.1 pp | -0.5 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −0.5% | 0.8% | 1.2% | 39.7% | 59.1% | 1.8% | 2.9% | 43% | 41% | -1.1 pp | -0.1 pp | -1.2 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −1% | 1.0% | 1.5% | 15.8% | 82.6% | 5.1% | 8.9% | 19% | 17% | -3.7 pp | -0.3 pp | -4.0 pp | halves agree | 0.00% | 0.01% | 117,662 |
| +3% / −2% | 1.1% | 1.7% | 3.8% | 94.4% | 19.1% | 31.3% | 6% | 6% | -12.2 pp | -0.7 pp | -12.9 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +3% / −3% | 1.1% | 1.8% | 1.1% | 97.1% | 39.1% | 60.9% | 3% | 3% | -21.7 pp | -0.9 pp | -22.6 pp | halves agree | 0.00% | 0.00% | 117,662 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
25 of 25 Benjamini–Hochberg survivors in the 4h horizon also agree across both halves of history.
Sampling check: re-running this horizon on 7,751 strictly non-overlapping entries (rather than 117,662 every-bar entries) moves the asymmetry by a median of 0.8 pp and at most 4.2 pp.
1,296 non-overlapping entries · 1,832 coverage-span equivalents · median 92 bars per window
At this horizon: 0.25% ≈ 0.11σ · 0.5% ≈ 0.21σ · 1% ≈ 0.42σ · 2% ≈ 0.84σ · 3% ≈ 1.25σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 46.6% | 51.9% | 1.4% | -3.7 pp | -5.1 to -2.4 pp | 16823 |
| +0.25% / −0.5% | 63.8% | 33.9% | 2.2% | -3.9 pp | -5.5 to -2.5 pp | 11252 |
| +0.25% / −1% | 76.9% | 19.7% | 3.4% | -3.7 pp | -5.3 to -2.1 pp | 6923 |
| +0.25% / −2% | 83.9% | 8.5% | 7.6% | -1.6 pp | -2.9 to -0.3 pp | 4409 |
| +0.25% / −3% | 85.0% | 3.9% | 11.1% | -0.8 pp | -1.7 to +0.1 pp | 4276 |
| +0.5% / −0.25% | 30.4% | 67.5% | 2.1% | -3.8 pp | -5.4 to -2.4 pp | 11861 |
| +0.5% / −0.5% | 45.9% | 50.6% | 3.4% | -5.8 pp | -8.3 to -3.7 pp | 7077 |
| +0.5% / −1% | 61.5% | 32.2% | 6.3% | -5.8 pp | -8.4 to -3.3 pp | 4181 |
| +0.5% / −2% | 70.5% | 14.0% | 15.5% | -2.8 pp | -5.1 to -0.5 pp | 2534 |
| +0.5% / −3% | 72.0% | 6.4% | 21.6% | -1.7 pp | -3.4 to +0.0 pp | 2170 |
| +1% / −0.25% | 17.0% | 80.0% | 3.0% | -3.6 pp | -5.2 to -1.9 pp | 7616 |
| +1% / −0.5% | 28.4% | 65.8% | 5.8% | -5.8 pp | -8.4 to -3.2 pp | 4426 |
| +1% / −1% | 41.2% | 44.8% | 14.0% | -6.6 pp | -10.6 to -2.7 pp | 2544 |
| +1% / −2% | 48.9% | 19.9% | 31.2% | -4.2 pp | -8.0 to -0.1 pp | 1562 |
| +1% / −3% | 50.3% | 9.0% | 40.7% | -3.0 pp | -6.0 to +0.1 pp | 1192 |
| +2% / −0.25% | 7.8% | 85.7% | 6.6% | -1.6 pp | -2.9 to -0.4 pp | 5475 |
| +2% / −0.5% | 13.0% | 73.2% | 13.8% | -2.8 pp | -5.1 to -0.5 pp | 3073 |
| +2% / −1% | 18.9% | 51.2% | 29.8% | -4.2 pp | -8.0 to -0.1 pp | 1783 |
| +2% / −2% | 22.8% | 23.2% | 54.0% | -4.2 pp | -10.5 to +2.4 pp | 915 |
| +2% / −3% | 23.6% | 10.8% | 65.6% | -4.5 pp | -10.6 to +2.3 pp | 604 |
| +3% / −0.25% | 3.6% | 86.5% | 9.9% | -0.8 pp | -1.7 to +0.1 pp | 4935 |
| +3% / −0.5% | 5.9% | 74.2% | 19.8% | -1.6 pp | -3.3 to +0.0 pp | 2778 |
| +3% / −1% | 8.4% | 52.4% | 39.2% | -3.0 pp | -6.0 to +0.1 pp | 1516 |
| +3% / −2% | 10.2% | 24.2% | 65.6% | -4.5 pp | -10.6 to +2.3 pp | 710 |
| +3% / −3% | 10.8% | 11.4% | 77.8% | -6.6 pp | -15.3 to +2.9 pp | 431 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 46.6% | 49.0% | 49.6% | 1.4% | 47.3% | 49.7% | 99% | 99% | -2.4 pp | -1.3 pp | -3.7 pp | halves agree | 2.97% | 2.97% | 117,662 |
| +0.25% / −0.5% | 63.8% | 66.3% | 31.6% | 2.1% | 65.3% | 67.7% | 98% | 98% | -2.4 pp | -1.5 pp | -3.9 pp | halves agree | 1.27% | 1.20% | 117,662 |
| +0.25% / −1% | 76.9% | 79.5% | 17.5% | 3.0% | 79.6% | 82.0% | 97% | 97% | -2.3 pp | -1.3 pp | -3.7 pp | halves agree | 0.49% | 0.45% | 117,662 |
| +0.25% / −2% | 83.9% | 85.6% | 7.8% | 6.6% | 90.8% | 91.6% | 92% | 93% | -0.8 pp | -0.8 pp | -1.6 pp | halves agree | 0.06% | 0.06% | 117,662 |
| +0.25% / −3% | 85.0% | 86.5% | 3.6% | 9.9% | 95.7% | 96.0% | 89% | 90% | -0.3 pp | -0.5 pp | -0.8 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +0.5% / −0.25% | 30.4% | 32.7% | 65.1% | 2.2% | 31.0% | 33.4% | 98% | 98% | -2.4 pp | -1.5 pp | -3.8 pp | halves agree | 1.20% | 1.27% | 117,662 |
| +0.5% / −0.5% | 45.9% | 49.8% | 46.8% | 3.4% | 47.6% | 51.5% | 97% | 97% | -4.0 pp | -1.9 pp | -5.8 pp | halves agree | 0.86% | 0.86% | 117,662 |
| +0.5% / −1% | 61.5% | 65.4% | 28.8% | 5.8% | 65.7% | 69.4% | 94% | 94% | -3.8 pp | -2.1 pp | -5.8 pp | halves agree | 0.37% | 0.36% | 117,662 |
| +0.5% / −2% | 70.5% | 73.1% | 13.1% | 13.8% | 83.4% | 84.8% | 85% | 86% | -1.4 pp | -1.5 pp | -2.8 pp | halves agree | 0.07% | 0.07% | 117,662 |
| +0.5% / −3% | 72.0% | 74.2% | 5.9% | 19.8% | 91.9% | 92.6% | 78% | 80% | -0.7 pp | -0.9 pp | -1.7 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +1% / −0.25% | 17.0% | 19.2% | 77.4% | 3.4% | 17.6% | 19.9% | 97% | 97% | -2.3 pp | -1.3 pp | -3.6 pp | halves agree | 0.45% | 0.49% | 117,662 |
| +1% / −0.5% | 28.4% | 31.8% | 61.9% | 6.3% | 30.2% | 33.9% | 94% | 94% | -3.8 pp | -2.1 pp | -5.8 pp | halves agree | 0.36% | 0.37% | 117,662 |
| +1% / −1% | 41.2% | 44.6% | 41.4% | 14.0% | 47.9% | 51.9% | 86% | 86% | -4.0 pp | -2.6 pp | -6.6 pp | halves agree | 0.20% | 0.20% | 117,662 |
| +1% / −2% | 48.9% | 51.2% | 19.0% | 29.8% | 71.1% | 73.0% | 69% | 70% | -1.8 pp | -2.4 pp | -4.2 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +1% / −3% | 50.3% | 52.4% | 8.4% | 39.2% | 84.8% | 86.2% | 59% | 61% | -1.4 pp | -1.6 pp | -3.0 pp | halves agree | 0.01% | 0.00% | 117,662 |
| +2% / −0.25% | 7.8% | 8.4% | 84.0% | 7.6% | 8.3% | 9.1% | 93% | 92% | -0.8 pp | -0.8 pp | -1.6 pp | halves agree | 0.06% | 0.06% | 117,662 |
| +2% / −0.5% | 13.0% | 13.9% | 70.6% | 15.5% | 15.1% | 16.5% | 86% | 85% | -1.4 pp | -1.5 pp | -2.8 pp | halves agree | 0.07% | 0.07% | 117,662 |
| +2% / −1% | 18.9% | 19.8% | 49.0% | 31.2% | 27.0% | 28.8% | 70% | 69% | -1.8 pp | -2.4 pp | -4.2 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +2% / −2% | 22.8% | 23.2% | 22.8% | 54.0% | 49.5% | 50.5% | 46% | 46% | -1.0 pp | -3.2 pp | -4.2 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +2% / −3% | 23.6% | 24.2% | 10.2% | 65.6% | 68.6% | 70.3% | 34% | 34% | -1.7 pp | -2.9 pp | -4.5 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +3% / −0.25% | 3.6% | 3.8% | 85.1% | 11.1% | 4.0% | 4.3% | 90% | 89% | -0.3 pp | -0.5 pp | -0.8 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −0.5% | 5.9% | 6.4% | 72.0% | 21.6% | 7.4% | 8.1% | 80% | 78% | -0.7 pp | -0.9 pp | -1.6 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −1% | 8.4% | 9.0% | 50.3% | 40.7% | 13.8% | 15.2% | 61% | 59% | -1.4 pp | -1.6 pp | -3.0 pp | halves agree | 0.00% | 0.01% | 117,662 |
| +3% / −2% | 10.2% | 10.8% | 23.6% | 65.6% | 29.7% | 31.4% | 34% | 34% | -1.7 pp | -2.9 pp | -4.5 pp | halves agree | 0.00% | 0.00% | 117,662 |
| +3% / −3% | 10.8% | 11.4% | 10.8% | 77.8% | 48.5% | 51.5% | 22% | 22% | -3.0 pp | -3.5 pp | -6.6 pp | halves agree | 0.00% | 0.00% | 117,662 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
13 of 13 Benjamini–Hochberg survivors in the 1d horizon also agree across both halves of history.
Sampling check: re-running this horizon on 1,296 strictly non-overlapping entries (rather than 117,662 every-bar entries) moves the asymmetry by a median of 0.8 pp and at most 5.6 pp.
525 non-overlapping entries · 610 coverage-span equivalents · median 224 bars per window
At this horizon: 0.25% ≈ 0.07σ · 0.5% ≈ 0.14σ · 1% ≈ 0.27σ · 2% ≈ 0.54σ · 3% ≈ 0.80σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 47.3% | 52.6% | 0.1% | -3.5 pp | -5.0 to -2.3 pp | 16738 |
| +0.25% / −0.5% | 65.2% | 34.8% | 0.1% | -3.8 pp | -5.5 to -2.4 pp | 11191 |
| +0.25% / −1% | 79.1% | 20.7% | 0.2% | -3.6 pp | -5.3 to -2.1 pp | 6941 |
| +0.25% / −2% | 88.1% | 10.7% | 1.2% | -2.0 pp | -3.5 to -0.6 pp | 4240 |
| +0.25% / −3% | 90.6% | 6.5% | 3.0% | -1.3 pp | -2.5 to -0.0 pp | 3787 |
| +0.5% / −0.25% | 31.2% | 68.7% | 0.1% | -3.7 pp | -5.4 to -2.3 pp | 12109 |
| +0.5% / −0.5% | 47.6% | 52.3% | 0.2% | -5.7 pp | -8.2 to -3.6 pp | 7252 |
| +0.5% / −1% | 64.7% | 34.6% | 0.6% | -6.0 pp | -8.8 to -3.5 pp | 4044 |
| +0.5% / −2% | 77.9% | 18.9% | 3.2% | -3.5 pp | -6.1 to -0.9 pp | 2310 |
| +0.5% / −3% | 81.5% | 11.4% | 7.1% | -2.3 pp | -4.8 to -0.0 pp | 1965 |
| +1% / −0.25% | 18.0% | 81.8% | 0.2% | -3.6 pp | -5.3 to -2.0 pp | 7973 |
| +1% / −0.5% | 30.5% | 68.9% | 0.6% | -5.9 pp | -8.7 to -3.5 pp | 4524 |
| +1% / −1% | 46.7% | 51.0% | 2.3% | -6.9 pp | -11.0 to -3.0 pp | 2386 |
| +1% / −2% | 61.6% | 29.4% | 9.0% | -4.0 pp | -8.5 to +0.4 pp | 1359 |
| +1% / −3% | 66.2% | 17.6% | 16.2% | -2.6 pp | -6.8 to +1.6 pp | 1127 |
| +2% / −0.25% | 9.8% | 89.2% | 1.0% | -2.0 pp | -3.5 to -0.6 pp | 5473 |
| +2% / −0.5% | 17.5% | 79.8% | 2.6% | -3.4 pp | -6.1 to -0.8 pp | 2777 |
| +2% / −1% | 28.8% | 63.1% | 8.0% | -4.0 pp | -8.5 to +0.4 pp | 1467 |
| +2% / −2% | 39.5% | 38.2% | 22.3% | -2.3 pp | -9.2 to +4.5 pp | 787 |
| +2% / −3% | 43.1% | 23.1% | 33.7% | -1.9 pp | -9.2 to +5.3 pp | 593 |
| +3% / −0.25% | 6.2% | 91.0% | 2.8% | -1.3 pp | -2.5 to -0.0 pp | 4615 |
| +3% / −0.5% | 10.9% | 82.7% | 6.4% | -2.3 pp | -4.8 to -0.0 pp | 2522 |
| +3% / −1% | 17.8% | 66.8% | 15.4% | -2.6 pp | -6.8 to +1.6 pp | 1287 |
| +3% / −2% | 24.5% | 41.0% | 34.5% | -1.9 pp | -9.2 to +5.3 pp | 633 |
| +3% / −3% | 26.6% | 25.1% | 48.4% | -2.4 pp | -12.1 to +6.9 pp | 436 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 47.3% | 49.6% | 50.4% | 0.1% | 47.4% | 49.6% | 100% | 100% | -2.2 pp | -1.3 pp | -3.5 pp | halves agree | 3.03% | 3.03% | 117,662 |
| +0.25% / −0.5% | 65.2% | 67.4% | 32.5% | 0.1% | 65.2% | 67.5% | 100% | 100% | -2.3 pp | -1.5 pp | -3.8 pp | halves agree | 1.32% | 1.25% | 117,662 |
| +0.25% / −1% | 79.1% | 81.3% | 18.4% | 0.2% | 79.3% | 81.5% | 100% | 100% | -2.3 pp | -1.4 pp | -3.6 pp | halves agree | 0.50% | 0.47% | 117,662 |
| +0.25% / −2% | 88.1% | 89.1% | 9.8% | 1.0% | 89.2% | 90.1% | 99% | 99% | -0.9 pp | -1.1 pp | -2.0 pp | halves agree | 0.07% | 0.07% | 117,662 |
| +0.25% / −3% | 90.6% | 91.0% | 6.2% | 2.8% | 93.3% | 93.7% | 97% | 97% | -0.3 pp | -1.0 pp | -1.3 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +0.5% / −0.25% | 31.2% | 33.4% | 66.5% | 0.1% | 31.3% | 33.5% | 100% | 100% | -2.2 pp | -1.5 pp | -3.7 pp | halves agree | 1.25% | 1.32% | 117,662 |
| +0.5% / −0.5% | 47.6% | 51.4% | 48.4% | 0.2% | 47.6% | 51.5% | 100% | 100% | -3.8 pp | -1.9 pp | -5.7 pp | halves agree | 0.88% | 0.88% | 117,662 |
| +0.5% / −1% | 64.7% | 68.5% | 30.9% | 0.6% | 65.1% | 68.9% | 99% | 99% | -3.8 pp | -2.2 pp | -6.0 pp | halves agree | 0.38% | 0.39% | 117,662 |
| +0.5% / −2% | 77.9% | 79.8% | 17.6% | 2.6% | 80.4% | 81.9% | 97% | 97% | -1.5 pp | -2.0 pp | -3.5 pp | halves agree | 0.10% | 0.08% | 117,662 |
| +0.5% / −3% | 81.5% | 82.7% | 10.9% | 6.4% | 87.7% | 88.4% | 93% | 94% | -0.6 pp | -1.7 pp | -2.3 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +1% / −0.25% | 18.0% | 20.2% | 79.6% | 0.2% | 18.0% | 20.2% | 100% | 100% | -2.2 pp | -1.3 pp | -3.6 pp | halves agree | 0.47% | 0.50% | 117,662 |
| +1% / −0.5% | 30.5% | 34.3% | 65.1% | 0.6% | 30.7% | 34.5% | 99% | 99% | -3.8 pp | -2.1 pp | -5.9 pp | halves agree | 0.39% | 0.38% | 117,662 |
| +1% / −1% | 46.7% | 50.8% | 46.9% | 2.3% | 47.8% | 52.0% | 98% | 98% | -4.2 pp | -2.7 pp | -6.9 pp | halves agree | 0.22% | 0.22% | 117,662 |
| +1% / −2% | 61.6% | 63.1% | 28.8% | 8.0% | 67.7% | 68.6% | 91% | 92% | -1.0 pp | -3.1 pp | -4.0 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +1% / −3% | 66.2% | 66.8% | 17.8% | 15.4% | 79.0% | 78.9% | 84% | 85% | +0.1 pp | -2.7 pp | -2.6 pp | HALVES DISAGREE | 0.01% | 0.00% | 117,662 |
| +2% / −0.25% | 9.8% | 10.6% | 88.2% | 1.2% | 9.9% | 10.8% | 99% | 99% | -0.9 pp | -1.1 pp | -2.0 pp | halves agree | 0.07% | 0.07% | 117,662 |
| +2% / −0.5% | 17.5% | 18.8% | 78.0% | 3.2% | 18.0% | 19.5% | 97% | 97% | -1.5 pp | -2.0 pp | -3.4 pp | halves agree | 0.08% | 0.10% | 117,662 |
| +2% / −1% | 28.8% | 29.4% | 61.6% | 9.0% | 31.3% | 32.3% | 92% | 91% | -1.0 pp | -3.1 pp | -4.0 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +2% / −2% | 39.5% | 38.1% | 39.5% | 22.3% | 50.9% | 49.1% | 78% | 78% | +1.8 pp | -4.1 pp | -2.3 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +2% / −3% | 43.1% | 41.0% | 24.5% | 34.5% | 65.1% | 62.6% | 66% | 66% | +2.5 pp | -4.4 pp | -1.9 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +3% / −0.25% | 6.2% | 6.4% | 90.6% | 3.0% | 6.3% | 6.6% | 97% | 97% | -0.3 pp | -1.0 pp | -1.3 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −0.5% | 10.9% | 11.4% | 81.6% | 7.1% | 11.6% | 12.2% | 94% | 93% | -0.6 pp | -1.7 pp | -2.3 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −1% | 17.8% | 17.6% | 66.2% | 16.2% | 21.1% | 21.0% | 85% | 84% | +0.1 pp | -2.7 pp | -2.6 pp | HALVES DISAGREE | 0.00% | 0.01% | 117,662 |
| +3% / −2% | 24.5% | 23.1% | 43.1% | 33.7% | 37.4% | 34.9% | 66% | 66% | +2.5 pp | -4.4 pp | -1.9 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +3% / −3% | 26.6% | 25.1% | 26.6% | 48.4% | 51.5% | 48.5% | 52% | 52% | +2.9 pp | -5.3 pp | -2.4 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
13 of 13 Benjamini–Hochberg survivors in the 3d horizon also agree across both halves of history.
Sampling check: re-running this horizon on 525 strictly non-overlapping entries (rather than 117,662 every-bar entries) moves the asymmetry by a median of 2.0 pp and at most 6.8 pp.
263 non-overlapping entries · 366 coverage-span equivalents · median 276 bars per window
At this horizon: 0.25% ≈ 0.06σ · 0.5% ≈ 0.12σ · 1% ≈ 0.24σ · 2% ≈ 0.49σ · 3% ≈ 0.72σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 47.4% | 52.6% | 0.0% | -3.5 pp | -4.9 to -2.2 pp | 16707 |
| +0.25% / −0.5% | 65.2% | 34.8% | 0.0% | -3.8 pp | -5.5 to -2.3 pp | 11243 |
| +0.25% / −1% | 79.2% | 20.8% | 0.0% | -3.6 pp | -5.2 to -2.0 pp | 6935 |
| +0.25% / −2% | 88.8% | 11.2% | 0.1% | -2.1 pp | -3.7 to -0.7 pp | 4202 |
| +0.25% / −3% | 92.0% | 7.3% | 0.7% | -1.4 pp | -2.7 to -0.1 pp | 3733 |
| +0.5% / −0.25% | 31.3% | 68.7% | 0.0% | -3.7 pp | -5.4 to -2.2 pp | 12068 |
| +0.5% / −0.5% | 47.6% | 52.3% | 0.0% | -5.7 pp | -8.2 to -3.6 pp | 7267 |
| +0.5% / −1% | 65.1% | 34.9% | 0.0% | -5.8 pp | -8.6 to -3.4 pp | 3986 |
| +0.5% / −2% | 79.5% | 20.2% | 0.3% | -3.7 pp | -6.5 to -1.0 pp | 2260 |
| +0.5% / −3% | 84.6% | 13.3% | 2.1% | -2.5 pp | -5.1 to -0.1 pp | 1883 |
| +1% / −0.25% | 18.1% | 81.9% | 0.0% | -3.5 pp | -5.2 to -2.0 pp | 7824 |
| +1% / −0.5% | 30.8% | 69.2% | 0.0% | -5.8 pp | -8.6 to -3.4 pp | 4476 |
| +1% / −1% | 47.9% | 52.1% | 0.1% | -6.7 pp | -10.7 to -2.8 pp | 2341 |
| +1% / −2% | 65.5% | 32.8% | 1.8% | -4.2 pp | -9.0 to +0.3 pp | 1284 |
| +1% / −3% | 72.0% | 21.6% | 6.4% | -2.8 pp | -7.6 to +1.5 pp | 1068 |
| +2% / −0.25% | 10.2% | 89.7% | 0.0% | -2.1 pp | -3.6 to -0.6 pp | 5214 |
| +2% / −0.5% | 18.6% | 81.2% | 0.2% | -3.7 pp | -6.4 to -0.9 pp | 2672 |
| +2% / −1% | 31.9% | 66.8% | 1.4% | -4.2 pp | -9.0 to +0.3 pp | 1416 |
| +2% / −2% | 46.7% | 44.8% | 8.5% | -2.3 pp | -9.6 to +4.0 pp | 777 |
| +2% / −3% | 52.2% | 29.6% | 18.2% | -1.5 pp | -9.4 to +5.3 pp | 605 |
| +3% / −0.25% | 7.1% | 92.2% | 0.7% | -1.3 pp | -2.7 to -0.1 pp | 4461 |
| +3% / −0.5% | 13.0% | 85.1% | 1.9% | -2.4 pp | -5.1 to -0.1 pp | 2366 |
| +3% / −1% | 22.1% | 72.0% | 5.9% | -2.8 pp | -7.6 to +1.5 pp | 1220 |
| +3% / −2% | 32.3% | 49.4% | 18.3% | -1.5 pp | -9.4 to +5.3 pp | 647 |
| +3% / −3% | 36.0% | 32.8% | 31.2% | -1.2 pp | -10.9 to +7.7 pp | 471 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 47.4% | 49.6% | 50.4% | 0.0% | 47.4% | 49.6% | 100% | 100% | -2.2 pp | -1.3 pp | -3.5 pp | halves agree | 3.04% | 3.04% | 117,662 |
| +0.25% / −0.5% | 65.2% | 67.5% | 32.5% | 0.0% | 65.2% | 67.5% | 100% | 100% | -2.3 pp | -1.5 pp | -3.8 pp | halves agree | 1.32% | 1.25% | 117,662 |
| +0.25% / −1% | 79.2% | 81.4% | 18.6% | 0.0% | 79.2% | 81.4% | 100% | 100% | -2.2 pp | -1.4 pp | -3.6 pp | halves agree | 0.50% | 0.47% | 117,662 |
| +0.25% / −2% | 88.8% | 89.7% | 10.3% | 0.0% | 88.8% | 89.7% | 100% | 100% | -0.9 pp | -1.2 pp | -2.1 pp | halves agree | 0.09% | 0.08% | 117,662 |
| +0.25% / −3% | 92.0% | 92.2% | 7.1% | 0.7% | 92.7% | 92.8% | 99% | 99% | -0.2 pp | -1.2 pp | -1.4 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +0.5% / −0.25% | 31.3% | 33.5% | 66.5% | 0.0% | 31.3% | 33.5% | 100% | 100% | -2.2 pp | -1.5 pp | -3.7 pp | halves agree | 1.25% | 1.32% | 117,662 |
| +0.5% / −0.5% | 47.6% | 51.5% | 48.5% | 0.0% | 47.7% | 51.5% | 100% | 100% | -3.8 pp | -1.9 pp | -5.7 pp | halves agree | 0.89% | 0.89% | 117,662 |
| +0.5% / −1% | 65.1% | 68.8% | 31.2% | 0.0% | 65.1% | 68.8% | 100% | 100% | -3.7 pp | -2.2 pp | -5.8 pp | halves agree | 0.38% | 0.40% | 117,662 |
| +0.5% / −2% | 79.5% | 81.1% | 18.7% | 0.2% | 79.7% | 81.3% | 100% | 100% | -1.6 pp | -2.1 pp | -3.7 pp | halves agree | 0.10% | 0.08% | 117,662 |
| +0.5% / −3% | 84.6% | 85.1% | 13.0% | 1.9% | 86.4% | 86.7% | 98% | 98% | -0.4 pp | -2.1 pp | -2.5 pp | halves agree | 0.02% | 0.01% | 117,662 |
| +1% / −0.25% | 18.1% | 20.3% | 79.7% | 0.0% | 18.1% | 20.3% | 100% | 100% | -2.2 pp | -1.3 pp | -3.5 pp | halves agree | 0.47% | 0.50% | 117,662 |
| +1% / −0.5% | 30.8% | 34.5% | 65.5% | 0.0% | 30.8% | 34.5% | 100% | 100% | -3.7 pp | -2.1 pp | -5.8 pp | halves agree | 0.40% | 0.38% | 117,662 |
| +1% / −1% | 47.9% | 51.8% | 48.1% | 0.1% | 47.9% | 51.9% | 100% | 100% | -4.0 pp | -2.7 pp | -6.7 pp | halves agree | 0.22% | 0.22% | 117,662 |
| +1% / −2% | 65.5% | 66.7% | 31.9% | 1.4% | 66.6% | 67.7% | 98% | 99% | -1.0 pp | -3.2 pp | -4.2 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +1% / −3% | 72.0% | 72.0% | 22.1% | 5.9% | 76.9% | 76.5% | 94% | 94% | +0.4 pp | -3.2 pp | -2.8 pp | HALVES DISAGREE | 0.01% | 0.00% | 117,662 |
| +2% / −0.25% | 10.2% | 11.1% | 88.8% | 0.1% | 10.2% | 11.1% | 100% | 100% | -0.9 pp | -1.2 pp | -2.1 pp | halves agree | 0.08% | 0.09% | 117,662 |
| +2% / −0.5% | 18.6% | 20.1% | 79.6% | 0.3% | 18.7% | 20.2% | 100% | 100% | -1.5 pp | -2.1 pp | -3.7 pp | halves agree | 0.08% | 0.10% | 117,662 |
| +2% / −1% | 31.9% | 32.7% | 65.5% | 1.8% | 32.3% | 33.3% | 99% | 98% | -1.0 pp | -3.2 pp | -4.2 pp | halves agree | 0.02% | 0.02% | 117,662 |
| +2% / −2% | 46.7% | 44.8% | 46.7% | 8.5% | 51.0% | 49.0% | 91% | 91% | +2.0 pp | -4.3 pp | -2.3 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +2% / −3% | 52.2% | 49.4% | 32.3% | 18.3% | 63.8% | 60.4% | 82% | 82% | +3.4 pp | -4.9 pp | -1.5 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +3% / −0.25% | 7.1% | 7.3% | 92.0% | 0.7% | 7.1% | 7.3% | 99% | 99% | -0.2 pp | -1.2 pp | -1.3 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −0.5% | 13.0% | 13.3% | 84.6% | 2.1% | 13.2% | 13.6% | 98% | 98% | -0.3 pp | -2.1 pp | -2.4 pp | halves agree | 0.01% | 0.02% | 117,662 |
| +3% / −1% | 22.1% | 21.6% | 72.0% | 6.4% | 23.5% | 23.1% | 94% | 94% | +0.4 pp | -3.2 pp | -2.8 pp | HALVES DISAGREE | 0.00% | 0.01% | 117,662 |
| +3% / −2% | 32.3% | 29.6% | 52.2% | 18.2% | 39.6% | 36.2% | 82% | 82% | +3.4 pp | -4.9 pp | -1.5 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
| +3% / −3% | 36.0% | 32.8% | 36.0% | 31.2% | 52.4% | 47.6% | 69% | 69% | +4.8 pp | -5.9 pp | -1.2 pp | HALVES DISAGREE | 0.00% | 0.00% | 117,662 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
13 of 13 Benjamini–Hochberg survivors in the 5d horizon also agree across both halves of history.
Sampling check: re-running this horizon on 263 strictly non-overlapping entries (rather than 117,662 every-bar entries) moves the asymmetry by a median of 3.6 pp and at most 17.8 pp.
Across the 4 horizons this window publishes, 64 of 68 cells that look significant uncorrected survive Benjamini–Hochberg at q=0.05, and 64 of those 64 Benjamini–Hochberg survivors also agree across both halves of history — survives Benjamini-Hochberg AND both halves of history agree in sign. Only confirmed cells are shown in bold and ringed on the charts; a cell that is significant but not confirmed keeps its colour, its number and its interval, marked “not confirmed” rather than hidden. The gaps between all three counts are themselves the point.
Not every window in the barrier grid above resolves: especially at shorter horizons and wider rungs, many windows touch neither barrier before the horizon runs out. This section measures what those windows did anyway, resolved or not — how far price ran in the trader's favour (MFE) and how much heat it took against them (MAE) — from intrabar high and low, never from closes: a stop is hit by the low, not the close. See methodology.
MFE is signed positive and MAE is signed negative. For MFE the higher quantile is the better outcome; for MAE it is the opposite — the lower (more negative) quantile is the worse drawdown, and the higher quantile is the mildest one. The tables below label each MAE column by what it means, not just by its quantile, so the two are never read the same way by mistake.
At the 4h horizon, a +1% / −1% bracket resolves — hits target or stop — in only 32.1% of long entries; the other 67.9% touch neither barrier inside the horizon. Across every 4h window regardless of whether it resolved, the median run in the trader's favour (MFE) was +0.39% and the best tenth of windows reached +1.29%, while the median heat taken against them (MAE) was -0.41% and the worst tenth of windows drew down -1.38% — that is what the rest of the grid's “neither” column was doing instead of resolving.
7,751 non-overlapping entries · 10,993 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 7,751 | +0.39% | +1.29% | -0.41% | -1.38% |
| Short | 7,751 | +0.41% | +1.40% | -0.38% | -1.27% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 76,540 | 27.2% | 15976 |
| +0.25% / −0.5% | 76,540 | 10.3% | 7634 |
| +0.25% / −1% | 76,540 | 2.4% | 5655 |
| +0.25% / −2% | 76,540 | 0.3% | 8788 |
| +0.25% / −3% | 76,540 | 0.1% | 12007 |
| +0.5% / −0.25% | 46,748 | 34.9% | 5892 |
| +0.5% / −0.5% | 46,748 | 15.0% | 3442 |
| +0.5% / −1% | 46,748 | 4.0% | 2754 |
| +0.5% / −2% | 46,748 | 0.5% | 4211 |
| +0.5% / −3% | 46,748 | 0.1% | 5935 |
| +1% / −0.25% | 18,677 | 42.6% | 1932 |
| +1% / −0.5% | 18,677 | 21.5% | 1144 |
| +1% / −1% | 18,677 | 6.9% | 938 |
| +1% / −2% | 18,677 | 1.1% | 1122 |
| +1% / −3% | 18,677 | 0.3% | 2052 |
| +2% / −0.25% | 4,555 | 50.9% | 540 |
| +2% / −0.5% | 4,555 | 29.4% | 317 |
| +2% / −1% | 4,555 | 11.9% | 211 |
| +2% / −2% | 4,555 | 3.2% | 226 |
| +2% / −3% | 4,555 | 1.2% | 394 |
| +3% / −0.25% | 1,392 | 53.2% | 150 |
| +3% / −0.5% | 1,392 | 33.7% | 105 |
| +3% / −1% | 1,392 | 16.5% | 77 |
| +3% / −2% | 1,392 | 5.7% | 110 |
| +3% / −3% | 1,392 | 2.8% | 139 |
1,296 non-overlapping entries · 1,832 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 1,296 | +1.03% | +3.15% | -1.07% | -3.17% |
| Short | 1,296 | +1.08% | +3.27% | -1.02% | -3.06% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 100,506 | 41.9% | 23737 |
| +0.25% / −0.5% | 100,506 | 23.8% | 15017 |
| +0.25% / −1% | 100,506 | 9.3% | 7021 |
| +0.25% / −2% | 100,506 | 1.7% | 6384 |
| +0.25% / −3% | 100,506 | 0.4% | 6946 |
| +0.5% / −0.25% | 85,383 | 56.5% | 13571 |
| +0.5% / −0.5% | 85,383 | 35.5% | 7757 |
| +0.5% / −1% | 85,383 | 14.7% | 3481 |
| +0.5% / −2% | 85,383 | 2.7% | 3248 |
| +0.5% / −3% | 85,383 | 0.7% | 3652 |
| +1% / −0.25% | 59,951 | 65.7% | 5844 |
| +1% / −0.5% | 59,951 | 43.5% | 2899 |
| +1% / −1% | 59,951 | 18.8% | 1675 |
| +1% / −2% | 59,951 | 3.9% | 1425 |
| +1% / −3% | 59,951 | 1.2% | 1660 |
| +2% / −0.25% | 28,484 | 67.6% | 1718 |
| +2% / −0.5% | 28,484 | 45.9% | 973 |
| +2% / −1% | 28,484 | 21.6% | 627 |
| +2% / −2% | 28,484 | 5.9% | 521 |
| +2% / −3% | 28,484 | 2.5% | 643 |
| +3% / −0.25% | 13,187 | 67.7% | 778 |
| +3% / −0.5% | 13,187 | 47.1% | 416 |
| +3% / −1% | 13,187 | 25.0% | 300 |
| +3% / −2% | 13,187 | 8.7% | 257 |
| +3% / −3% | 13,187 | 4.1% | 307 |
525 non-overlapping entries · 610 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 525 | +1.76% | +5.20% | -1.69% | -5.00% |
| Short | 525 | +1.71% | +5.26% | -1.73% | -4.94% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 108,019 | 45.1% | 21494 |
| +0.25% / −0.5% | 108,019 | 27.6% | 15137 |
| +0.25% / −1% | 108,019 | 13.3% | 8613 |
| +0.25% / −2% | 108,019 | 3.9% | 5730 |
| +0.25% / −3% | 108,019 | 1.3% | 5681 |
| +0.5% / −0.25% | 98,124 | 61.1% | 13217 |
| +0.5% / −0.5% | 98,124 | 41.9% | 8404 |
| +0.5% / −1% | 98,124 | 21.9% | 4494 |
| +0.5% / −2% | 98,124 | 6.5% | 3160 |
| +0.5% / −3% | 98,124 | 2.2% | 3301 |
| +1% / −0.25% | 80,500 | 73.0% | 7245 |
| +1% / −0.5% | 80,500 | 54.9% | 4092 |
| +1% / −1% | 80,500 | 31.4% | 2451 |
| +1% / −2% | 80,500 | 10.0% | 1551 |
| +1% / −3% | 80,500 | 3.2% | 1810 |
| +2% / −0.25% | 53,245 | 78.3% | 3303 |
| +2% / −0.5% | 53,245 | 61.1% | 1733 |
| +2% / −1% | 53,245 | 36.3% | 1027 |
| +2% / −2% | 53,245 | 12.7% | 795 |
| +2% / −3% | 53,245 | 4.7% | 804 |
| +3% / −0.25% | 33,355 | 78.3% | 1851 |
| +3% / −0.5% | 33,355 | 61.6% | 1035 |
| +3% / −1% | 33,355 | 37.0% | 646 |
| +3% / −2% | 33,355 | 13.6% | 436 |
| +3% / −3% | 33,355 | 6.3% | 421 |
263 non-overlapping entries · 366 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 263 | +2.35% | +6.69% | -2.11% | -5.97% |
| Short | 263 | +2.15% | +6.34% | -2.30% | -6.27% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 110,499 | 46.3% | 21184 |
| +0.25% / −0.5% | 110,499 | 29.2% | 16440 |
| +0.25% / −1% | 110,499 | 15.1% | 9707 |
| +0.25% / −2% | 110,499 | 5.4% | 6079 |
| +0.25% / −3% | 110,499 | 2.0% | 5259 |
| +0.5% / −0.25% | 103,110 | 62.9% | 14179 |
| +0.5% / −0.5% | 103,110 | 44.6% | 9484 |
| +0.5% / −1% | 103,110 | 25.3% | 5259 |
| +0.5% / −2% | 103,110 | 9.2% | 3544 |
| +0.5% / −3% | 103,110 | 3.5% | 3016 |
| +1% / −0.25% | 89,344 | 75.6% | 8318 |
| +1% / −0.5% | 89,344 | 58.9% | 4807 |
| +1% / −1% | 89,344 | 36.7% | 3014 |
| +1% / −2% | 89,344 | 13.8% | 1820 |
| +1% / −3% | 89,344 | 5.2% | 1833 |
| +2% / −0.25% | 65,858 | 81.6% | 4578 |
| +2% / −0.5% | 65,858 | 66.6% | 2512 |
| +2% / −1% | 65,858 | 43.0% | 1429 |
| +2% / −2% | 65,858 | 16.6% | 949 |
| +2% / −3% | 65,858 | 6.7% | 906 |
| +3% / −0.25% | 46,032 | 81.8% | 2680 |
| +3% / −0.5% | 46,032 | 66.8% | 1448 |
| +3% / −1% | 46,032 | 43.5% | 841 |
| +3% / −2% | 46,032 | 17.3% | 518 |
| +3% / −3% | 46,032 | 7.9% | 471 |
| Horizon | median move | p90 move | p99 move | non-overlapping entries | n | spread as share of median move |
|---|---|---|---|---|---|---|
| 4h | 0.37% | 1.30% | 3.26% | 7,751 | 117,392 | — |
| 1d | 1.00% | 3.18% | — | 1,296 | 117,397 | — |
| 3d | 1.65% | 5.07% | — | 525 | 117,651 | — |
| 5d | 2.19% | 6.52% | — | 263 | 117,661 | — |
No spread configured for this asset, so cost is not reported.
Sigma may be scaled by the square root of time; medians, stop distances and targets may not. Variance grows in proportion to time, so σ — its square root — grows with the square root of time. The typical move grows faster than that, so scaling a median from a short horizon to a longer one sets it too tight — on every asset and window this site publishes. The tails do not miss in one direction the way the median does, so a stop distance or a target is read off the measured table above, never scaled to it.
| Horizon | q | √t factor from 4h | median move, √t prediction | median move, measured | error | median ÷ σ | kurtosis | non-overlapping entries | n |
|---|---|---|---|---|---|---|---|---|---|
| 4h | 16 bars | ×1.00 | 0.369% | 0.369% | +0.0% | 0.400 | 30.7 | 7,751 | 117,392 |
| 1d | 92 bars | ×2.40 | 0.885% | 0.997% | +12.6% | 0.470 | 18.8 | 1,296 | 117,397 |
| 3d | 224 bars | ×3.74 | 1.381% | 1.649% | +19.4% | 0.486 | 12.3 | 525 | 117,651 |
| 5d | 276 bars | ×4.15 | 1.533% | 2.194% | +43.1% | 0.513 | 8.8 | 263 | 117,661 |
A positive error means the real move is larger than √t predicts. The error is relative — (measured − predicted) ÷ predicted — not itself a move size in percent of price, even though it sits between two columns that are. The reason a positive error is the norm is in the last two columns: aggregation pulls the distribution toward normal, where median ÷ σ is 0.674 and kurtosis is 3.0. By the longest horizon this table actually publishes (5d), median ÷ σ is 0.513 and kurtosis is 8.8 — read those against the Gaussian reference above rather than assuming the shape has normalised by then.
VR(q) = Var(q-bar move) ÷ (q × Var(1-bar move)). Above 1 moves extend and a distant target is more reachable than its distance suggests; below 1 they retrace. The null is exactly 1. q is derived from this asset's own bar calendar, so “5d” is however many bars a five-calendar-day window actually contains — usually fewer than five trading days, because it swallows a weekend.
| Horizon | q | VR | 95% interval | as move size vs random walk | robust z | p (robust) | blocks | homoskedastic z | verdict | split-half VR (first / second) |
|---|---|---|---|---|---|---|---|---|---|---|
| 4h | 16 bars ≈ 0.17 trading days | 0.955 | 0.872 to 1.037 | 93% to 102% | -1.08 | 0.280 | 7,410 | -3.56 | not significant | 0.968 / 0.950 — halves agree |
| 1d | 92 bars ≈ 1.00 trading days | 0.988 | 0.814 to 1.163 | 90% to 108% | -0.13 | 0.897 | 1,288 | -0.36 | not significant | 0.942 / 1.006 — HALVES DISAGREE |
| 3d | 224 bars ≈ 2.43 trading days | 0.975 | 0.725 to 1.225 | 85% to 111% | -0.19 | 0.846 | 529 | -0.50 | not significant | 0.956 / 0.983 — halves agree |
| 5d | 276 bars ≈ 3.00 trading days | 0.965 | 0.695 to 1.235 | 83% to 111% | -0.25 | 0.800 | 429 | -0.63 | not significant | 0.969 / 0.965 — halves agree |
Only the robust z (and its p, the column so labelled) is tested; it allows the size of moves to change over time, which this asset's does. The homoskedastic z column assumes it does not, and is shown to make the gap visible: 1 of these horizons would look significant under it against 0 that survive Benjamini–Hochberg on the robust one. That difference is volatility clustering read as mean reversion, not a finding.
The verdict column applies one rule and states it in full: survives Benjamini-Hochberg on the robust standard error AND both halves of history agree in sign of (VR - 1). The split-half column is the second half of that rule shown as data — the same VR recomputed on each disjoint half of the window, at the same q. The comparison is on the sign of (VR − 1), not of VR: VR is always positive, so comparing raw signs would mark every pair as agreeing and the check could never fail. Halves that disagree do not overturn a significant result on their own; they stop it being called confirmed.
How far price moved while you could not trade. Split three ways and never pooled: a venue halt and a weekend are different phenomena and only one of them carries real overnight risk. No stop protects you across either.
| Kind | n | median | p90 | p99 | >1× median bar | >3× |
|---|---|---|---|---|---|---|
| session | 0 | insufficient data | ||||
| weekend | 221 | -0.004 | 0.073 | 1.117 | 22.6% | 9.0% |
| daily break | 894 | -0.004 | 0.044 | 0.205 | 8.4% | 2.0% |
Gaps are in USD/oz, signed. Median bar range for comparison: 0.062.
Kept for placing an entry, not as a headline: the horizon this site reports at is hours to days, and a multi-day stop sized to a 15-minute range is the error that shakes a trader out of positions that would have worked.
| Unit | n | mean | p10 | p25 | p50 | p75 | p90 | p95 | p99 |
|---|---|---|---|---|---|---|---|---|---|
| USD/oz | 117662 | 0.1220 | 0.0270 | 0.0390 | 0.0620 | 0.1140 | 0.2570 | 0.4270 | 0.9720 |
| % of price | 117662 | 0.3040 | 0.1065 | 0.1530 | 0.2292 | 0.3599 | 0.5646 | 0.7504 | 1.3552 |
| Unit | n | mean | p10 | p25 | p50 | p75 | p90 | p95 | p99 |
|---|---|---|---|---|---|---|---|---|---|
| USD/oz | 116546 | 0.1223 | 0.0270 | 0.0390 | 0.0620 | 0.1140 | 0.2580 | 0.4280 | 0.9726 |
| % of price | 116546 | 0.3042 | 0.1067 | 0.1533 | 0.2294 | 0.3602 | 0.5648 | 0.7504 | 1.3554 |
23,545 bars · 2025-08-03T22:00:00+00:00 to 2026-08-02T23:45:00+00:00
Over this window the price drifted +57.8% a year. Each cell below shows how much of its long/short gap that drift accounts for; only the remainder is tested.
1,540 non-overlapping entries · 2,184 coverage-span equivalents · median 16 bars per window
At this horizon: 0.25% ≈ 0.15σ · 0.5% ≈ 0.30σ · 1% ≈ 0.61σ · 2% ≈ 1.21σ · 3% ≈ 1.81σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.0% | 54.8% | 2.3% | -5.0 pp not confirmed | -7.5 to -2.7 pp | 3581 |
| +0.25% / −0.5% | 59.2% | 35.0% | 5.8% | -5.4 pp not confirmed | -7.7 to -3.0 pp | 2934 |
| +0.25% / −1% | 70.0% | 18.2% | 11.9% | -4.7 pp not confirmed | -6.8 to -2.7 pp | 1783 |
| +0.25% / −2% | 74.5% | 6.1% | 19.5% | -2.2 pp not confirmed | -3.6 to -0.7 pp | 1235 |
| +0.25% / −3% | 75.1% | 2.7% | 22.2% | -1.5 pp not confirmed | -2.3 to -0.7 pp | 1284 |
| +0.5% / −0.25% | 28.0% | 66.4% | 5.6% | -5.4 pp not confirmed | -7.6 to -3.0 pp | 4244 |
| +0.5% / −0.5% | 40.6% | 47.1% | 12.3% | -7.4 pp not confirmed | -10.6 to -3.9 pp | 3486 |
| +0.5% / −1% | 50.3% | 26.1% | 23.6% | -7.3 pp not confirmed | -10.9 to -3.8 pp | 1695 |
| +0.5% / −2% | 54.7% | 8.9% | 36.4% | -3.8 pp not confirmed | -6.5 to -1.0 pp | 909 |
| +0.5% / −3% | 55.4% | 4.1% | 40.5% | -3.1 pp not confirmed | -4.6 to -1.6 pp | 845 |
| +1% / −0.25% | 14.3% | 74.5% | 11.1% | -4.5 pp not confirmed | -6.5 to -2.4 pp | 3052 |
| +1% / −0.5% | 21.6% | 55.7% | 22.7% | -7.2 pp not confirmed | -10.8 to -3.7 pp | 2204 |
| +1% / −1% | 27.8% | 32.1% | 40.1% | -9.1 pp not confirmed | -14.2 to -4.0 pp | 1406 |
| +1% / −2% | 31.1% | 11.4% | 57.5% | -6.9 pp not confirmed | -12.2 to -1.6 pp | 609 |
| +1% / −3% | 31.6% | 5.5% | 62.9% | -7.0 pp not confirmed | -10.2 to -3.5 pp | 499 |
| +2% / −0.25% | 4.7% | 77.1% | 18.2% | -2.0 pp not confirmed | -3.4 to -0.6 pp | 1853 |
| +2% / −0.5% | 7.2% | 58.8% | 34.0% | -3.7 pp not confirmed | -6.4 to -0.9 pp | 1126 |
| +2% / −1% | 9.4% | 34.6% | 56.0% | -6.9 pp not confirmed | -12.2 to -1.6 pp | 764 |
| +2% / −2% | 10.8% | 12.8% | 76.4% | -10.2 pp | -19.2 to +0.1 pp | 431 |
| +2% / −3% | 11.2% | 6.4% | 82.4% | -15.0 pp not confirmed | -23.1 to -6.7 pp | 333 |
| +3% / −0.25% | 1.7% | 77.5% | 20.9% | -1.4 pp not confirmed | -2.1 to -0.6 pp | 2127 |
| +3% / −0.5% | 2.5% | 59.2% | 38.3% | -3.0 pp not confirmed | -4.5 to -1.4 pp | 1164 |
| +3% / −1% | 3.3% | 35.0% | 61.7% | -6.9 pp not confirmed | -10.2 to -3.4 pp | 708 |
| +3% / −2% | 3.8% | 13.0% | 83.1% | -15.0 pp not confirmed | -23.1 to -6.7 pp | 351 |
| +3% / −3% | 4.0% | 6.6% | 89.4% | -25.9 pp not confirmed | -38.0 to -15.2 pp | 289 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.0% | 46.2% | 51.5% | 2.3% | 44.0% | 47.3% | 98% | 98% | -3.3 pp | -1.7 pp | -5.0 pp | not computed: this window is too short for two halves to detect instability | 8.56% | 8.56% | 23,364 |
| +0.25% / −0.5% | 59.2% | 62.8% | 31.6% | 5.6% | 62.8% | 66.5% | 94% | 94% | -3.7 pp | -1.7 pp | -5.4 pp | not computed: this window is too short for two halves to detect instability | 3.60% | 3.65% | 23,364 |
| +0.25% / −1% | 70.0% | 73.6% | 15.3% | 11.1% | 79.4% | 82.8% | 88% | 89% | -3.4 pp | -1.4 pp | -4.7 pp | not computed: this window is too short for two halves to detect instability | 1.16% | 0.98% | 23,364 |
| +0.25% / −2% | 74.5% | 77.0% | 4.8% | 18.2% | 92.5% | 94.1% | 81% | 82% | -1.6 pp | -0.6 pp | -2.2 pp | not computed: this window is too short for two halves to detect instability | 0.26% | 0.15% | 23,364 |
| +0.25% / −3% | 75.1% | 77.4% | 1.7% | 20.9% | 96.6% | 97.8% | 78% | 79% | -1.3 pp | -0.2 pp | -1.5 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.03% | 23,364 |
| +0.5% / −0.25% | 28.0% | 31.4% | 62.8% | 5.8% | 29.6% | 33.4% | 94% | 94% | -3.7 pp | -1.7 pp | -5.4 pp | not computed: this window is too short for two halves to detect instability | 3.65% | 3.60% | 23,364 |
| +0.5% / −0.5% | 40.6% | 45.2% | 42.4% | 12.3% | 46.3% | 51.6% | 88% | 88% | -5.3 pp | -2.1 pp | -7.4 pp | not computed: this window is too short for two halves to detect instability | 1.86% | 1.86% | 23,364 |
| +0.5% / −1% | 50.3% | 55.1% | 22.1% | 22.7% | 65.9% | 71.4% | 76% | 77% | -5.5 pp | -1.8 pp | -7.3 pp | not computed: this window is too short for two halves to detect instability | 0.68% | 0.58% | 23,364 |
| +0.5% / −2% | 54.7% | 58.7% | 7.3% | 34.0% | 86.0% | 88.9% | 64% | 66% | -2.9 pp | -0.9 pp | -3.8 pp | not computed: this window is too short for two halves to detect instability | 0.19% | 0.12% | 23,364 |
| +0.5% / −3% | 55.4% | 59.1% | 2.6% | 38.3% | 93.1% | 95.9% | 59% | 62% | -2.8 pp | -0.4 pp | -3.1 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.02% | 23,364 |
| +1% / −0.25% | 14.3% | 17.0% | 71.2% | 11.9% | 16.1% | 19.3% | 89% | 88% | -3.2 pp | -1.3 pp | -4.5 pp | not computed: this window is too short for two halves to detect instability | 0.98% | 1.16% | 23,364 |
| +1% / −0.5% | 21.6% | 25.4% | 51.0% | 23.6% | 27.9% | 33.3% | 77% | 76% | -5.4 pp | -1.8 pp | -7.2 pp | not computed: this window is too short for two halves to detect instability | 0.58% | 0.68% | 23,364 |
| +1% / −1% | 27.8% | 31.8% | 28.1% | 40.1% | 46.4% | 53.1% | 60% | 60% | -6.7 pp | -2.4 pp | -9.1 pp | not computed: this window is too short for two halves to detect instability | 0.29% | 0.29% | 23,364 |
| +1% / −2% | 31.1% | 34.5% | 9.5% | 56.0% | 73.1% | 78.4% | 43% | 44% | -5.3 pp | -1.6 pp | -6.9 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.07% | 23,364 |
| +1% / −3% | 31.6% | 35.0% | 3.3% | 61.7% | 85.2% | 91.4% | 37% | 38% | -6.2 pp | -0.8 pp | -7.0 pp | not computed: this window is too short for two halves to detect instability | 0.05% | 0.01% | 23,364 |
| +2% / −0.25% | 4.7% | 5.8% | 74.7% | 19.5% | 5.7% | 7.2% | 82% | 81% | -1.5 pp | -0.6 pp | -2.0 pp | not computed: this window is too short for two halves to detect instability | 0.15% | 0.26% | 23,364 |
| +2% / −0.5% | 7.2% | 8.7% | 54.9% | 36.4% | 10.9% | 13.7% | 66% | 64% | -2.8 pp | -0.9 pp | -3.7 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.19% | 23,364 |
| +2% / −1% | 9.4% | 11.4% | 31.2% | 57.5% | 21.4% | 26.7% | 44% | 43% | -5.3 pp | -1.6 pp | -6.9 pp | not computed: this window is too short for two halves to detect instability | 0.07% | 0.09% | 23,364 |
| +2% / −2% | 10.8% | 12.8% | 10.9% | 76.4% | 45.9% | 54.0% | 24% | 24% | -8.2 pp | -2.0 pp | -10.2 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.02% | 23,364 |
| +2% / −3% | 11.2% | 13.0% | 3.8% | 83.1% | 63.8% | 77.3% | 18% | 17% | -13.5 pp | -1.5 pp | -15.0 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.00% | 23,364 |
| +3% / −0.25% | 1.7% | 2.6% | 75.3% | 22.2% | 2.1% | 3.3% | 79% | 78% | -1.2 pp | -0.2 pp | -1.4 pp | not computed: this window is too short for two halves to detect instability | 0.03% | 0.12% | 23,364 |
| +3% / −0.5% | 2.5% | 4.0% | 55.5% | 40.5% | 4.1% | 6.7% | 62% | 59% | -2.6 pp | -0.4 pp | -3.0 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.09% | 23,364 |
| +3% / −1% | 3.3% | 5.5% | 31.7% | 62.9% | 8.6% | 14.7% | 38% | 37% | -6.1 pp | -0.8 pp | -6.9 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.05% | 23,364 |
| +3% / −2% | 3.8% | 6.4% | 11.2% | 82.4% | 22.7% | 36.2% | 17% | 18% | -13.4 pp | -1.5 pp | -15.0 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.01% | 23,364 |
| +3% / −3% | 4.0% | 6.6% | 4.0% | 89.4% | 37.9% | 62.1% | 11% | 11% | -24.3 pp | -1.6 pp | -25.9 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.00% | 23,364 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
24 Benjamini–Hochberg survivors in the 4h horizon — not computed: this window is too short for two halves to detect instability, so none can be confirmed here.
Sampling check: re-running this horizon on 1,540 strictly non-overlapping entries (rather than 23,364 every-bar entries) moves the asymmetry by a median of 1.2 pp and at most 7.7 pp.
258 non-overlapping entries · 364 coverage-span equivalents · median 92 bars per window
At this horizon: 0.25% ≈ 0.06σ · 0.5% ≈ 0.13σ · 1% ≈ 0.25σ · 2% ≈ 0.50σ · 3% ≈ 0.75σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.6% | 55.8% | 0.6% | -5.6 pp not confirmed | -8.2 to -3.1 pp | 3477 |
| +0.25% / −0.5% | 61.5% | 37.5% | 1.0% | -6.2 pp not confirmed | -9.1 to -3.3 pp | 2708 |
| +0.25% / −1% | 75.9% | 22.0% | 2.1% | -5.3 pp not confirmed | -8.1 to -2.1 pp | 2156 |
| +0.25% / −2% | 85.4% | 10.7% | 3.8% | -2.5 pp | -4.6 to -0.1 pp | 1440 |
| +0.25% / −3% | 88.1% | 6.8% | 5.1% | -2.2 pp | -4.2 to -0.1 pp | 1253 |
| +0.5% / −0.25% | 29.9% | 69.3% | 0.9% | -6.2 pp not confirmed | -9.1 to -3.4 pp | 3699 |
| +0.5% / −0.5% | 45.4% | 53.0% | 1.7% | -8.7 pp not confirmed | -12.8 to -4.0 pp | 2079 |
| +0.5% / −1% | 61.2% | 34.9% | 3.8% | -8.1 pp not confirmed | -12.5 to -2.8 pp | 1420 |
| +0.5% / −2% | 74.1% | 18.2% | 7.7% | -4.0 pp | -8.1 to +0.5 pp | 858 |
| +0.5% / −3% | 77.7% | 11.9% | 10.4% | -4.0 pp | -7.8 to -0.1 pp | 673 |
| +1% / −0.25% | 17.9% | 80.5% | 1.5% | -5.1 pp not confirmed | -7.9 to -1.9 pp | 2464 |
| +1% / −0.5% | 29.7% | 67.4% | 3.0% | -8.0 pp not confirmed | -12.4 to -2.6 pp | 1291 |
| +1% / −1% | 44.3% | 48.6% | 7.1% | -8.0 pp not confirmed | -14.0 to -1.1 pp | 894 |
| +1% / −2% | 57.9% | 28.2% | 13.9% | -4.8 pp | -11.6 to +2.5 pp | 507 |
| +1% / −3% | 62.1% | 18.5% | 19.4% | -5.4 pp | -11.9 to +1.8 pp | 390 |
| +2% / −0.25% | 9.6% | 87.2% | 3.3% | -2.4 pp | -4.5 to -0.0 pp | 2046 |
| +2% / −0.5% | 16.7% | 76.7% | 6.6% | -3.9 pp | -8.0 to +0.6 pp | 1047 |
| +2% / −1% | 26.6% | 58.7% | 14.7% | -4.8 pp | -11.6 to +2.5 pp | 678 |
| +2% / −2% | 36.9% | 36.4% | 26.7% | -3.3 pp | -13.7 to +6.9 pp | 345 |
| +2% / −3% | 40.1% | 24.3% | 35.6% | -5.9 pp | -17.1 to +5.2 pp | 252 |
| +3% / −0.25% | 5.8% | 89.0% | 5.3% | -2.1 pp | -4.1 to -0.1 pp | 1461 |
| +3% / −0.5% | 10.1% | 79.3% | 10.6% | -3.9 pp | -7.8 to -0.0 pp | 826 |
| +3% / −1% | 16.3% | 62.1% | 21.6% | -5.3 pp | -11.9 to +1.9 pp | 501 |
| +3% / −2% | 22.6% | 39.7% | 37.6% | -5.8 pp | -17.1 to +5.3 pp | 256 |
| +3% / −3% | 24.9% | 26.8% | 48.3% | -8.9 pp | -23.0 to +5.4 pp | 184 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.6% | 47.2% | 52.2% | 0.6% | 43.9% | 47.5% | 99% | 99% | -3.6 pp | -1.9 pp | -5.6 pp | not computed: this window is too short for two halves to detect instability | 8.57% | 8.57% | 23,364 |
| +0.25% / −0.5% | 61.5% | 65.6% | 33.5% | 0.9% | 62.1% | 66.2% | 99% | 99% | -4.1 pp | -2.1 pp | -6.2 pp | not computed: this window is too short for two halves to detect instability | 3.63% | 3.66% | 23,364 |
| +0.25% / −1% | 75.9% | 79.5% | 18.9% | 1.5% | 77.5% | 80.8% | 98% | 98% | -3.3 pp | -2.0 pp | -5.3 pp | not computed: this window is too short for two halves to detect instability | 1.19% | 0.99% | 23,364 |
| +0.25% / −2% | 85.4% | 87.0% | 9.7% | 3.3% | 88.9% | 90.0% | 96% | 97% | -1.1 pp | -1.4 pp | -2.5 pp | not computed: this window is too short for two halves to detect instability | 0.26% | 0.15% | 23,364 |
| +0.25% / −3% | 88.1% | 88.9% | 5.8% | 5.3% | 92.8% | 93.9% | 95% | 95% | -1.1 pp | -1.1 pp | -2.2 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.03% | 23,364 |
| +0.5% / −0.25% | 29.9% | 33.9% | 65.1% | 1.0% | 30.1% | 34.2% | 99% | 99% | -4.1 pp | -2.1 pp | -6.2 pp | not computed: this window is too short for two halves to detect instability | 3.66% | 3.63% | 23,364 |
| +0.5% / −0.5% | 45.4% | 51.1% | 47.3% | 1.7% | 46.1% | 51.9% | 98% | 98% | -5.8 pp | -2.9 pp | -8.7 pp | not computed: this window is too short for two halves to detect instability | 1.88% | 1.88% | 23,364 |
| +0.5% / −1% | 61.2% | 66.8% | 30.2% | 3.0% | 63.7% | 68.8% | 96% | 97% | -5.1 pp | -3.0 pp | -8.1 pp | not computed: this window is too short for two halves to detect instability | 0.68% | 0.58% | 23,364 |
| +0.5% / −2% | 74.1% | 76.5% | 16.8% | 6.6% | 80.3% | 82.0% | 92% | 93% | -1.7 pp | -2.3 pp | -4.0 pp | not computed: this window is too short for two halves to detect instability | 0.19% | 0.12% | 23,364 |
| +0.5% / −3% | 77.7% | 79.3% | 10.2% | 10.6% | 86.7% | 88.6% | 90% | 89% | -1.9 pp | -2.1 pp | -4.0 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.02% | 23,364 |
| +1% / −0.25% | 17.9% | 20.9% | 77.1% | 2.1% | 18.2% | 21.3% | 98% | 98% | -3.1 pp | -2.0 pp | -5.1 pp | not computed: this window is too short for two halves to detect instability | 0.99% | 1.19% | 23,364 |
| +1% / −0.5% | 29.7% | 34.2% | 61.9% | 3.8% | 30.6% | 35.6% | 97% | 96% | -5.0 pp | -2.9 pp | -8.0 pp | not computed: this window is too short for two halves to detect instability | 0.58% | 0.68% | 23,364 |
| +1% / −1% | 44.3% | 48.3% | 44.6% | 7.1% | 47.7% | 52.0% | 93% | 93% | -4.2 pp | -3.7 pp | -8.0 pp | not computed: this window is too short for two halves to detect instability | 0.30% | 0.30% | 23,364 |
| +1% / −2% | 57.9% | 58.6% | 26.7% | 14.7% | 67.2% | 68.7% | 86% | 85% | -1.5 pp | -3.4 pp | -4.8 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.07% | 23,364 |
| +1% / −3% | 62.1% | 62.1% | 16.3% | 21.6% | 77.0% | 79.2% | 81% | 78% | -2.2 pp | -3.2 pp | -5.4 pp | not computed: this window is too short for two halves to detect instability | 0.05% | 0.01% | 23,364 |
| +2% / −0.25% | 9.6% | 10.5% | 85.7% | 3.8% | 9.9% | 10.9% | 97% | 96% | -1.0 pp | -1.4 pp | -2.4 pp | not computed: this window is too short for two halves to detect instability | 0.15% | 0.26% | 23,364 |
| +2% / −0.5% | 16.7% | 18.0% | 74.3% | 7.7% | 17.9% | 19.5% | 93% | 92% | -1.6 pp | -2.3 pp | -3.9 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.19% | 23,364 |
| +2% / −1% | 26.6% | 28.1% | 58.0% | 13.9% | 31.2% | 32.6% | 85% | 86% | -1.4 pp | -3.4 pp | -4.8 pp | not computed: this window is too short for two halves to detect instability | 0.07% | 0.09% | 23,364 |
| +2% / −2% | 36.9% | 36.4% | 36.9% | 26.7% | 50.3% | 49.6% | 73% | 73% | +0.7 pp | -3.9 pp | -3.3 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.02% | 23,364 |
| +2% / −3% | 40.1% | 39.7% | 22.6% | 37.6% | 62.2% | 63.7% | 64% | 62% | -1.5 pp | -4.3 pp | -5.9 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.00% | 23,364 |
| +3% / −0.25% | 5.8% | 6.7% | 88.2% | 5.1% | 6.1% | 7.1% | 95% | 95% | -1.0 pp | -1.1 pp | -2.1 pp | not computed: this window is too short for two halves to detect instability | 0.03% | 0.12% | 23,364 |
| +3% / −0.5% | 10.1% | 11.8% | 77.8% | 10.4% | 11.3% | 13.1% | 89% | 90% | -1.8 pp | -2.1 pp | -3.9 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.09% | 23,364 |
| +3% / −1% | 16.3% | 18.5% | 62.1% | 19.4% | 20.7% | 22.9% | 78% | 81% | -2.2 pp | -3.2 pp | -5.3 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.05% | 23,364 |
| +3% / −2% | 22.6% | 24.3% | 40.1% | 35.6% | 36.3% | 37.8% | 62% | 64% | -1.5 pp | -4.3 pp | -5.8 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.01% | 23,364 |
| +3% / −3% | 24.9% | 26.8% | 24.9% | 48.3% | 48.2% | 51.8% | 52% | 52% | -3.6 pp | -5.2 pp | -8.9 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.00% | 23,364 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
9 Benjamini–Hochberg survivors in the 1d horizon — not computed: this window is too short for two halves to detect instability, so none can be confirmed here.
Sampling check: re-running this horizon on 258 strictly non-overlapping entries (rather than 23,364 every-bar entries) moves the asymmetry by a median of 5.7 pp and at most 10.2 pp.
105 non-overlapping entries · 121 coverage-span equivalents · median 222 bars per window
At this horizon: 0.25% ≈ 0.04σ · 0.5% ≈ 0.08σ · 1% ≈ 0.16σ · 2% ≈ 0.32σ · 3% ≈ 0.49σ — the same percentage barrier is a far bigger move at 4h than at 5d, which is most of why resolution rates differ across horizons.
Shares of all long entries, not of the ones that resolved: the three add to 100%. “Neither” is the entry still open when the horizon expired — its share is the column below, and across this site it runs from near zero to nearly every entry depending on the horizon and how wide the rungs are. What happened inside those windows is measured in the next section, from intrabar highs and lows.
| target / stop | long target first | long stop first | long neither | excess (long − short) | 95% CI | long effective n |
|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.9% | 56.1% | 0.0% | -5.5 pp not confirmed | -8.2 to -3.0 pp | 3450 |
| +0.25% / −0.5% | 62.0% | 38.0% | 0.0% | -6.1 pp not confirmed | -9.1 to -3.1 pp | 2740 |
| +0.25% / −1% | 77.1% | 22.8% | 0.1% | -5.1 pp not confirmed | -8.1 to -1.9 pp | 2175 |
| +0.25% / −2% | 87.9% | 11.7% | 0.4% | -2.2 pp | -4.7 to +0.4 pp | 1570 |
| +0.25% / −3% | 91.4% | 7.6% | 1.0% | -1.4 pp | -3.7 to +1.1 pp | 1256 |
| +0.5% / −0.25% | 30.3% | 69.7% | 0.0% | -6.0 pp not confirmed | -9.0 to -3.1 pp | 3534 |
| +0.5% / −0.5% | 46.3% | 53.7% | 0.0% | -8.3 pp not confirmed | -12.5 to -3.8 pp | 2080 |
| +0.5% / −1% | 63.2% | 36.4% | 0.4% | -7.7 pp not confirmed | -12.5 to -2.2 pp | 1370 |
| +0.5% / −2% | 78.7% | 20.2% | 1.2% | -3.2 pp | -7.6 to +1.5 pp | 830 |
| +0.5% / −3% | 84.1% | 13.5% | 2.3% | -2.3 pp | -6.7 to +2.1 pp | 659 |
| +1% / −0.25% | 18.6% | 81.3% | 0.1% | -4.9 pp not confirmed | -7.9 to -1.6 pp | 2251 |
| +1% / −0.5% | 31.1% | 68.6% | 0.3% | -7.6 pp not confirmed | -12.3 to -2.1 pp | 1175 |
| +1% / −1% | 47.9% | 51.2% | 0.9% | -6.6 pp | -12.8 to +0.5 pp | 805 |
| +1% / −2% | 65.5% | 31.7% | 2.8% | -2.0 pp | -9.2 to +5.6 pp | 462 |
| +1% / −3% | 72.7% | 22.3% | 4.9% | -1.8 pp | -9.0 to +6.2 pp | 366 |
| +2% / −0.25% | 11.1% | 88.6% | 0.3% | -2.1 pp | -4.5 to +0.5 pp | 1866 |
| +2% / −0.5% | 19.7% | 79.2% | 1.1% | -3.2 pp | -7.6 to +1.5 pp | 940 |
| +2% / −1% | 33.3% | 63.4% | 3.3% | -2.0 pp | -9.3 to +5.6 pp | 573 |
| +2% / −2% | 48.7% | 43.2% | 8.1% | +1.5 pp | -9.6 to +12.0 pp | 306 |
| +2% / −3% | 56.1% | 31.7% | 12.2% | -0.3 pp | -12.8 to +11.6 pp | 233 |
| +3% / −0.25% | 7.8% | 91.1% | 1.0% | -1.3 pp | -3.7 to +1.1 pp | 1415 |
| +3% / −0.5% | 14.2% | 83.0% | 2.8% | -2.2 pp | -6.6 to +2.2 pp | 778 |
| +3% / −1% | 24.3% | 68.8% | 6.9% | -1.8 pp | -8.9 to +6.2 pp | 450 |
| +3% / −2% | 35.9% | 49.4% | 14.8% | -0.3 pp | -12.8 to +11.6 pp | 242 |
| +3% / −3% | 41.6% | 37.4% | 21.0% | -1.7 pp | -16.7 to +12.6 pp | 174 |
| target / stop | long target first | short target first | short stop first | short neither | long, of resolved | short, of resolved | long resolved | short resolved | raw gap | change after detrending | detrended gap | stability | long ties | short ties | n |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +0.25% / −0.25% | 43.9% | 47.4% | 52.6% | 0.0% | 43.9% | 47.4% | 100% | 100% | -3.5 pp | -2.0 pp | -5.5 pp | not computed: this window is too short for two halves to detect instability | 8.67% | 8.67% | 23,364 |
| +0.25% / −0.5% | 62.0% | 66.0% | 34.0% | 0.0% | 62.0% | 66.0% | 100% | 100% | -4.0 pp | -2.1 pp | -6.1 pp | not computed: this window is too short for two halves to detect instability | 3.75% | 3.74% | 23,364 |
| +0.25% / −1% | 77.1% | 80.3% | 19.7% | 0.1% | 77.2% | 80.3% | 100% | 100% | -3.1 pp | -2.0 pp | -5.1 pp | not computed: this window is too short for two halves to detect instability | 1.24% | 1.02% | 23,364 |
| +0.25% / −2% | 87.9% | 88.5% | 11.2% | 0.3% | 88.3% | 88.7% | 100% | 100% | -0.5 pp | -1.7 pp | -2.2 pp | not computed: this window is too short for two halves to detect instability | 0.26% | 0.17% | 23,364 |
| +0.25% / −3% | 91.4% | 91.1% | 7.9% | 1.0% | 92.4% | 92.0% | 99% | 99% | +0.3 pp | -1.7 pp | -1.4 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.03% | 23,364 |
| +0.5% / −0.25% | 30.3% | 34.2% | 65.8% | 0.0% | 30.3% | 34.2% | 100% | 100% | -3.9 pp | -2.1 pp | -6.0 pp | not computed: this window is too short for two halves to detect instability | 3.74% | 3.75% | 23,364 |
| +0.5% / −0.5% | 46.3% | 51.7% | 48.3% | 0.0% | 46.3% | 51.7% | 100% | 100% | -5.4 pp | -2.9 pp | -8.3 pp | not computed: this window is too short for two halves to detect instability | 1.98% | 1.98% | 23,364 |
| +0.5% / −1% | 63.2% | 68.0% | 31.7% | 0.3% | 63.4% | 68.2% | 100% | 100% | -4.8 pp | -3.0 pp | -7.7 pp | not computed: this window is too short for two halves to detect instability | 0.71% | 0.60% | 23,364 |
| +0.5% / −2% | 78.7% | 79.1% | 19.8% | 1.1% | 79.6% | 79.9% | 99% | 99% | -0.3 pp | -2.9 pp | -3.2 pp | not computed: this window is too short for two halves to detect instability | 0.19% | 0.12% | 23,364 |
| +0.5% / −3% | 84.1% | 83.0% | 14.2% | 2.8% | 86.1% | 85.4% | 98% | 97% | +0.7 pp | -3.0 pp | -2.3 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.02% | 23,364 |
| +1% / −0.25% | 18.6% | 21.5% | 78.3% | 0.1% | 18.7% | 21.5% | 100% | 100% | -2.9 pp | -2.0 pp | -4.9 pp | not computed: this window is too short for two halves to detect instability | 1.02% | 1.24% | 23,364 |
| +1% / −0.5% | 31.1% | 35.7% | 63.9% | 0.4% | 31.2% | 35.9% | 100% | 100% | -4.6 pp | -2.9 pp | -7.6 pp | not computed: this window is too short for two halves to detect instability | 0.60% | 0.71% | 23,364 |
| +1% / −1% | 47.9% | 50.9% | 48.2% | 0.9% | 48.3% | 51.4% | 99% | 99% | -3.0 pp | -3.6 pp | -6.6 pp | not computed: this window is too short for two halves to detect instability | 0.30% | 0.30% | 23,364 |
| +1% / −2% | 65.5% | 63.3% | 33.4% | 3.3% | 67.4% | 65.5% | 97% | 97% | +1.9 pp | -3.9 pp | -2.0 pp | not computed: this window is too short for two halves to detect instability | 0.09% | 0.07% | 23,364 |
| +1% / −3% | 72.7% | 68.8% | 24.3% | 6.9% | 76.5% | 73.9% | 95% | 93% | +2.6 pp | -4.4 pp | -1.8 pp | not computed: this window is too short for two halves to detect instability | 0.05% | 0.01% | 23,364 |
| +2% / −0.25% | 11.1% | 11.4% | 88.1% | 0.4% | 11.1% | 11.5% | 100% | 100% | -0.4 pp | -1.7 pp | -2.1 pp | not computed: this window is too short for two halves to detect instability | 0.17% | 0.26% | 23,364 |
| +2% / −0.5% | 19.7% | 20.0% | 78.9% | 1.2% | 19.9% | 20.2% | 99% | 99% | -0.3 pp | -2.9 pp | -3.2 pp | not computed: this window is too short for two halves to detect instability | 0.12% | 0.19% | 23,364 |
| +2% / −1% | 33.3% | 31.6% | 65.6% | 2.8% | 34.5% | 32.5% | 97% | 97% | +1.9 pp | -3.9 pp | -2.0 pp | not computed: this window is too short for two halves to detect instability | 0.07% | 0.09% | 23,364 |
| +2% / −2% | 48.7% | 43.2% | 48.8% | 8.1% | 53.0% | 47.0% | 92% | 92% | +6.0 pp | -4.5 pp | +1.5 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.02% | 23,364 |
| +2% / −3% | 56.1% | 49.4% | 35.9% | 14.8% | 63.9% | 57.9% | 88% | 85% | +6.0 pp | -6.3 pp | -0.3 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.00% | 23,364 |
| +3% / −0.25% | 7.8% | 7.4% | 91.6% | 1.0% | 7.9% | 7.5% | 99% | 99% | +0.4 pp | -1.7 pp | -1.3 pp | not computed: this window is too short for two halves to detect instability | 0.03% | 0.12% | 23,364 |
| +3% / −0.5% | 14.2% | 13.5% | 84.2% | 2.3% | 14.6% | 13.8% | 97% | 98% | +0.8 pp | -3.0 pp | -2.2 pp | not computed: this window is too short for two halves to detect instability | 0.02% | 0.09% | 23,364 |
| +3% / −1% | 24.3% | 22.3% | 72.8% | 4.9% | 26.1% | 23.4% | 93% | 95% | +2.7 pp | -4.5 pp | -1.8 pp | not computed: this window is too short for two halves to detect instability | 0.01% | 0.05% | 23,364 |
| +3% / −2% | 35.9% | 31.7% | 56.1% | 12.2% | 42.1% | 36.1% | 85% | 88% | +6.0 pp | -6.3 pp | -0.3 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.01% | 23,364 |
| +3% / −3% | 41.6% | 37.4% | 41.6% | 21.0% | 52.7% | 47.3% | 79% | 79% | +5.4 pp | -7.0 pp | -1.7 pp | not computed: this window is too short for two halves to detect instability | 0.00% | 0.00% | 23,364 |
The three gap columns add: raw + change after detrending = detrended gap. Only the detrended gap is tested and only it carries the interval in the table above. “Of resolved” is the target-first rate among entries that reached either barrier — on cells where little resolves it can read near 100% while the unconditional rate is modest.
8 Benjamini–Hochberg survivors in the 3d horizon — not computed: this window is too short for two halves to detect instability, so none can be confirmed here.
Sampling check: re-running this horizon on 105 strictly non-overlapping entries (rather than 23,364 every-bar entries) moves the asymmetry by a median of 6.2 pp and at most 15.9 pp.
Insufficient data — 53 non-overlapping 5d entries in this coverage span, below the floor of 100. A hit rate estimated on fewer cannot distinguish the deviations this statistic exists to detect.
Across the 3 horizons this window publishes, 41 of 47 cells that look significant uncorrected survive Benjamini–Hochberg at q=0.05, and 0 of those 41 Benjamini–Hochberg survivors also agree across both halves of history — survives Benjamini-Hochberg AND both halves of history agree in sign. Only confirmed cells are shown in bold and ringed on the charts; a cell that is significant but not confirmed keeps its colour, its number and its interval, marked “not confirmed” rather than hidden. The gaps between all three counts are themselves the point.
Not every window in the barrier grid above resolves: especially at shorter horizons and wider rungs, many windows touch neither barrier before the horizon runs out. This section measures what those windows did anyway, resolved or not — how far price ran in the trader's favour (MFE) and how much heat it took against them (MAE) — from intrabar high and low, never from closes: a stop is hit by the low, not the close. See methodology.
MFE is signed positive and MAE is signed negative. For MFE the higher quantile is the better outcome; for MAE it is the opposite — the lower (more negative) quantile is the worse drawdown, and the higher quantile is the mildest one. The tables below label each MAE column by what it means, not just by its quantile, so the two are never read the same way by mistake.
At the 4h horizon, a +1% / −1% bracket resolves — hits target or stop — in only 59.9% of long entries; the other 40.1% touch neither barrier inside the horizon. Across every 4h window regardless of whether it resolved, the median run in the trader's favour (MFE) was +0.60% and the best tenth of windows reached +2.15%, while the median heat taken against them (MAE) was -0.66% and the worst tenth of windows drew down -2.32% — that is what the rest of the grid's “neither” column was doing instead of resolving.
1,540 non-overlapping entries · 2,184 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 1,540 | +0.60% | +2.15% | -0.66% | -2.32% |
| Short | 1,540 | +0.66% | +2.38% | -0.59% | -2.10% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 17,637 | 31.7% | 6397 |
| +0.25% / −0.5% | 17,637 | 16.8% | 2799 |
| +0.25% / −1% | 17,637 | 5.7% | 1661 |
| +0.25% / −2% | 17,637 | 1.0% | 2297 |
| +0.25% / −3% | 17,637 | 0.3% | 2956 |
| +0.5% / −0.25% | 13,025 | 43.3% | 2848 |
| +0.5% / −0.5% | 13,025 | 23.9% | 1453 |
| +0.5% / −1% | 13,025 | 8.5% | 999 |
| +0.5% / −2% | 13,025 | 1.5% | 1278 |
| +0.5% / −3% | 13,025 | 0.5% | 1935 |
| +1% / −0.25% | 7,455 | 52.1% | 1217 |
| +1% / −0.5% | 7,455 | 30.7% | 656 |
| +1% / −1% | 7,455 | 12.1% | 509 |
| +1% / −2% | 7,455 | 2.4% | 575 |
| +1% / −3% | 7,455 | 0.8% | 1028 |
| +2% / −0.25% | 2,675 | 57.8% | 484 |
| +2% / −0.5% | 2,675 | 36.3% | 259 |
| +2% / −1% | 2,675 | 17.0% | 175 |
| +2% / −2% | 2,675 | 5.1% | 181 |
| +2% / −3% | 2,675 | 2.1% | 323 |
| +3% / −0.25% | 974 | 59.1% | 155 |
| +3% / −0.5% | 974 | 38.8% | 95 |
| +3% / −1% | 974 | 20.6% | 68 |
| +3% / −2% | 974 | 8.0% | 108 |
| +3% / −3% | 974 | 4.0% | 131 |
258 non-overlapping entries · 364 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 258 | +1.64% | +5.10% | -1.51% | -5.57% |
| Short | 258 | +1.53% | +5.90% | -1.62% | -4.86% |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 20,951 | 41.8% | 4385 |
| +0.25% / −0.5% | 20,951 | 27.4% | 7140 |
| +0.25% / −1% | 20,951 | 14.0% | 2642 |
| +0.25% / −2% | 20,951 | 4.4% | 1998 |
| +0.25% / −3% | 20,951 | 1.6% | 1757 |
| +0.5% / −0.25% | 18,722 | 58.2% | 3889 |
| +0.5% / −0.5% | 18,722 | 41.0% | 3196 |
| +0.5% / −1% | 18,722 | 22.7% | 1353 |
| +0.5% / −2% | 18,722 | 7.3% | 1089 |
| +0.5% / −3% | 18,722 | 2.9% | 990 |
| +1% / −0.25% | 15,212 | 70.9% | 2762 |
| +1% / −0.5% | 15,212 | 53.5% | 1241 |
| +1% / −1% | 15,212 | 31.4% | 756 |
| +1% / −2% | 15,212 | 11.0% | 537 |
| +1% / −3% | 15,212 | 4.6% | 523 |
| +2% / −0.25% | 10,061 | 77.5% | 1246 |
| +2% / −0.5% | 10,061 | 60.9% | 610 |
| +2% / −1% | 10,061 | 38.0% | 375 |
| +2% / −2% | 10,061 | 14.3% | 254 |
| +2% / −3% | 10,061 | 6.9% | 316 |
| +3% / −0.25% | 6,361 | 78.7% | 760 |
| +3% / −0.5% | 6,361 | 62.7% | 425 |
| +3% / −1% | 6,361 | 40.2% | 259 |
| +3% / −2% | 6,361 | 17.0% | 181 |
| +3% / −3% | 6,361 | 8.4% | 208 |
105 non-overlapping entries · 121 coverage-span equivalents.
| Side | non-overlapping entries | MFE median | MFE p90 — best case | MAE median — typical heat | MAE p10 — worst-case heat |
|---|---|---|---|---|---|
| Long | 105 | +2.97% | — | -2.45% | — |
| Short | 105 | +2.51% | — | -2.88% | — |
Long-side reading only — there is no short-side version of this table. Of long entries whose maximum favourable excursion reached at least the target rung, what share drew down past the stop rung before the target was reached — i.e. would the stop already have closed a trade that went on to work. A drawdown that happens after the target was reached does not count: the position was already closed by then. Keyed by the same target/stop rungs as the barrier grid above, so the two tables read side by side. See methodology.
| target / stop | qualifying long entries | share drew down past stop first | effective n |
|---|---|---|---|
| +0.25% / −0.25% | 22,167 | 44.6% | 3109 |
| +0.25% / −0.5% | 22,167 | 30.7% | 5857 |
| +0.25% / −1% | 22,167 | 17.4% | 4594 |
| +0.25% / −2% | 22,167 | 7.1% | 2557 |
| +0.25% / −3% | 22,167 | 3.5% | 2354 |
| +0.5% / −0.25% | 20,962 | 62.1% | 2979 |
| +0.5% / −0.5% | 20,962 | 46.2% | 2968 |
| +0.5% / −1% | 20,962 | 28.8% | 2198 |
| +0.5% / −2% | 20,962 | 12.1% | 1255 |
| +0.5% / −3% | 20,962 | 6.1% | 1152 |
| +1% / −0.25% | 18,877 | 75.7% | 3189 |
| +1% / −0.5% | 18,877 | 60.7% | 1879 |
| +1% / −1% | 18,877 | 40.4% | 1292 |
| +1% / −2% | 18,877 | 18.8% | 617 |
| +1% / −3% | 18,877 | 9.9% | 556 |
| +2% / −0.25% | 15,129 | 82.6% | 2338 |
| +2% / −0.5% | 15,129 | 69.4% | 1178 |
| +2% / −1% | 15,129 | 48.4% | 639 |
| +2% / −2% | 15,129 | 24.7% | 332 |
| +2% / −3% | 15,129 | 13.4% | 277 |
| +3% / −0.25% | 11,579 | 84.1% | 1325 |
| +3% / −0.5% | 11,579 | 71.4% | 768 |
| +3% / −1% | 11,579 | 51.0% | 390 |
| +3% / −2% | 11,579 | 27.6% | 209 |
| +3% / −3% | 11,579 | 16.0% | 186 |
Insufficient data — 53 non-overlapping 5d entries in this coverage span, below the floor of 100. An excursion quantile estimated on fewer cannot distinguish real heat from noise.
| Horizon | median move | p90 move | p99 move | non-overlapping entries | n | spread as share of median move |
|---|---|---|---|---|---|---|
| 4h | 0.60% | 2.15% | — | 1,540 | 23,310 | — |
| 1d | 1.63% | 5.17% | — | 258 | 23,310 | — |
| 3d | 2.62% | — | — | 105 | 23,362 | — |
| 5d | insufficient data (53 non-overlapping entries) | |||||
No spread configured for this asset, so cost is not reported.
Sigma may be scaled by the square root of time; medians, stop distances and targets may not. Variance grows in proportion to time, so σ — its square root — grows with the square root of time. The typical move grows faster than that, so scaling a median from a short horizon to a longer one sets it too tight — on every asset and window this site publishes. The tails do not miss in one direction the way the median does, so a stop distance or a target is read off the measured table above, never scaled to it.
| Horizon | q | √t factor from 4h | median move, √t prediction | median move, measured | error | median ÷ σ | kurtosis | non-overlapping entries | n |
|---|---|---|---|---|---|---|---|---|---|
| 4h | 16 bars | ×1.00 | 0.600% | 0.600% | +0.0% | 0.398 | 20.2 | 1,540 | 23,310 |
| 1d | 92 bars | ×2.40 | 1.438% | 1.631% | +13.4% | 0.472 | 12.5 | 258 | 23,310 |
| 3d | 222 bars | ×3.72 | 2.234% | 2.622% | +17.4% | 0.475 | 8.0 | 105 | 23,362 |
| 5d | insufficient data — the horizon-move table did not publish this horizon | ||||||||
A positive error means the real move is larger than √t predicts. The error is relative — (measured − predicted) ÷ predicted — not itself a move size in percent of price, even though it sits between two columns that are. The reason a positive error is the norm is in the last two columns: aggregation pulls the distribution toward normal, where median ÷ σ is 0.674 and kurtosis is 3.0. By the longest horizon this table actually publishes (3d), median ÷ σ is 0.475 and kurtosis is 8.0 — read those against the Gaussian reference above rather than assuming the shape has normalised by then.
VR(q) = Var(q-bar move) ÷ (q × Var(1-bar move)). Above 1 moves extend and a distant target is more reachable than its distance suggests; below 1 they retrace. The null is exactly 1. q is derived from this asset's own bar calendar, so “5d” is however many bars a five-calendar-day window actually contains — usually fewer than five trading days, because it swallows a weekend.
| Horizon | q | VR | 95% interval | as move size vs random walk | robust z | p (robust) | blocks | homoskedastic z | verdict | split-half VR (first / second) |
|---|---|---|---|---|---|---|---|---|---|---|
| 4h | 16 bars ≈ 0.17 trading days | 0.957 | 0.811 to 1.103 | 90% to 105% | -0.58 | 0.565 | 1,471 | -1.49 | not significant | not computed: this window is too short for two halves to detect instability |
| 1d | 92 bars ≈ 1.00 trading days | 1.032 | 0.723 to 1.341 | 85% to 116% | +0.20 | 0.840 | 255 | +0.44 | not significant | not computed: this window is too short for two halves to detect instability |
| 3d | 222 bars ≈ 2.41 trading days | 1.008 | 0.567 to 1.449 | 75% to 120% | +0.04 | 0.972 | 106 | +0.07 | not significant | not computed: this window is too short for two halves to detect instability |
| 5d | insufficient data — 85 non-overlapping q-blocks at this horizon, below the floor of 100 | |||||||||
Only the robust z (and its p, the column so labelled) is tested; it allows the size of moves to change over time, which this asset's does. The homoskedastic z column assumes it does not, and is shown to make the gap visible: 0 of these horizons would look significant under it against 0 that survive Benjamini–Hochberg on the robust one. The two counts do not diverge on this window, so there is no such gap here for volatility clustering to have manufactured.
The verdict column applies one rule and states it in full: survives Benjamini-Hochberg on the robust standard error AND both halves of history agree in sign of (VR - 1). The split-half column is the second half of that rule shown as data — the same VR recomputed on each disjoint half of the window, at the same q. The comparison is on the sign of (VR − 1), not of VR: VR is always positive, so comparing raw signs would mark every pair as agreeing and the check could never fail. Halves that disagree do not overturn a significant result on their own; they stop it being called confirmed.
How far price moved while you could not trade. Split three ways and never pooled: a venue halt and a weekend are different phenomena and only one of them carries real overnight risk. No stop protects you across either.
| Kind | n | median | p90 | p99 | >1× median bar | >3× |
|---|---|---|---|---|---|---|
| session | 0 | insufficient data | ||||
| weekend | 26 | insufficient data | ||||
| daily break | 113 | 0.029 | 0.156 | 0.441 | 3.5% | 0.0% |
Gaps are in USD/oz, signed. Median bar range for comparison: 0.239.
Kept for placing an entry, not as a headline: the horizon this site reports at is hours to days, and a multi-day stop sized to a 15-minute range is the error that shakes a trader out of positions that would have worked.
| Unit | n | mean | p10 | p25 | p50 | p75 | p90 | p95 | p99 |
|---|---|---|---|---|---|---|---|---|---|
| USD/oz | 23364 | 0.3476 | 0.0710 | 0.1320 | 0.2390 | 0.4232 | 0.7060 | 0.9740 | 1.9084 |
| % of price | 23364 | 0.4924 | 0.1519 | 0.2330 | 0.3708 | 0.5894 | 0.9251 | 1.2484 | 2.3038 |
| Unit | n | mean | p10 | p25 | p50 | p75 | p90 | p95 | p99 |
|---|---|---|---|---|---|---|---|---|---|
| USD/oz | 23224 | 0.3479 | 0.0710 | 0.1320 | 0.2400 | 0.4240 | 0.7057 | 0.9730 | 1.9090 |
| % of price | 23224 | 0.4925 | 0.1516 | 0.2330 | 0.3710 | 0.5898 | 0.9248 | 1.2446 | 2.3063 |