Screener
Single-name candidates ranked by confluence · as of 2026-09-11
The screener reads the full S&P 900 universe, runs every name through each dislocation archetype, and shows every name that fired an archetype today (the full candidate log).How the screener works →
Longs (3)
182.87 → 188.52+3.1% vs SPY
- ·locked 2026-09-01 (evening)
- ·entry 182.87 @ 2026-09-02
- ·invalidates: close < 175.38
day 6/20
85.36 → 84.66+0.2% vs SPY
- ·locked 2026-09-02 (evening)
- ·entry 85.36 @ 2026-09-03
- ·invalidates: close < 81.10
day 5/20
45.13 → —
- ·locked 2026-09-10 (evening)
- ·entry 45.13 @ 2026-09-11
- ·invalidates: close < 42.92
day 0/20
Shorts (2)
17.23 → 16.44+3.6% vs SPY
- ·locked 2026-09-02 (evening)
- ·entry 17.23 @ 2026-09-03
- ·invalidates: close > 17.83
day 5/5
176.39 → 172.40+1.3% vs SPY
- ·locked 2026-09-02 (evening)
- ·entry 176.39 @ 2026-09-03
- ·invalidates: close > 182.39
day 5/5
Competing for open slots (4)· open: 0L / 0S · locks at the next screener run
Board-pick snapshot: longs 44% (n=228) · shorts 51% (n=147) (n=375)
1-2 week swing, not day-trading · calm-conditional · graded, not advice ·what the badges mean
What Company Action Board — the pick-contract position book: picks lock on a ripe + slot win and hold to horizon or invalidation, immune to day-to-day trigger flicker.
So-what Use as a watchlist filter, not a sizing directive. The countdown chip is the hold clock; the stop chip is the mechanical invalidation; competing names lock only when a slot frees.
Each archetype earns its sizing tier from realized entry-day cohorts — every name it flagged on the same day is graded as ONE observation, because same-day picks move together. PROBATION at 60%, BETTABLE at 80% plus 8 cohorts earned live after the ledger lock (2026-07-28).
- · Evidence is counted in ENTRY-DAY COHORTS, never episodes — names flagged the same day move together and are not independent bets.
- · Imported pre-lock evidence is CALM-REGIME ONLY: every historical cohort was entered in a single low-drawdown stretch, so no archetype here has been tested in its own bad weather.
- · Imported cohorts are recorded with a provenance weight (see n_backfill vs n_live); BETTABLE cannot be reached on imported evidence at all — it requires matured cohorts earned after the ledger lock.
- · Tiers inform manual sizing only. Nothing on this board places, sizes, or closes an order.
Contract-graded record — each pick held to its ACTUAL exit (horizon / price-invalidation / adverse-regime), market-adjusted vs SPY. The strictest board lens: a live, growing ledger over a calm regime — not a forecast, not a proven track record. The credibility label is a display state; it licenses no sizing.
Closed record — held to exit
Contract-graded record — each locked pick graded as a held position to its exit · graded, not advice
Candidates
0
filtered: positioning_extreme
Universe
906
S&P 900 members (500+400)
Archetypes
0
none
As of
2026-09-11
screener trade date
not active — macro drawdown read is low (score 44.0); the hedge shortlist lights up only when the macro drawdown read is elevated (['elevated', 'high'])
Archetype calibration
per-horizon · market-adjustedForward-grading (locked 2026-06-23). This is a young, single-episode substrate — the apparent in-sample edge is mostly sector beta, not yet stock-selection skill. Rates read at one horizon per pill (never blended across horizons), excess over SPY, intent-aware. First out-of-sample checkpoint ~2026-07-21. Calibrated rows are solid; still-accruing rows are dimmed.How this is graded →
286966 predictions accruing · first maturations ~Jul 6 · hit-rates unlock at 30 matured per archetype.
1N2Overextended Blowoffshort76.5% @5d-1.56% meann=17calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 62.5% | -0.61% | 16 | 10–50% |
| 3d | 62.5% | -1.55% | 16 | 23–67% |
| 5d | 76.5% | -1.56% | 17 | 31–74% |
| 10d | 62.5% | -1.49% | 16 | 28–72% |
| 20d | 64.3% | -2.51% | 14 | 16–61% |
| 30d | 63.6% | -2.44% | 11 | 15–65% |
| 45d | 80.0% | -4.10% | 5 | 12–77% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
2N1Distribution Topshort51.7% @5d-0.06% meann=516calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 51.5% | -0.14% | 515 | 47–56% |
| 3d | 52.8% | -0.13% | 494 | 48–57% |
| 5d | 51.7% | -0.06% | 516 | 45–53% |
| 10d | 54.3% | -0.49% | 484 | 40–49% |
| 20d | 59.2% | -1.08% | 449 | 39–48% |
| 30d | 67.4% | -2.19% | 362 | 39–49% |
| 45d | 71.2% | -3.39% | 267 | 37–49% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
3N3Failed Breakoutshort54.9% @5d-0.34% meann=1865calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 58.6% | -0.42% | 1947 | 44–49% |
| 3d | 55.4% | -0.27% | 1819 | 46–50% |
| 5d | 54.9% | -0.34% | 1865 | 43–47% |
| 10d | 54.6% | -0.47% | 1761 | 45–49% |
| 20d | 55.6% | -0.57% | 1540 | 46–51% |
| 30d | 55.2% | -0.46% | 1278 | 46–52% |
| 45d | 61.2% | -1.25% | 899 | 47–54% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
4N6PEAD Miss Driftshort45.2% @5d1.12% meann=31calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 5d | 45.2% | 1.12% | 31 | 14–43% |
| 10d | 58.1% | 0.39% | 31 | 21–53% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
5P1Pair Divergencelong54.0% @5d0.19% meann=150calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 48.7% | -0.11% | 152 | 38–53% |
| 3d | 50.0% | -0.13% | 150 | 36–51% |
| 5d | 54.0% | 0.19% | 150 | 38–54% |
| 10d | 53.5% | 0.03% | 144 | 39–55% |
| 20d | 53.0% | 0.09% | 134 | 40–56% |
| 30d | 50.0% | 0.63% | 124 | 45–63% |
| 45d | 37.2% | -1.38% | 43 | 35–63% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
6A2Oversold Sympathylong49.1% @5d0.08% meann=1015calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 47.5% | -0.05% | 1039 | 52–58% |
| 3d | 48.7% | -0.14% | 975 | 47–54% |
| 5d | 49.1% | 0.08% | 1015 | 47–54% |
| 10d | 48.2% | 0.15% | 956 | 50–56% |
| 20d | 48.9% | 0.09% | 838 | 52–58% |
| 30d | 49.9% | 0.35% | 690 | 48–55% |
| 45d | 50.1% | -0.03% | 499 | 44–53% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
7N5Revision Collapseshort51.2% @5d0.70% meann=41calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 52.4% | 0.44% | 42 | 31–60% |
| 3d | 58.5% | 0.30% | 41 | 34–64% |
| 5d | 51.2% | 0.70% | 41 | 32–61% |
| 10d | 41.0% | 2.15% | 39 | 48–77% |
| 20d | 38.5% | 1.22% | 39 | 36–66% |
| 30d | 51.4% | 3.31% | 37 | 38–69% |
| 45d | 36.8% | 8.01% | 19 | 41–81% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
8A8PEAD Beat Driftlong49.3% @5d0.01% meann=227calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 5d | 49.3% | 0.01% | 227 | 48–61% |
| 10d | 45.1% | -0.71% | 226 | 47–60% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
9A4Pre-Catalyst Coilinglonggraded at partial40.0% @5d-0.38% meann=610calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 42.8% | -0.27% | 677 | 58–65% |
| 3d | 38.7% | -0.47% | 604 | 56–64% |
| 5d | 40.0% | -0.38% | 610 | 58–66% |
| 10d | 42.0% | -0.52% | 512 | 53–61% |
| 20d | 41.2% | -0.55% | 376 | 50–60% |
| 30d | 44.4% | -0.11% | 284 | 42–53% |
| 45d | 40.8% | -1.62% | 213 | 43–56% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
10A5Mechanical Reversionlong37.6% @5d-0.51% meann=359calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 37.8% | -0.81% | 270 | 44–56% |
| 3d | 24.1% | -1.84% | 270 | 37–49% |
| 5d | 37.6% | -0.51% | 359 | 42–53% |
| 10d | 41.6% | -0.79% | 358 | 46–56% |
| 20d | 42.6% | -1.98% | 357 | 44–55% |
| 30d | 43.8% | -1.03% | 356 | 43–53% |
| 45d | 43.1% | -1.76% | 355 | 51–61% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
11N4Structural-Derate Continuationshort28.6% @5d2.15% meann=7calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 50.0% | -0.19% | 6 | 19–81% |
| 3d | 50.0% | -0.15% | 6 | 19–81% |
| 5d | 28.6% | 2.15% | 7 | 8–64% |
| 10d | 42.9% | 0.99% | 7 | 16–75% |
| 20d | 28.6% | 4.71% | 7 | 8–64% |
| 30d | 20.0% | 4.50% | 5 | 4–62% |
| 45d | 25.0% | 1.66% | 4 | 5–70% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
12A3Positioning Extremelong31.1% @5d-2.21% meann=61calibrated›
| Horizon | Excess hit | Mean excess | n | 95% CI |
|---|---|---|---|---|
| 1d | 36.9% | -0.61% | 65 | 32–55% |
| 3d | 29.5% | -2.10% | 61 | 60–82% |
| 5d | 31.1% | -2.21% | 61 | 60–82% |
Experimental — confidence dials (not validated, watched forward) ›
The screener's own dials vs excess hit-rate, at the 5d horizon. Reported factually for forward grading — no edge is asserted.
·A1Post-Earnings Overreactionlongaccruing (0/30)~Sep 11accruing›
No matured horizons yet — 7 predictions recorded, 7 pending.
·A6Vol Mispricinglongaccruing (0/30)~Sep 11accruing›
No matured horizons yet — 637 predictions recorded, 601 pending.
Excess = return minus SPY over the same window; hit = excess > 0 for long archetypes, < 0 for short. An empty bucket shows "—", not 0%.
No screener candidates yet.
The screener runs twice each trading day (10:30 and 17:30 ET); the board populates after the first run.
You are filtering by archetype positioning_extreme. Clear filter to see all.
Screener runs twice each trading day (10:30 and 17:30 ET). Candidates decay after their per-archetype relevance window. Sort and archetype filters update via URL params (SSR re-fetch).