Pattern Engine · PEAD
PEAD live signals.
Long Setups and Short Setups from the Closelook PEAD pipeline. Each entry passed all three architectural layers — regime, trend, pattern — before becoming a live alert. The 63-day holding window starts at T+3 (the third trading day after the print).
No PEAD signals in the Engine yet.
The Pattern Engine fires PEAD cards Mon–Fri at 11:30 / 20:30 / 21:00 / 21:30 UTC. First batch lands at the evening c-fire crons.
-
→ PEAD NONE — ↗ T+5CAR — vs SPY Hold through — Indices RUBIN Aug 25, 11:00 PM
T+3 said ↓SHORT at -20.24% CAR; T+5 extends to -24.49%. Drift survived the classical post-earnings window.
-
↑ DE PEAD LONG 70 T+3CAR +9.03% vs SPY Hold through Nov 17, 2026 Indices HALO Aug 25, 08:31 PM
Deere & Co reported Q3 FY2026 EPS of $5.10 on 2026-08-20, beating the $4.79 consensus by $0.31 (+6.5%).
-
↑ NDSN PEAD LONG 80 T+3CAR +10.29% vs SPY Hold through Nov 16, 2026 Indices RUBIN Aug 25, 11:31 AM
Nordson reported Q3 FY2026 EPS of $3.25 on 2026-08-19, beating the $3.09 consensus by $0.16 (+5.2%).
-
→ ADI PEAD NONE 0 T+3CAR -0.93% vs SPY Hold through Nov 16, 2026 Indices NASDAQ Aug 25, 11:31 AM
Analog Devices reported Q3 FY2026 EPS of $3.45 on 2026-08-19, beating the $3.33 consensus by $0.12 (+3.6%).
-
→ PEAD NONE — ↗ T+5CAR — vs SPY Hold through — Indices RUBIN · NASDAQ Aug 20, 11:00 PM
T+3 said ↓SHORT at -5.52% CAR; T+5 extends to -8.2%. Drift survived the classical post-earnings window.
-
→ PEAD NONE — ↗ T+5CAR — vs SPY Hold through — Indices NASDAQ Aug 20, 11:00 PM
T+3 said ↓SHORT at -6.53% CAR; T+5 extends to -8.01%. Drift survived the classical post-earnings window.
-
↓ FN PEAD SHORT 80 T+3CAR -20.24% vs SPY Hold through Nov 12, 2026 Indices RUBIN Aug 20, 08:31 PM
Fabrinet reported Q4 FY2026 EPS of $3.92 on 2026-08-17, beating the $3.69 consensus by $0.23 (+6.2%).
-
→ 035720.KO PEAD NONE 4 T+3CAR -2.36% vs SPY Hold through Nov 11, 2026 Indices AW25 Aug 20, 11:32 AM
Kakao reported Q2 2026 EPS of ₩387 on 2026-08-14, missing the ₩388.91 consensus by ₩1.91 (−0.49%); prior-year EPS comparison is not available in our feed.
-
→ PEAD NONE — ↘ T+5CAR — vs SPY Hold through — Indices RUBIN Aug 19, 11:00 PM
T+3 said ↑LONG at +6.64% CAR; T+5 reverses to -12.31%. The market re-priced the print.
-
→ PEAD NONE — ↗ T+5CAR — vs SPY Hold through — Indices NASDAQ Aug 19, 11:00 PM
T+3 said ↓SHORT at -6.53% CAR; T+5 extends to -8.01%. Drift survived the classical post-earnings window.
Hold-horizon legend
- T-1
- Close before the announcement.
- T+0
- Announcement day close (or next trading day for AMC reporters).
- T+3
- Third trading day after the announcement — when the CAR is measured and the signal becomes ready to fire.
- T+63
- ~3 calendar months after T+0. Hard exit per Stage A spec; later stages add adaptive exits.
Candidates, not recommendations
These are systematic candidate signals from a documented academic anomaly. They are not buy or sell recommendations. The Pattern Engine generates the candidates; the decision to act, position sizing, risk management, and portfolio context are entirely the reader's responsibility. Past PEAD spreads have weakened materially since the 1980s and may continue to evolve. Short-side signals carry asymmetric risk including borrow costs and squeeze potential. See the full Lab Read for the complete failure-mode treatment.
FAQ · from the current data
PEAD — live — Q&A
What does the PEAD engine show?
Post-earnings-announcement drift setups from the nightly scan: long and short candidates with the 3-day cumulative abnormal return vs benchmark, the 63-day hold window and ±3% direction thresholds of the Pattern Lab methodology. Methodology: /lab/patterns/pead/.
How often does it run?
Nightly after the US close on trading days, during and after earnings season. Part of the Pattern Engine, the live counterpart of Pattern Lab — five engines on the Signal desk.
Is a signal here a recommendation?
No. Every engine output is a pattern hit in an investment research diary — a watch-list label with its trigger state, never advice. Base rates and the methodology live on the Pattern Lab page.