Interaction Miner#219tier 2experimentalNew

`combo analyst revision x pairs cointegration`

cadence: Dailydata: mediumlong short

paper

2026

Source

interaction-miner discovery (2026-05-20): analyst_revision_jump x pairs_cointegration, lift_ratio=1.297

Citation only — paper link pending.

What it checks

Big gap up plus the stock is unusually cheap vs its sector ETF = double-confirmed drift entry.

Mechanism

A gap-without-revision on a stock that's also dislocated from its sector ETF (high |pair-cointegration z-score|) sets up an unusually-strong drift candidate. Two mechanisms point the same direction: the gap reflects new private information (So-Wang 2023 JAR) and the cointegration spread reflects temporary mispricing vs the sector benchmark (Gatev-Goetzmann-Rouwenhorst 2006). pair_corr=-0.086 (very orthogonal).

No production champion data for this family yet. Stats appear once the discovery pipeline promotes at least one strategy with this family tag, or once a multi-family blend that includes it earns a champion slot.

Signal rule

position = clip(1.5 * (z(gap_signed_score) + z(-pair_cointeg_z)), -1, +1)

Data dependencies

daily_prices
Adjusted-close OHLCV for every US-listed ticker; primary price feed.
sector_etf_prices
Worker data table — see services/worker schema.

Expected edge

Paper alpha: lift=1.297x over either signal alone
Paper window: 10-day forward

Combined IC=0.275 vs 0.21/0.20 for either alone - lift 1.30x.

Related families

analyst revision jumpSentiment

So & Wang (2023) JAR "News-Implied Analyst Revisions and Drift": a large overnight gap on day t that isn't preceded by an analyst revision is mispriced — the revision arrives around T+5 and the price continues to drift through it. Reported: Sharpe 1.5, 30-day drift, robust 2003-2022. v1 approximation flags gaps ≥3% with no earnings event in trailing 5d as "unrevised"; long on upside gaps, short on downside gaps, hold 21-63d.

pairs cointegrationPairs

Engle-Granger 2-step cointegration test (Gatev-Goetzmann-Rouwenhorst 2006): tests the stock vs its sector ETF for cointegration before trading the spread. Without cointegration the z-score spread is spurious and pairs trading breaks down. Run cointegration on the first 70% of history; only trade the spread signal on the remaining holdout if the p-value < 0.05.

Explore `combo analyst revision x pairs cointegration` on alphactor.ai

See which tickers this family is currently firing on, with live signals and rankings.

Open universe with this filter →Browse all 251 families

combo analyst revision x pairs cointegration