AAC $9.90 ACHR $2.91 -9.35% ▼ AERO $15.30 -3.35% ▼ AIN $62.20 -4.15% ▼ AIR $101.84 -2.99% ▼ AM $19.73 -1.79% ▼ AMP $480.29 -0.65% ▼ AMTM $20.34 -1.88% ▼ APH $69.72 -7.25% ▼ ASB $29.06 -1.86% ▼ ATI $165.60 -5.01% ▼ ATRO $55.12 -5.50% ▼ AVAV $86.79 -9.45% ▼ AVNW $15.69 -0.06% ▼ AXON $291.14 -5.14% ▼ BAB $25.87 +0.94% ▲ BAH $81.02 -1.97% ▼ BBAI $1.58 -9.18% ▼ BDL $49.02 -5.18% ▼ BELFB $196.38 -10.93% ▼ BHE $55.63 -9.35% ▼ BKSY $16.37 -11.75% ▼ BWXT $108.54 -7.28% ▼ CACI $587.81 -2.89% ▼ CAE $26.71 -1.73% ▼ CDRE $29.45 -3.63% ▼ CICN $0.00 CMTL $1.02 -4.67% ▼ CNC $63.78 -1.83% ▼ CODA $7.50 -3.23% ▼ AAC $9.90 ACHR $2.91 -9.35% ▼ AERO $15.30 -3.35% ▼ AIN $62.20 -4.15% ▼ AIR $101.84 -2.99% ▼ AM $19.73 -1.79% ▼ AMP $480.29 -0.65% ▼ AMTM $20.34 -1.88% ▼ APH $69.72 -7.25% ▼ ASB $29.06 -1.86% ▼ ATI $165.60 -5.01% ▼ ATRO $55.12 -5.50% ▼ AVAV $86.79 -9.45% ▼ AVNW $15.69 -0.06% ▼ AXON $291.14 -5.14% ▼ BAB $25.87 +0.94% ▲ BAH $81.02 -1.97% ▼ BBAI $1.58 -9.18% ▼ BDL $49.02 -5.18% ▼ BELFB $196.38 -10.93% ▼ BHE $55.63 -9.35% ▼ BKSY $16.37 -11.75% ▼ BWXT $108.54 -7.28% ▼ CACI $587.81 -2.89% ▼ CAE $26.71 -1.73% ▼ CDRE $29.45 -3.63% ▼ CICN $0.00 CMTL $1.02 -4.67% ▼ CNC $63.78 -1.83% ▼ CODA $7.50 -3.23% ▼

Monday’s Analyst Calls: Sell-Side Noise or Alpha Signals? Backtest Before You Trade

October 26, 2026 · by stackcalder

Analyst research calls make excellent case studies in why systematic traders build their own signals. Monday’s roundup from 24/7 Wall St. — Block, Boise Cascade, Equifax, FirstEnergy, Lockheed Martin, Lumentum, NetApp, Roper, Spotify, and assorted others — is the market-data equivalent of a momentum-free feed: titles, no timestamps on the underlying information edge, no measureable alpha.

The core problem with sell-side upgrades and downgrades is decay. Event studies on analyst rating changes consistently show the announcement-day drift concentrated in the first minutes after publication — exactly when Python developers running a `yfinance` or Polygon.io websocket get the headline before any retail reader scrolling a roundup does. If your pipeline ingests analyst calls at 9:35 a.m., you’re trading residual noise, not signal. Build the event window properly: pull the timestamped release, estimate abnormal returns against a sector benchmark over [-5, +30] minute windows, and check whether the post-announcement drift survives transaction costs. Most of it doesn’t.

There’s also a selection-bias problem worth quantifying. Aggregated “top analyst calls” lists overrepresent names with recent price action — Block and Spotify appear because they’re volatile enough to generate headlines, not because rating changes there carry more information. A quick sanity check in pandas: compute the correlation between a stock’s 30-day realized volatility and its frequency in these roundups. You’ll find it’s embarrassingly high. Volatility generates coverage; coverage generates clicks; clicks generate roundups. None of it is a factor.

For those who insist on using analyst actions systematically, treat them as a weak feature, not a strategy. Encode rating changes as ordinal events, regress forward returns against the change magnitude with sector and beta controls, and expect an information coefficient in the 0.01–0.03 range at best — indistinguishable from zero once you correct for multiple hypothesis testing across analysts and horizons. Anything a backtest vendor claims beyond that is curve-fitted until demonstrated otherwise.

The trades worth monitoring from this batch are the ones with observable catalysts: Equifax into data-release cycles, Lumentum around optical transceiver order commentary, NetApp on enterprise storage capex guidance. Those map to checkable fundamental series. A rating change on vibes does not.

The uncomfortable truth: if sell-side calls were reliably predictive, the analysts would trade them quietly instead of broadcasting them to a syndication feed.