AVANT INSIDER SIGNAL

The cluster-buy signal

Avant Signal watches for a specific Form 4 pattern: qualifying open-market purchases by multiple reported insiders in the same company.

See the latest delayed clusters Read the full research

ONE PRECISE DEFINITION

How the live cluster methodology works

A cluster-buy signal forms when qualifying Form 4 purchases with transaction code P, each valued at least $100,000, bring the same ticker from fewer than two to at least two distinct reported insider names.

Qualifying filings may be no more than 11 calendar days apart, based on their filing dates. The signal forms when the later qualifying filing makes the pattern observable. Its formation time is the SEC acceptance time of that filing.

The public preview contains only published delayed clusters from the period where our source data meets the required completeness checks. Each cluster is published after three completed NYSE sessions. A qualifying purchase filed after the cluster forms has its own three-completed-session publication delay before it can appear. Records that fail source, eligibility, integrity, completeness, or publication checks are omitted, so the preview is not exhaustive market coverage. A cluster is a disclosed activity pattern, not evidence of coordination, investment advice, or a forecast.

HISTORICAL RESEARCH

What the historical studies reported

Two related studies measured mean excess returns versus SPY at 5, 10, and 30 days:

Original retrospective study

440 transaction-row observations

5-day mean excess
+2.02%
10-day mean excess
+2.41%
30-day mean excess
+2.29%

Formation-aligned study

1,328 formation events

5-day mean excess
+1.18%
10-day mean excess
+1.27%
30-day mean excess
+0.10%

How to interpret these numbers

These studies use related but different cohorts and event definitions, so the difference should not be attributed solely to lookahead. They differ in retrospective earliest-member versus observable formation anchoring, transaction-row observations versus one formation event, cohort and insider-role composition, and episode construction and quality selection.

Historical acceptance timestamps were unavailable for the aligned study period, so all aligned events used the first official NYSE regular-session open strictly after filing_date. The historical aligned study is not acceptance-time precise.

The aligned median excess return was approximately flat at 5 and 10 days and was −1.23% at 30 days. Slightly fewer than half of aligned events with valid returns had positive excess returns, so the mean should not be treated as a typical-event prediction. These historical results are not forecasts of future performance.

Read the full research and aligned-study addendum.

PUBLISHED DELAYED CLUSTERS

Latest published delayed clusters

  2. HEI, HEI.A

    Cluster filing date

    Qualifying purchases
    8
    Distinct reported insiders
    8

    Reported insiders

    • CHERUVATATH NANDAKUMAR
    • HENRIQUES ADOLFO
    • SCHRIESHEIM ALAN
    • Fine Carol F.
    • Neitzel Julie
    • Hildebrandt Mark H
    • CULLIGAN THOMAS M
    • MENDELSON VICTOR H

    Source SEC filings

  5. BBD

    Cluster filing date

    Qualifying purchases
    27
    Distinct reported insiders
    3

    Reported insiders

    • Di Marcello Francesco
    • Alvarez Denise Aguiar
    • Alvarez Rubens Aguiar

    Source SEC filings

Continue in Explore

Review delayed Form 4 transaction context with the existing Cluster only filter applied.

Explore delayed clusters

AVANT SIGNAL

Get the cluster-buy signal when it forms.

The public clusters above are delayed. Signal is the existing plan for the real-time cluster-buy signal.

This exact signal in real time — $19.99/month
Chat with Shawn the Founder