Pillar 1.5 · Bipartite Resolver

Spoliation Radar

“What isn’t said is often the most damning evidence.”

The Spoliation Radar deploys bipartite edge discounting across fractional multi-channel communication graphs. It autonomously detects unnatural silences — a 5-day communication gap on Slack immediately following a critical internal email — and instantly drafts Rule 37(e) motions for sanctions based on targeted evidence destruction.

How It Works

01

Ingest communication events from all channels: email, Slack, Teams, SMS, and any other source in the discovery production.

02

Classify each event by channel type: one-to-one direct message, group conversation, or broadcast channel.

03

Apply the 1/√N broadcast discount: direct messages carry weight 1.0, broadcast channels carry weight 1/√(member_count). This prevents ambient channel noise from corrupting the baseline.

04

Merge all weighted edges into a single unified bipartite communication graph.

05

Run an isolation forest anomaly detector on the time series of pairwise communication weights between custodian pairs.

06

Gaps exceeding the configurable severity threshold are flagged with per-channel breakdowns and a draft Rule 37(e) motion is generated automatically.

Technical Detail

edge_weight = 1.0 / sqrt(N) // N = recipient count. DM: N=1 → weight=1.0. Channel with 100 members: weight=0.1

HAYES ↔ JENKINS · Aug 8–13, 2025 · 5-day gap

EMAIL: 4/85 expected · SLACK: 8/130 expected · TEAMS: 1/22 expected — 94% volume drop with broadcast discount applied. Rule 37(e) draft generated in 2.3 seconds.

94%

Volume drop detected (Hayes ↔ Jenkins)

5

Day gap across all three channels simultaneously

2.3s

Time to generate Rule 37(e) motion draft