← litmanintelligence.com  |  ← Counsel PDFs index  |  Counsel dashboard

Kfu Email Count Reconciliation

KFU@4PATENT.COM Email Count Reconciliation Memo

Date: 2026-04-06 Audience: Counsel Status: URGENT — must be resolved before BOP filing on 04/02/2026 Author: Evidence pipeline (reconciliation pass)

1. The Discrepancy

Two different figures for the volume of email traffic involving kfu@4patent.com are circulating in the case file:

Figure Source file Stated scope
19,232 messages RESEARCH_LOG.md, line 419 (entry 2026-03-24) "ND0002 archive (2020–2025). Explosion in 2023–2024 (8,733 + 8,948)."
29,862 messages output/BOP_RESPONSE_DRAFT.md, line 97 (Category 2 — Client-Facing Publications) No scope stated. Listed alongside other KFU-related counts.

A separate count (24,526 emails) appears in BOP_RESPONSE_DRAFT.md ¶10 attributed to Martha Long specifically — that figure measures a different population (Martha-as-sender, not all kfu@4patent.com traffic) and is not part of this discrepancy.

2. Source Data Check

Raw grep -c -i "kfu@4patent.com" against the canonical metadata exports (any field — From, To, Cc, Bcc, Subject, etc.):

File Lines (rows) kfu@4patent.com hits
output/EMAIL_METADATA_ND0001.csv 181,571 482
output/EMAIL_METADATA_ND0002.csv 95,330 29,708
Combined 276,901 30,190

Observations: - Neither published figure (19,232 nor 29,862) matches the raw combined total (30,190). - 29,862 is within ~150 of the ND0002-only raw count (29,708). It is almost certainly an ND0002-only, any-field count generated at a slightly different point in time (the full-body extraction was actively rewriting ND0002 during the 03/24 session — see RESEARCH_LOG entry 2026-03-24, "ND0002 at 463MB up from 63MB"). It is not the combined count. - 19,232 is materially smaller than even the ND0002-only raw count (29,708). It is consistent with a narrower scope — most likely From/To only, excluding Cc/Bcc, or possibly deduped on Message-ID. The RESEARCH_LOG explicitly scopes it to "ND0002 archive (2020–2025)." - ND0001 contributes only 482 hits (1.6% of the total), so the ND0001 vs. ND0002+combined distinction is small in absolute terms but explains why nobody has flagged this as a "missing archive" issue.

3. What's Likely Going On

Both numbers are defensible measurements of different things:

  1. 29,862 = ND0002-only, any-field (From/To/Cc/Bcc) message count. This counts every email in which kfu@4patent.com appears anywhere in the headers — i.e., every message the alias touched. Snapshot taken mid-extraction, hence the small drift from the current 29,708.
  2. 19,232 = ND0002-only, narrower scope — most likely direct-recipient only (From or To), excluding bystander Cc traffic. RESEARCH_LOG entry implies this is a recipient-volume count tied to Martha Long → KFU email chains, not the total touch count.

Neither figure includes ND0001 (which adds 482), and neither has been re-run against the post-03/24 full-body-extracted dataset.

A second agent is independently producing the definitive counts. Until those land, treat both numbers as preliminary.

4. Recommendation for the BOP Filing

The BOP is a verified pleading. It cannot contain a number that the underlying CSV will not support on cross-examination.

Recommendation:

  1. Do not file 29,862 as written. It is unscoped in the BOP draft, it does not match the current raw count of either archive individually or combined, and it cannot be reproduced from the metadata files as they stand today.
  2. Do not file 19,232 either without restating its scope. As written in RESEARCH_LOG it implies a smaller universe than what the CSV will yield on a plain grep — opposing counsel will easily find more.
  3. Preferred fix for the 04/02/2026 BOP: Use a conservative, reproducible, scope-defined statement, e.g.:

    "Based on metadata produced from the ND0001 and ND0002 archives (2020–2025), kfu@4patent.com appears as sender, recipient, or copy-recipient on approximately 30,000 messages, the substantial majority of which were sent or received after June 15, 2020. Final counts will be confirmed in discovery."

  4. Lock the methodology before the filing: pick one canonical query (recommended: From OR To OR Cc OR Bcc, both archives, deduped on Message-ID) and re-run it against the current metadata files. Cite the resulting integer in the BOP, RESEARCH_LOG, and every downstream memo. Save the query and result file under output/ so it is reproducible.
  5. Coordinate with the second agent producing definitive counts; do not file until those land and match.

Per BOP strategy guidance, the BOP should be lean — it is not the place to litigate the precise email count. A round, conservative, scoped figure is the safest approach.

5. Files That Cite Either Number and Need to Be Updated for Consistency

Once a single canonical figure is locked, the following files must be updated in lockstep. Failure to do so will create internal contradictions opposing counsel can exploit.

Cites 29,862:

Cites 19,232:

Also referenced indirectly (verify):

6. Action Items Before 04/02/2026

This memo does not edit BOP_RESPONSE_DRAFT.md or RESEARCH_LOG.md. The counsel/lead-author should make the final call on the canonical figure before propagating the change.