🎙️ Episode Overview

Tonight’s 45-minute Master Class addresses the methodology problem that almost nobody teaches systematically: when a subject gives multiple conflicting accounts of the same events, which elements do you credit, weight, or discard — and on what basis? The instinctive answer — trust the most recent account — is wrong often enough to be dangerous. Morgan builds a five-principle framework for parsing conflicting confessions and applies it in real time to all four Watts iterations to produce a reconstruction more accurate than any single version provides.

🔍 In This Episode

Morgan opens with the core problem: most investigators default to recency when accounts conflict, and most of the time that default works. But the Watts case demonstrates exactly where it fails and why. He then walks through five principles for treating each confession account as data rather than truth — mapping self-serving elements, running corroboration as a weight machine, building a contradiction map, and identifying what stays consistent across iterations under different conditions. The session closes by applying all five principles to the Watts case simultaneously, producing a synthesized reconstruction with explicit confidence levels for each element.

🧠 Key Concept

The Confession as Reconstruction Tool — A confession account is not a disclosure of truth. It is a document produced by a specific person under specific conditions with specific incentives at a specific moment. When a subject produces multiple accounts, the investigator or analyst is not looking for the correct version among incorrect ones. They are treating each account as data — each one revealing what the subject needed to disclose or conceal at that moment. The lie is as informative as the truth. Sometimes more so.

⚠️ Why This Matters

Conflicting accounts are a routine feature of serious investigations. Every investigator who has worked a case involving multiple interview sessions, a changing story, or a post-conviction disclosure has faced this methodology problem. The default — assume the latest account is most accurate — has no principled basis and fails in documented cases. A systematic framework for weighting conflicting accounts produces reconstructions that are demonstrably more accurate than any single account provides.

🔑 The Five Principles

Principle 1: Recency Is Not Accuracy

The most recent account earns additional weight only when it meets specific criteria. Four criteria apply:

* The context removed or reduced the subject’s incentive to lie (post-conviction, no remaining legal exposure)

* The account was given to analysts rather than prosecutors (different dynamic, different disclosure patterns)

* New details are corroborated by independent evidence the subject didn’t control

* New details increase the subject’s culpability rather than reduce it (self-incriminating additions are more reliable than exculpatory ones)

The February 2019 Watts FBI interview meets all four criteria. It earns the highest weight among the four iterations — not because it came last, but because it passes the test.

Principle 2: Map the Self-Serving Elements First

Every account is produced with a goal beyond truth-telling. Identify that goal before evaluating content.

* Iteration 1 (Grief Performance): Goal is cooperative innocence. Every element serves that goal. Full provisional weight, no more.

* Iteration 2 (Polygraph Break): Goal is to acknowledge presence while redirecting the murder charge for the girls. The self-serving element — “Shanann killed the girls” — is discarded. Everything else earns more weight.

* Iteration 3 (Plea): Goal is to resolve the legal matter while controlling information. The least informative account precisely because its goal is maximum information control.

* Iteration 4 (FBI Interview): Goal is psychological narrative control — present a man who was overwhelmed, not predatory. The oxycodone detail undermines his own narrative. That’s why it earns the highest weight.

Principle 3: Corroboration Is the Weight Machine

Every element of every account runs through three tiers:

* Full weight: Corroborated by independent evidence the subject didn’t control (surveillance footage, physical evidence, cell records, cadaver dogs)

* Provisional weight: No independent corroboration available, but consistent with known facts and not contradicted by physical evidence

* Discarded: Contradicted by physical evidence or physically implausible

Full weight examples: 5:15am truck departure (Trinastich surveillance), Shanann’s strangulation in the master bedroom (cadaver dog hit on mattress), bodies at Cervi 319 (physical recovery), 8-inch tank openings (documented at recovery).

Provisional weight examples: Oxycodone attempt (no corroboration available; baby survived and was not tested), Bella’s question on the drive (no corroboration possible; consistent with physical evidence that girls were alive in the home).

Discarded: Shanann strangled Bella and Celeste (contradicted by physical implausibility — a 15-weeks-pregnant woman strangling two children is inconsistent with the physical evidence of the crime).

Principle 4: Build the Contradiction Map

Where accounts contradict each other, the contradiction is evidence. The direction of the lie tells you what the subject needed to hide at that specific moment.

Who killed the girls: Iteration 2 says Shanann; Iteration 4 says Watts. The contradiction reveals that Iteration 2 was constructed specifically to redirect the most serious murder exposure. The direction of the lie tells you exactly what he was protecting.

Girls’ condition at departure: Iteration 2 is silent; Iteration 4 says alive. Physical evidence (no child death trace evidence found in the home, cadaver dog hits confined to master bedroom and garage) is consistent with Iteration 4. Full weight.

Location of murders: Iteration 2 is unspecified; Iteration 4 places Shanann in the master bedroom. Corroborated by the cadaver dog. Full weight on the Iteration 4 location.

Principle 5: What Doesn’t Change Is What You Keep

Elements consistent across multiple iterations given under different conditions and incentives are the foundation of the reconstruction. They have survived multiple pressures and emerged unchanged.

Watts consistent elements: the drive to Anadarko Cervi 319 (consistent from first acknowledgment through final account, corroborated by physical recovery), the timeline of Shanann’s return at approximately 1:48am (consistent across accounts, corroborated by doorbell camera and cell records), the fact of the marital argument (content shifted, fact consistent across all accounts).

These elements form the skeleton. Build from the consistent corroborated elements outward. Add provisional weight elements at lower confidence. Discard what the physical evidence eliminates.

🧩 Synthesized Reconstruction

What all four accounts, run through the five principles, produce:

Night of August 12 into August 13: Watts attempts to administer oxycodone through Shanann’s food — establishing a planning horizon that predates August 13th. Provisional weight. Shanann returns from Phoenix at approximately 1:48am. Full weight. An argument about the marriage occurs. High weight.

Early morning, August 13: Watts strangles Shanann in the master bedroom. Full weight. He goes to Bella’s room, then Celeste’s room. The girls are alive. Full weight — consistent with physical evidence at the home.

5:15am: Truck departs. Full weight. Girls are alive during the drive. Full weight. Bella asks where they are going. Full weight on the condition; provisional weight on the specific question.

Cervi 319: Shanann’s body buried near the site. Girls placed through 8-inch diameter tank openings. Full weight — corroborated by physical recovery and documented tank dimensions.

The permanent Unknowable: Why — at the level of psychological truth — remains outside the record.

📄 Companion Article

The Thursday morning Known vs. Knowable post provides the factual foundation for tonight’s Master Class. Both are published on the Crime: Reconstructed Substack.

🎧 About the Show

Crime: Reconstructed applies First Principles reasoning, constraint-based elimination, and systems thinking to criminal investigation. The platform isn’t about honoring victims or identifying villains — it’s about whether the investigation was done correctly. Hosted by Morgan Wright: former state trooper, detective, intelligence professional, and analyst with four decades of law enforcement experience.

Audio establishes the frame. Writing does the work.

🔗 Continue the Investigation

The full Week 9 reconstruction — all four confession iterations, the five-principle framework, and the synthesized reconstruction — is published on the Crime: Reconstructed Substack.

❓ Listener Question

The most self-incriminating elements of the February 2019 account — the ones that made Watts look worse, not better — are the ones that earn the most weight in the reconstruction. What does that tell us about how to read any account given under reduced legal exposure? Leave your answer in the comments.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit crimereconstructed.substack.com