The Work Nobody Sees

Why Root Cause Analysis and Timeline Reconstruction Take So Long, How Much Evidence Is Enough, What Happens When a Piece Doesn’t Fit, and How to Build the Story That Wins

May 14, 2026

Why Root Cause Analysis and Timeline Reconstruction Take So Long, How Much Evidence Is Enough, What Happens When a Piece Doesn’t Fit, and How to Build the Story That Wins

A Forensic Engineer’s Honest Account of the Process Behind the Conclusion

The attorney calls on a Tuesday. Two vehicles. One intersection. Four photographs, a police report, and two depositions. The attorney asks how long the analysis will take. The engineer says six to eight weeks. The line goes silent for a moment, and when the attorney speaks again the question is the same one every attorney asks: “Why does it take that long? I can see the damage in the photographs. The police report says who ran the light. What am I paying you to do for six weeks?”

It is a fair question, and it deserves an honest answer—not because the attorney is wrong to ask it, but because the answer reveals everything the attorney needs to understand about the difference between looking at evidence and analyzing it, between having enough information to form an impression and having enough to defend a conclusion under oath, and between a narrative that sounds right and a story that the physics cannot contradict.

The six to eight weeks is not the engineer staring at photographs. It is the engineer doing the work that nobody sees—the work that transforms a stack of documents into a conclusion that survives deposition, survives cross-examination, survives a Daubert challenge, and survives the opposing expert’s best effort to dismantle it. That work cannot be rushed without being compromised, and it cannot be skipped without being discovered. What follows is an honest account of where the time goes, how much evidence the process actually requires, what happens when a piece of evidence refuses to cooperate, and how the surviving analysis becomes the story that wins the case.

Why It Takes So Long: The Work Behind the Conclusion

The attorney sees the output: a report, a set of annotated exhibits, and an expert who can explain what happened in plain language. What the attorney does not see is the iterative, nonlinear, frequently frustrating process that produced that output. Forensic analysis is not a pipeline where evidence enters one end and conclusions emerge from the other. It is a cycle of observation, hypothesis, testing, revision, and retesting that the engineer repeats until the conclusion stabilizes—until the same answer emerges regardless of which analytical path the engineer takes to reach it.

Phase 1: Evidence Acquisition and Inventory

Before a single calculation is performed, the engineer must acquire, organize, catalog, and verify every piece of available evidence. This phase alone routinely consumes one to two weeks, and the time is not optional.

The evidence arrives in fragments. The police report arrives first, often incomplete. The photographs arrive in batches—scene photographs from the officer, damage photographs from the adjuster, supplemental photographs from a party’s attorney—with inconsistent resolution, unknown provenance, and no guarantee of completeness. Depositions arrive on their own schedule, sometimes months apart. The EDR data arrives after a download that must be coordinated with the vehicle’s custodian. The NCAP test data must be identified, downloaded, and verified against the specific vehicle year, make, and model. The Google Street View imagery must be captured at multiple dates and documented with coordinates and timestamps. Weather records, signal timing data, EMS logs, and cellular records each require separate requests to separate custodians on separate timelines.

A short note before the break. Inventor's Mind is free, and subscribing stays free — no payment, no upsell, just an email. I am being honest with you: for months I have been writing into a feed without knowing who was on the other side of it.

The free sub is how I close that gap. It tells me who showed up, what landed, and where to take the next column. It is not a paywall. It is a signal flare. Please don't disappear — join and stay so I can make the work that makes you happy. Thank you for reading.

Each piece of evidence must be verified for authenticity and completeness before it enters the analysis. A photograph with missing metadata must be flagged. An EDR download with a chain-of-custody gap must be noted. A deposition that references documents not produced in discovery must trigger a request to counsel. The engineer who begins the analysis before the evidence is complete risks building conclusions on a foundation that changes when the missing pieces arrive—which means starting over, not adjusting.

Phase 2: Systematic Review and Extraction

With the evidence cataloged, the engineer begins the multi-pass review process described throughout this series. Each evidence category is reviewed systematically. Photographs undergo the six-pass review: inventory and chain of custody, global damage assessment, localized deformation analysis, contact and transfer evidence, safety system evidence, and pre-existing condition identification. Depositions undergo the four-pass review: survey, extraction, gap analysis, and consistency check. The police report is parsed for absolute time anchors, spatial data, and preliminary observations that require independent verification.

This phase produces the engineer’s working evidence matrix—a comprehensive catalog of every fact, every measurement, every observation, and every data point extracted from every source, cross-referenced by source and tagged with reliability indicators. Building this matrix is painstaking, unglamorous work that no one outside the engineering practice will ever see. It is also the work that makes everything that follows possible, because every subsequent calculation, every hypothesis test, and every cross-check traces back to this matrix. The engineer who shortcuts this phase discovers the gaps in trial preparation, not in the analysis phase where they can be addressed.

Phase 3: Hypothesis Generation and Initial Analysis

With the evidence matrix complete, the engineer generates hypotheses—multiple competing explanations for the event that are each consistent with at least some of the evidence. The generation process is informed by the patterns identified during the systematic review, by the engineer’s experience with similar failures, and by the known physics of the event type.

Each hypothesis is then subjected to initial analytical testing. The engineer performs the calculations required to evaluate whether each hypothesis is physically consistent with the quantitative evidence: speed estimates from crush analysis, momentum calculations from post-impact trajectories, deceleration rates from skid mark measurements, sight-distance calculations from roadway geometry, and severity assessments from NCAP data comparison. Each calculation requires input values, each input value requires a documented source or a justified assumption, and each assumption must be tested for sensitivity.

This phase is where the majority of the analytical time is consumed, because it is iterative. The first calculation produces a result that must be checked against other evidence. If the result is inconsistent—if the speed estimate from crush analysis does not align with the speed indicated by the EDR data—the engineer must investigate the inconsistency. The investigation may reveal an error in the calculation, an incorrect assumption, a limitation in the analytical method, or a more complex event sequence than initially hypothesized. Each discovery triggers a revision, and each revision triggers a new round of cross-checks. The iteration continues until all elements of the analysis are mutually consistent.

This iterative process cannot be scheduled with precision, because the engineer cannot predict in advance how many iterations will be required. A straightforward rear-end collision with consistent evidence across all sources may converge in a single pass. A complex multi-vehicle, multi-impact event with contradictory witness accounts and ambiguous physical evidence may require a dozen iterations before the analysis stabilizes. The six-to-eight-week estimate accounts for this uncertainty. It is not padding. It is the realistic duration of a process that must be allowed to run until it converges.

Phase 4: Cross-Validation and Stress Testing

Once the analysis converges on a conclusion, the engineer’s work is not finished. The conclusion must be tested against every evidence source that was not directly used in the primary analysis—the independent cross-checks that either corroborate or challenge the result. If the primary speed estimate was derived from crush analysis, the engineer cross-checks it against momentum calculations, EDR data, skid mark analysis, and NCAP comparison. Each independent check that corroborates the conclusion strengthens it. Each check that produces a conflicting result requires investigation and resolution.

The engineer also performs the sensitivity analysis: systematically varying each critical assumption within its reasonable range to determine how the conclusion changes. If the speed estimate is 42 mph when the friction coefficient is 0.72, what is it when the coefficient is 0.65? What about 0.80? The resulting range defines the confidence bounds on the conclusion—the honest answer to “how fast was the vehicle going?” is not a single number but a range that reflects the analytical uncertainty.

The stress test goes further. The engineer deliberately attempts to break their own conclusion. What is the weakest link in the causal chain? What assumption, if wrong, would change the answer? What evidence, if reinterpreted, would support an alternative hypothesis? The engineer who has stress-tested their own conclusion knows exactly where the opposing expert will attack, because the engineer has already attacked the same points. This preparation is invisible to the attorney—it does not appear in the report—but it is the preparation that makes the expert unflappable under cross-examination.

Phase 5: Report Preparation and Exhibit Development

The final phase translates the analytical work into deliverables: the expert report, the supporting calculations, the annotated exhibits, and—where appropriate—the visualization elements. The report must document every opinion, every basis for every opinion, every assumption, every evidence source considered, and every alternative hypothesis evaluated and eliminated. Federal Rule of Civil Procedure 26(a)(2)(B) requires this disclosure, and any opinion not documented in the report is subject to exclusion at trial.

Report preparation is not a transcription exercise. It is a communication challenge. The analysis must be presented in a logical sequence that a non-engineer can follow, with sufficient technical detail to withstand expert scrutiny and sufficient clarity to be understood by the trier of fact. Exhibits must be designed to communicate specific analytical points without visual overload. Every sentence in the report must be defensible, every number must be traceable, and every conclusion must be connected to its evidentiary foundation by an unbroken chain of reasoning.

This is where the six to eight weeks went. Not in staring at photographs. Not in billing hours on obvious conclusions. In building an analytical structure so thoroughly tested, so completely documented, and so clearly communicated that it can withstand every challenge the adversarial system can produce.

How Much Evidence Is Enough: The Sufficiency Threshold

The question every attorney asks second—after asking why it takes so long—is how much evidence the engineer actually needs. The answer is unsatisfying in its imprecision but honest in its accuracy: the engineer needs enough evidence to eliminate all alternative hypotheses except one, or enough to bound the conclusion within a range narrow enough to be useful.

This is not a fixed quantity. It depends on the complexity of the event, the number of plausible alternative explanations, and the specificity of the conclusion required. But the concept of sufficiency can be structured in a way that gives the attorney a practical framework for evaluating whether the expert has enough to work with.

The Minimum: One Conclusion Can Be Supported

At the minimum sufficiency threshold, the available evidence is sufficient to support at least one conclusion to a reasonable degree of engineering certainty. This does not mean the evidence eliminates all alternatives—it means the evidence supports one explanation more strongly than any competing explanation, and the engineer can articulate why. The conclusion carries uncertainty. The range is wide. But the opinion is defensible because the evidence points more strongly in one direction than in any other.

For a straightforward collision case, the minimum evidence set typically includes: photographs of all involved vehicles showing damage from multiple angles, a police report with a scene diagram and measured positions, at least one witness account of the event sequence, and sufficient vehicle identification to locate applicable NCAP test data. With these elements, the engineer can generally establish the impact configuration, estimate a severity range, determine the principal direction of force, and offer a preliminary reconstruction of the event sequence.

The Target: All Alternatives Are Eliminated

The target sufficiency threshold—the level every forensic engineer aims for—is the point at which the evidence eliminates every plausible alternative hypothesis and the surviving conclusion is corroborated by multiple independent sources. At this level, the conclusion is not merely the most probable explanation. It is the only explanation consistent with all available evidence.

Reaching this threshold requires evidence breadth—multiple independent data types that each constrain the analysis from a different direction. Physical evidence constrains the impact mechanics. Electronic data constrains the pre-impact vehicle behavior. Environmental evidence constrains the conditions that influenced driver decisions. Testimonial evidence provides the human narrative that physical evidence cannot capture. Each additional evidence category reduces the space of possible explanations and increases the precision of the surviving conclusion.

The practical implication for attorneys is that evidence breadth matters more than evidence depth within any single category. Ten photographs from the same angle are less valuable than five photographs from five different angles. Three depositions that cover the same ground are less valuable than three depositions that cover different aspects of the event. The attorney who obtains a wide range of evidence types gives the engineer the tools to build a conclusion supported by convergence. The attorney who obtains a deep collection of a single evidence type gives the engineer the tools to build a conclusion supported by only one pillar.

The Insufficiency Line: When There Is Not Enough

Below the minimum threshold, the available evidence is insufficient to support any conclusion to a reasonable degree of engineering certainty. The engineer can describe the physical evidence. The engineer can identify possible explanations. But the engineer cannot select one explanation over another because the evidence does not discriminate between them.

This is not a failure of the engineer. It is a reality of the evidence, and the honest response is to report the insufficiency rather than to speculate past it. The engineer who offers a conclusion when the evidence is insufficient has crossed the line between analysis and advocacy. The conclusion may be correct by coincidence, but it cannot be defended as an engineering opinion because the analytical process did not produce it—assumption and judgment filled the gaps that evidence should have occupied.

For the attorney, evidence insufficiency is not the end of the case. It is a discovery problem. The engineer’s identification of specific evidence gaps—“I need friction testing on the actual roadway surface,” “I need the EDR data downloaded from Vehicle B,” “I need photographs of the undercarriage showing the suspension components”—provides the attorney with a targeted discovery list. Each gap filled moves the analysis closer to the sufficiency threshold. The forensic engineer’s role is not limited to analyzing what exists. It extends to identifying what is missing and advising counsel on how to obtain it.

The Diminishing Returns Line: When There Is More Than Enough

At the opposite end of the spectrum, there is a point where additional evidence does not materially change the conclusion. If four independent methods produce speed estimates within a five-mph range and the conclusion is supported by physical evidence, electronic data, and environmental documentation, a fifth method that produces a result within the same range adds confirmation but does not change the analysis. The attorney should understand when additional discovery expense is justified by analytical need and when it is producing diminishing returns.

The engineer’s responsibility is to communicate clearly where on this spectrum the current evidence places the analysis. “With the evidence we have, I can offer a conclusion within this range. With additional evidence—specifically these items—I can narrow the range to this. Beyond that, additional evidence is unlikely to change the conclusion.” This communication allows the attorney to make an informed cost-benefit decision about additional discovery. It is a conversation that too few experts initiate and too few attorneys request.

When a Piece of Evidence Does Not Fit: The Outlier Problem

This is the moment that defines the integrity of the forensic analysis. The engineer has assembled the evidence, generated and tested hypotheses, and converged on a conclusion supported by the majority of the evidence. And then one piece of evidence refuses to cooperate. One data point, one witness statement, one measurement, one photograph contradicts the conclusion that every other piece of evidence supports.

What the engineer does next determines whether the analysis is honest or curated. The temptation—and it is a genuine temptation, amplified by the pressure of litigation timelines and retaining party expectations—is to set the outlier aside. To minimize it. To rationalize it. To mention it in passing and move on to the conclusion that the rest of the evidence supports. Every forensic engineer who has practiced long enough has felt this temptation. The ones worth retaining are the ones who resist it.

The Outlier Is Not the Enemy: It Is the Teacher

An outlier—a piece of evidence that does not fit the working conclusion—is one of the most valuable findings in the entire analysis, because it forces the engineer to confront a possibility that the convergent evidence alone does not reveal: the possibility that the conclusion is wrong, or incomplete, or simpler than the actual event.

The forensic engineer’s response to an outlier is not to explain it away. It is to investigate it with the same rigor applied to every other piece of evidence. The investigation follows a structured decision tree that leads to one of four conclusions about the outlier, each with different implications for the analysis.

Conclusion 1: The Outlier Contains an Error

The most common resolution is that the outlier evidence itself is unreliable. A witness’s account of the event sequence contradicts the physical evidence because human memory is reconstructive and imperfect. A measurement in the police report is inconsistent with the photographic evidence because the officer estimated rather than measured. An EDR data point is anomalous because the module recorded a voltage spike rather than a genuine vehicle parameter.

When the outlier is attributable to a documented reliability limitation in the source, the engineer may appropriately reduce the weight given to that evidence—but must disclose both the outlier and the basis for the reduced weighting. The key word is “documented.” The engineer must identify a specific, articulable reason why the evidence is unreliable. “This data point does not fit my conclusion” is not a reliability assessment. It is circular reasoning. “This witness’s account of the vehicle’s direction of travel is contradicted by the paint transfer location, the crush geometry, and the debris scatter pattern, all of which independently indicate a different direction” is a reliability assessment grounded in multiple corroborating sources.

Conclusion 2: The Working Conclusion Contains an Error

The outlier may be correct, and the engineer’s working conclusion may be wrong. This is the resolution that the ego resists and the methodology demands. If the outlier evidence is reliable—if its source is credible, its measurement is verified, and no documented basis exists for discounting it—the engineer must reopen the analysis and determine what conclusion is consistent with all of the evidence, including the outlier.

This reopening is the most time-consuming resolution and the most important. It is also the resolution that explains why the six-to-eight-week timeline includes contingency. An analysis that was converging on a conclusion may need to be partially or fully revised when a reliable outlier forces reconsideration. The engineer who has built the analysis on a solid evidence matrix can perform this revision systematically. The engineer who shortcut the matrix-building phase must essentially start over.

The attorney’s response to an engineer who reports, “I found an outlier that forced me to revise my initial conclusion,” should not be disappointment. It should be confidence. An engineer who revised a conclusion based on evidence is demonstrating exactly the kind of intellectual honesty that survives cross-examination. An engineer who never revised anything either got it right the first time—possible but uncommon in complex cases—or never tested their own work rigorously enough to find the problem.

Conclusion 3: The Event Is More Complex Than Initially Hypothesized

Sometimes the outlier does not indicate an error in the evidence or an error in the conclusion. It indicates that the event involved a complexity that the initial hypotheses did not capture. A speed estimate from crush analysis is inconsistent with the EDR data—not because either is wrong, but because the vehicle struck a curb before the primary impact, absorbing energy in a pre-impact event that the crush analysis did not account for. A witness’s description of the vehicle’s trajectory is inconsistent with the momentum calculation—not because the witness is mistaken, but because the vehicle struck a second object after the primary collision, altering its post-impact path in a way the single-impact momentum model did not predict.

The outlier, in these cases, is the evidence that reveals the hidden complexity. It is the data point that forces the engineer to develop a more sophisticated model of the event—a model that accounts for the pre-impact curb strike, the secondary collision, or the multi-phase trajectory that the simpler model missed. The revised model, incorporating the outlier, produces a conclusion that is more accurate than the original because it captures the full complexity of the physical event.

This is forensic engineering at its most valuable. The attorney who understands this process recognizes that the outlier investigation is not a delay or a complication. It is the work that turns a decent analysis into an exceptional one. The conclusion that emerged from resolving the outlier is stronger than the conclusion that would have existed without the outlier, because it accounts for a dimension of the event that would otherwise have been invisible.

Conclusion 4: The Outlier Is Genuine and Irreconcilable

In rare cases, a piece of evidence is reliable, the working conclusion is well-supported, and no hidden complexity explains the discrepancy. The outlier simply does not fit, and the engineer cannot determine why.

The honest response is to report the irreconcilable outlier explicitly. The report states: the conclusion is supported by these evidence sources, this evidence source is inconsistent with the conclusion, and the investigation was unable to determine the cause of the inconsistency. The conclusion remains the most probable explanation because it is supported by the preponderance of evidence, but the engineer discloses the limitation.

This disclosure is not a weakness. It is a demonstration of integrity that judges and juries recognize. An expert who acknowledges an irreconcilable data point is an expert who is not hiding anything. An expert who presents a seamless, contradiction-free analysis in a complex case either did extraordinary work or is not reporting what they found. The trier of fact will evaluate which explanation is more likely.

How to Build the Story: From Analysis to Courtroom Narrative

The analysis is complete. The conclusion has survived cross-validation, stress testing, sensitivity analysis, and outlier investigation. The evidence matrix is cataloged, the calculations are documented, and the report meets Rule 26 disclosure requirements. Now the engineer and attorney face a different challenge: translating the analytical work into a story that twelve non-engineers can follow, understand, and remember.

The word “story” makes some engineers uncomfortable. It sounds like fiction. It sounds like advocacy. It sounds like the opposite of objective analysis. But the discomfort is misplaced. A story is simply a sequence of events connected by causation and presented in a logical order. That is exactly what the forensic analysis produced. The task is not to fictionalize the analysis. It is to present the analysis in the narrative structure that human cognition is designed to process.

The Architecture: Three Acts

Every effective forensic presentation follows a three-act structure, not because courtrooms are theaters but because human comprehension organizes information into beginnings, middles, and endings. The forensic story has a natural three-act structure that mirrors the event itself.

Act One: The Setup. What were the conditions before the accident? What was the physical environment? What were the vehicles doing? What should the drivers have perceived, and what decisions should they have made? Act One establishes the world the accident happened in—the roadway geometry documented by Street View, the traffic control devices captured in archived imagery, the weather conditions recorded by the National Weather Service, and the vehicle conditions established by maintenance records and pre-accident inspections. Act One answers the question: what did the physics require the drivers to do?

Act Two: The Failure. What went wrong? Where did the causal chain begin? What decisions were made that deviated from what the environment required? What mechanical failures occurred that altered the vehicles’ behavior? What conditions aligned to create the path from hazard to harm? Act Two is the root cause analysis presented as narrative. Each link in the why-chain becomes a scene in the story. The driver did not brake because the driver did not see the hazard because the sight line was obstructed because the vegetation was not maintained. Each scene follows from the previous one with the logical inevitability that the five-year-old test validated.

Act Three: The Proof. How do we know this is what happened? Act Three is the evidence convergence presented as confirmation. The crush analysis confirms the severity. The EDR data confirms the speed. The Street View imagery confirms the sight-line obstruction. The NCAP comparison calibrates the damage against a known benchmark. The witness testimony corroborates the sequence. Each independent evidence source is a witness that testifies to the same conclusion from a different vantage point. By the end of Act Three, the jury has heard the same story told by the metal, the electronics, the environment, the testing data, and the human observers—and they have heard it told consistently.

The Narrative Engine: Causation, Not Chronology

The most common mistake in forensic presentation is organizing the story chronologically—starting at the beginning of the timeline and moving forward through each event in sequence. Chronological organization is logical but not persuasive, because it buries the most important information in the middle of the narrative where the jury’s attention is weakest.

The more effective structure organizes around causation—leading with the root cause, then showing how the root cause produced the chain of events that led to the harm. This structure leverages the primacy effect by placing the most important analytical conclusion at the beginning, where it anchors the jury’s understanding of everything that follows. The driver could not see the oncoming vehicle. That is the root cause. Now let me show you why the driver could not see it, what happened because the driver could not see it, and how the physical evidence proves that this is exactly what occurred.

The causal structure also creates a natural framework for introducing each piece of evidence at the moment it is most relevant. The sight-line analysis is introduced when the narrative explains why the driver could not see. The crush analysis is introduced when the narrative explains the severity of the impact. The NCAP data is introduced when the narrative needs a calibration benchmark. Each piece of evidence arrives in the story at the moment the jury needs it to understand the next link in the causal chain. Nothing is introduced before its context has been established. Nothing is introduced after the jury has stopped caring about its category.

The Anchor Points: Where the Story Becomes Unforgettable

Every effective forensic narrative contains three to five anchor points—moments in the presentation where the engineering conclusion is crystallized into a single image, comparison, or statement that the jury will remember during deliberation. These are not the most technical moments in the presentation. They are the simplest.

An anchor point might be the side-by-side comparison: the NCAP test vehicle at 35 mph next to the field vehicle, with crush measurements on both. The visual contrast communicates the severity relationship in a single glance that no verbal explanation can match. An anchor point might be the Street View capture showing the sight line, with the obstruction circled and the available sight distance labeled. The jury can see what the driver could see—and what the driver could not. An anchor point might be the five-year-old explanation of the root cause: “The arm that holds the wheel broke because it was made with a weak spot.” Seven words that capture a metallurgical failure analysis the jury will remember when they cannot remember anything else.

The engineer and attorney should identify the anchor points collaboratively before the presentation is designed. Every exhibit, every diagram, every visualization should serve one of two purposes: establishing context for an anchor point or being an anchor point. Exhibits that do neither are candidates for elimination, because they consume the jury’s cognitive resources without advancing the narrative toward a memorable conclusion.

The Credibility Thread: Transparency as Persuasion

The story must include its own limitations. This sounds counterintuitive—why would the presentation voluntarily disclose weaknesses?—but the strategic logic is well-established in persuasion research and trial practice. An expert who discloses a limitation before opposing counsel raises it controls the narrative around that limitation. An expert who is forced to disclose a limitation on cross-examination has lost control.

The credibility thread runs throughout the three-act structure. In Act One, the engineer acknowledges the temporal gap between the Street View imagery date and the accident date. In Act Two, the engineer identifies the assumption that carries the most uncertainty and explains the sensitivity range. In Act Three, the engineer discloses the irreconcilable outlier and explains why it does not change the overall conclusion. Each disclosure is brief, matter-of-fact, and immediately followed by the evidence that supports the conclusion despite the limitation.

The cumulative effect of these disclosures is trust. The jury sees an expert who is not hiding anything. The jury sees an analysis that has been tested and found to be robust despite its limitations. The jury sees a story that acknowledges imperfection and survives it—which is exactly how reality works and exactly how jurors expect an honest expert to present it.

The Closing Frame: What the Evidence Requires

The story ends where the analysis ends: with the conclusion that the evidence requires. Not suggests. Not indicates. Requires. The language is deliberate. The engineer has tested every alternative. The engineer has stress-tested the conclusion. The engineer has disclosed the limitations. And the evidence, subjected to all of that scrutiny, requires this specific conclusion because no other explanation survives the testing.

The closing frame should echo the root cause in its simplest form—the five-year-old version, delivered without apology, without qualification, and without the technical scaffolding that supported the analysis but is no longer needed now that the conclusion has been reached. The arm broke because it had a weak spot. The driver could not stop because they could not see. The vehicle left the road because the brakes failed. Simple. True. Supported by every piece of evidence the jury has spent the last hour examining. The story does not end with a flourish. It ends with inevitability.

The Pitfalls: Where the Process Breaks Down

Rushing to Conclusion Before the Evidence Stabilizes

The most damaging failure is issuing a report before the analytical iterations have converged. An attorney pressing for an early opinion may receive a preliminary conclusion that changes when additional evidence arrives or additional cross-checks are performed. An opinion that changes between the report and the deposition is a cross-examination weapon that opposing counsel will wield with precision. The engineer’s obligation is to communicate honestly about the timeline and to resist pressure to deliver premature conclusions. The attorney’s obligation is to provide the time the analysis requires, because the cost of a revised opinion is always greater than the cost of a delayed one.

Ignoring the Outlier to Meet a Deadline

A forensic report filed without investigating an identified outlier is a ticking time bomb. If the engineer noticed the inconsistency and chose not to investigate it, the omission is a professional and ethical failure. If opposing counsel’s expert finds the outlier and raises it during cross-examination, the engineer must explain why it was not addressed—and no explanation will be satisfactory. The outlier investigation must be completed before the report is issued, even if it means requesting additional time.

Building the Story Before the Analysis Is Complete

When the attorney develops the case narrative before the engineering analysis is finished, the narrative drives the analysis rather than the reverse. The engineer feels implicit or explicit pressure to produce conclusions that support the pre-existing story. The analytical process is contaminated by the destination it is expected to reach. The remedy is simple in principle and difficult in practice: the engineering analysis must be completed independently before the narrative is constructed. The story must emerge from the analysis. The analysis must never be shaped to fit the story.

Overcomplicating the Narrative

An engineer who cannot resist presenting every calculation, every cross-check, every sensitivity analysis, and every alternative hypothesis to the jury has confused thoroughness with communication. The analysis must be thorough. The presentation must be selective. The report documents the complete analytical work. The courtroom presentation extracts the narrative thread, the anchor points, and the credibility disclosures that the jury needs to reach a verdict. Everything else belongs in the technical appendix, available for cross-examination but not volunteered during direct.

The Work Nobody Sees Is the Work That Wins

The attorney who asks why the analysis takes six to eight weeks is asking the right question. The answer is that every day of that timeline is occupied by work that the attorney will never see but that the opposing expert will test, the opposing counsel will challenge, and the trier of fact will ultimately evaluate—even without knowing it exists. The evidence matrix that no one reads is the foundation the entire report stands on. The outlier investigation that delayed the timeline by ten days is the reason the conclusion survives the cross-examination question that would have destroyed it. The sensitivity analysis that never appears in the presentation is the preparation that allows the expert to answer the “what if” question without hesitation.

The evidence threshold is not a number. It is a functional standard: enough evidence to eliminate alternatives, corroborate the conclusion through independent sources, and define the boundaries of what must have happened and what could not have happened. The attorney who obtains broad evidence across multiple categories gives the engineer the material to reach that standard. The attorney who obtains narrow evidence in a single category does not.

The outlier that does not fit is not an inconvenience. It is a gift—the one piece of evidence that forces the analysis to be better than it would have been without it. The engineer who investigates the outlier honestly produces a conclusion that is fortified against the exact challenge the outlier represents. The engineer who ignores it produces a conclusion with a hidden crack that opposing counsel will find.

The story that wins is not the most dramatic, the most technical, or the most expensive to produce. It is the simplest true account of what the evidence requires—organized around causation, anchored by three to five unforgettable moments, threaded with the credibility disclosures that earn the jury’s trust, and closed with a conclusion so thoroughly tested that it feels inevitable.

The work that produced that story is the work nobody sees. It is also the only work that matters.

The Inventor's Mind Blog's Substack

Discussion about this post

Ready for more?