▸ Demonstrating how high incident overlap emerges naturally from workload patterns—without wrongdoing.
⚠️ NOT FOR JUDICIAL USE
This software is a demonstrative aid for educational purposes only. It is not a validated forensic instrument.
Any application to real-world case data requires independent verification by a qualified expert (e.g., CStat/PStat accredited). See Documentation §1.1 for full liability disclaimer.
Analysis Mode (?)
Standard Monte Carlo: Fixed suspect (Nurse 38), random incident placement, compare overlap distributions.
Selection Effect: Simulates investigator bias by generating a larger pool of events, then selecting incidents where suspect was present. Shows how circular selection inflates apparent coincidence.
No Pre-designation: All 38 nurses drawn from same distribution. Reports maximum overlap for any nurse—answers "how unusual is it for someone to overlap this much?"
Advanced Configuration
▼
Definitions:
• 730 shifts = 365 days × 2 shifts/day
• Incidents placed without replacement
• Fixed Totals: round(p × 730) shifts per nurse
• Day/night mode keeps exact shift headcounts and lets nurses differ by day-vs-night exposure
Ward Staffing Profile (?)
Shift distribution preview (sorted 1→38). 12h shifts, 14 slots/week cap.
Suspect Nurse Exposure (?)
Percentage of all shifts worked by the suspect nurse.
≈ 59 hrs/week ·
2.2× ward average
Probability boost for high-risk shifts (experienced nurses get sicker patients).
Selection Effect Parameters (?)
Larger pool from which "suspicious" incidents are selected.
Relative weighting for selecting incidents from the pool. Models investigator preference, not evidence fabrication.
Simulation Parameters (?)
Running Simulation...
Mean Overlap
5th–95th Range
99th Percentile
Max Peak (most extreme observed)
ILLUSTRATIVE FRAMING
⚖️ Prosecution's Presentation
Loading...
ILLUSTRATIVE FRAMING
🛡️ Defence's Presentation
Loading...
Expected vs. Observed Overlap
Expected (Given Hours)
—
Simulated Mean
—
Ratio (Obs/Exp)
—
When ratio ≈ 1.0, observed overlaps match statistical expectation given workload.
Ward Mean Fill—
Average nurse workload
Highest Other Nurse—
Maximum non-suspect
Suspect Workload—
Nurse 38's exposure
P(Innocent Match) (?)—
95% CI: —
📊 Multiple Comparisons Problem
—
In — of simulations, at least one other nurse matched or exceeded the suspect's overlap count.
The "Prosecutor's Gap": Note the distance between the 99th percentile (amber) and the max peak (red).
Prosecutors may seize upon this outlier—which occurs naturally in large datasets—and present it as impossible coincidence.
But someone will always be the maximum; the question is whether that maximum is unusual given their exposure.