About

About PTEM Labs

PTEM Labs builds governed structural measurement instruments for catastrophe risk.

We deliver Phase Atlas structural climatology surfaces, PAC conditioning artifacts, and Capital Impact decision artifacts for institutional portfolio workflows.

PTEM Labs is an independent developer of governed structural measurement systems for catastrophe risk.

The work is led by Nathan Howell, Founder, based in Florida, United States. PTEM originated as an independent systems analysis effort and has developed into a structured instrument and evaluation platform for institutional risk workflows.

PTEM Labs operates as a founder-led, single-instrument organization: narrow in scope, strict in governance, and focused on versioned, auditable, reproducible outputs.

The operating orientation remains systems analysis: decomposing complex behavior into observable temporal structure and organization-state transitions. That orientation is applied operationally as institutional measurement infrastructure rather than a forecasting product.

Current System State

PTEM currently operates across four governed layers:

  1. governed structural measurement
  2. Phase Atlas structural climatology
  3. PAC portfolio-level conditioning
  4. Capital Impact tail metric translation

PTEM does not predict events; it measures structural regime availability and translates that into deterministic constraints on portfolio-level risk outputs.

Why PTEM exists

Institutional catastrophe workflows already rely on strong outcome models for intensity, hazard, and loss. Those systems are essential, but they do not directly measure structural organization through time.

PTEM exists to provide a governed structural layer for regime formation, persistence, drift monitoring, and structural-capacity constraints used in portfolio conditioning and model-risk review.

What PTEM is

PTEM is run as an instrument company, not a bespoke model service. Freeze control, forward-only scoring, and explicit change management keep releases inside a reproducible governance envelope.

Canonical releases include manifests, validation receipts, and versioned structural artifacts. Atlantic hurricanes are the proving ground: PTEM is validated on the frozen Atlantic v1.2 structural spine (1851–2024) at 6-hour cadence, from which Phase Atlas surfaces are deterministically produced.

PTEM is built to be evaluated, not assumed.

Validation posture

PTEM is governed before it is persuasive. Development and release discipline is pre-declared and replayable.

  • Frozen structural datasets anchor canonical runs and climatology releases
  • Forward-only computation with no look-ahead or retroactive edits
  • Pre-declared operating points and thresholds filed before release
  • Deterministic replay from manifests, validation receipts, and hash references
  • Audit trails published for institutional review and committee workflows
  • EPAC portability executed under identical governance rules

Lead distributions and mechanism-oriented evidence are published on /validation as evaluation material, not as operating-point selection tooling.

Public research artifacts supporting external evaluation are available in the Research section.

Two co-equal horizons

Structural measurement layer

  • Activation episodes and regime-entry sequences under forward-only discipline
  • Episode behavior traces and lead distributions for event interrogation
  • Audit-shaped outputs for challenge-layer review against CAT assumptions

Phase Atlas + PAC + Capital Impact layer

  • Phase Atlas structural climatology surfaces with fixed thresholds and frozen inputs
  • PAC deterministic conditioning for portfolio-level event-set workflows
  • Capital Impact translation into capital-relevant tail metrics and one-page decision artifacts
  • No storm-level structural exposure through conditioning channels

Measurement and conditioning layers are separately governed so institutions can use portfolio constraints without expanding storm-level exposure.

How institutions engage

PTEM is not distributed as an open feed or a generic model endpoint. Engagement is structured and governance-bound.

  • Structured evaluation with frozen Phase Atlas bundles and governance documentation
  • Replay-safe read-only SDK access, PAC interfaces, and Capital Impact reports
  • Validation packs for model-risk, portfolio, and controls review
  • Optional pilot integrations under institutional licensing terms
  • No source-code disclosure; proprietary internals remain sealed

Engagement path: evaluation → optional pilot → governed integration.

What we focus on

  • Maintaining a freeze-controlled Atlantic structural record with reproducible artifacts
  • Producing governed Phase Atlas structural climatology surfaces
  • Operating PAC as a deterministic conditioning layer for portfolio outputs
  • Producing one-page capital impact sheets from deterministic post-conditioning analytics
  • Applying explicit governance gates and change-control for every release
  • Delivering a single SDK surface for institutional measurement and conditioning access

The scope remains narrow: structural measurement, conditioning artifacts, and governed delivery for institutional risk workflows.

Who we serve

  • Reinsurers and portfolio-risk owners who need structural capacity constraints alongside CAT outcomes
  • Model validation and risk governance teams requiring replayable structural challenge-layer evidence
  • Public programs and evaluators reviewing structural climatology without storm-level exposure

Institutional evaluation access is available via /contact; access scope is outlined on /access.