Decay-Curve Fit and Half-Life Estimate¶
Analysis — instantiates Signal Persistence and Refresh Design
Fits observed strength across time or distance to a decay model, reporting the half-life, regime changes, uncertainty, and the predicted point where strength crosses the usable threshold.
Decay-Curve Fit and Half-Life Estimate is the modeling step. It takes measured strength-versus-time (or versus-distance) points and fits a decay law — exponential, power-law, piecewise, empirical — to them, yielding a half-life, an uncertainty band, detected regime changes, and a forecast of when strength will fall below a supplied threshold. The one idea that makes it this mechanism: it converts scattered readings into a predictive curve. It does not collect the readings, and it does not act on the forecast; it turns evidence into an actionable projection.
Example¶
A municipal water utility must keep a disinfectant residual above the regulatory minimum at the farthest tap. Chlorine decays as it travels the distribution main, so operators pull residual samples at increasing detention times and distances and fit a first-order decay curve, extracting a bulk decay coefficient and an effective half-life. The residuals reveal a regime change — decay accelerates through an old cast-iron section with high wall demand — so a single exponential is split into two segments. The fit then projects that residual crosses the 0.2 mg/L floor at roughly the 46-hour mark in a specific dead-end zone, carried forward as a range rather than a point. That dated, localized forecast tells the utility where a booster is needed. The fit itself adds no chlorine and sets no rule; it forecasts.
How it works¶
- Anchor to a baseline. Establish the initial strength, units, and measurement uncertainty at the origin so decay is measured against a known reference, not guessed from mid-curve points.
- Choose a model family. Exponential, power-law, piecewise, or tabulated — selected for an actionable fit, not mathematical elegance.
- Fit and detect regimes. Estimate the rate and half-life, flag breakpoints where the regime shifts, and quantify residual scatter as an uncertainty band.
- Project the crossing. Extrapolate to where the curve meets the supplied threshold, carrying the uncertainty forward so the answer is a window, not a false-precise instant.
Tuning parameters¶
- Model family — flexible forms fit the data better but overfit and extrapolate poorly; rigid forms are robust but can miss real regimes.
- Baseline precision — a tighter reference sharpens the whole fit but costs careful origin measurement.
- Regime sensitivity — how readily the curve is split into segments; more segments capture reality but chase noise.
- Uncertainty representation — point estimate, band, or full distribution carried into the crossing forecast.
- Extrapolation horizon — how far beyond the data the projection is trusted.
When it helps, and when it misleads¶
Its strength is turning "it seems to be fading" into a dated, uncertainty-tagged crossing forecast that refresh, relay, and expiry decisions can schedule against. A good fit does not merely say the signal is decaying; it says when, how fast, and how sure — the three things a plan actually needs. It misleads when the fitted regime does not hold: extrapolating an exponential past a plateau or a second, faster driver produces a confident wrong date, and a clean R² can mask an overfit[1] that generalizes badly. The costs that bite are precisely the unmodeled regimes. The guarding discipline is to prefer the simplest model the residuals allow, treat the crossing as a band, and re-fit as new points arrive rather than freezing the first curve.
How it implements the components¶
decay_curve_half_life_and_regime_model— its output: the fitted decay law, the half-life, and the segmentation into regimes.baseline_strength_and_reference_state— it establishes and normalizes to the initial level, units, and uncertainty that the decay is measured against.
It does not gather the strength readings at real receivers (receiver_and_endpoint_observability — that's Endpoint Strength Probe Network, whose measurements it consumes), and it does not enforce a stop rule on the forecast crossing (usable_strength_threshold_and_expiry_rule — that's Staleness TTL and Expiry Gate); it models and forecasts.
Related¶
- Instantiates: Signal Persistence and Refresh Design — supplies the decay model and crossing forecast the rest of the lifecycle plans against.
- Consumes: Endpoint Strength Probe Network — the endpoint readings it fits.
- Sibling mechanisms: Endpoint Strength Probe Network · Staleness TTL and Expiry Gate · Scheduled Reinforcement Cadence · Threshold-Triggered Refresh Controller · Relay and Repeater Placement Model · Adaptive Gain and Pre-Emphasis · Multichannel Redundant Delivery · Persistence Stress and Shadow Test · Refresh Burden and Accumulation Audit
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: Decay-Curve Fit and Half-Life Estimate operates as a computation, comparison, model, or analytic representation used to infer, estimate, or choose because it fits observed strength across time or distance to a decay model, reporting the half-life, regime changes, uncertainty, and the predicted point where strength crosses the usable threshold.
Independent corroboration: The frozen evidence defines Decay-Curve Fit and Half-Life Estimate as 'Fits observed strength across time or distance to a decay model, reporting the half-life, regime changes, uncertainty, and the predicted point where strength crosses the usable threshold', so its operative form is Analysis, Modeling & Optimization.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Single lineage
Present-day reach: Universal
Rationale: Statistical modeling cohered fitting exponential, power-law, piecewise, or empirical decay functions to observations and estimating half-life, uncertainty, regime changes, and threshold-crossing time.
Related originating lineages:
- Physics — Radioactive decay and relaxation processes supplied the half-life parameter and canonical exponential model.
Review resolution: Statistical modeling cohered fitting exponential, power-law, piecewise, or empirical decay functions to observations and estimating half-life, uncertainty, regime changes, and threshold-crossing time.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The fit is an input, not a decision. Keeping the model separate from the mechanisms that consume it — the cadence that reads the half-life, the controller that reads the crossing forecast, the gate that reads the staleness rate — lets a team improve the estimate with better sampling without re-litigating every downstream policy at once.
References¶
[1] Babyak, M. A. "What you see may not be what you get: a brief, nontechnical introduction to overfitting in regression-type models". Psychosomatic Medicine 66(3), 411–421 (2004). Shows that overfit models can achieve persuasive in-sample R² values from noise yet fail to replicate in new samples. registry ↩