MicroPulse · Methodology

Methodology & science basis

Rules decide, AI explains. Here is the basis for every threshold in MicroPulse, honestly sorted into three tiers: solid (cited), literature-informed (debated), and tunable default (internal). Nothing here is a black box.

SolidLiterature-informedTunable default

Solid

Estimated 1RM (e1RM)Solid

From a logged set, RIR-aware (RIR = 10 − RPE). Epley is the default; Brzycki/Lombardi available.

Basis: Epley 1985 · Brzycki 1993 · Lombardi 1989 (standard formulas)

src/lib/client/oneRepMaxFormulas.ts

Velocity-based 1RM (LV profile)Solid

1RM from the load–velocity line at each lift's minimum-velocity threshold (e.g. bench 0.15, squat 0.30 m/s).

Basis: González-Badillo & Sánchez-Medina 2010 · Banyard 2017

src/lib/lvProfile/index.ts

PR evidence gate (RPE ≥ 9)Solid

A set only counts as 1RM/PR evidence at RPE ≥ 9 — self-reported effort is unreliable further from failure.

Basis: Zourdos 2016 (RIR) · Foster 1998 (RPE). Conservative by design.

src/lib/client/oneRepMaxFormulas.ts

Confidence & baseline maturitySolid

Needs 8 sessions for a mature baseline, full sensitivity at 14. Thin-data signals show as low confidence, never as a verdict.

Basis: Operationalises the manifesto: show confidence, don't fake it.

src/lib/micropulse/trainingRead/index.ts

Literature-informed

ACWR (acute:chronic workload)Literature-informed

Reference band ≈ 0.8–1.3; we flag elevated fatigue above 1.3.

Basis: Gabbett 2016.

Caveat: Heavily debated since 2019 (Impellizzeri 2020; Lolli 2019; Wang 2020). We present ACWR as workload context, NOT an injury predictor.

src/lib/client/loadQuadrant.ts

Foster monotony & strainLiterature-informed

Monotony = mean/SD of daily load; strain = total × monotony. Deload recommended at monotony ≥1.5 & strain ≥4000; required at ≥2.0 or ≥6000.

Basis: Foster 1998.

Caveat: The formula is Foster's; the composite cutoffs are adapted for a 7-day window and are tunable.

src/lib/trainer/loadIntelligence.ts

HR load vs sRPE cross-checkLiterature-informed

Objective HR load (Edwards' summated-heart-rate-zone method) indexed to the player's own norm and compared to his sRPE. A personal-norm gap beyond ±25 is called diverging: heart harder than the rating (hidden load) or rating higher than the heart (low cardiac demand).

Basis: Edwards 1993 (summated-HR-zone TRIMP) · Buchheit 2024 (HR monitoring).

Caveat: Edwards' 5 zones are adapted to Catapult's 8 ordinal bands weighted by band order (NOT calibrated %HRmax cuts), so the AU is comparable only against the player's own baseline — never as an absolute vs sRPE. The ±25 gap and band weights are tunable. Belt-only: no belt = no data, never 0.

src/lib/micropulse/hrLoad/index.ts

HRmax estimate (for %HRmax)Literature-informed

When HRmax isn't set and no maximal effort is captured, %HRmax uses an age estimate: Tanaka 208 − 0.7·age (men/mixed) or Gulati 206 − 0.88·age (women). We prefer a coach-set value, then the observed belt peak, and mark an age estimate as approximate (“≈”).

Basis: Tanaka 2001 · Gulati 2010. More accurate than the old Fox 220−age.

Caveat: A population estimate (SEE ≈ 7–11 bpm), never a measured value — so an age-estimated %HRmax does NOT raise confidence and isn't stored; only measured/coach-set HRmax is persisted.

src/lib/micropulse/hrLoad/index.ts

Heavy-lift intensity (≥80% 1RM)Literature-informed

A set counts as 'heavy' at or above 80% of 1RM.

Basis: Pareja-Blanco 2017 · Schoenfeld 2017.

src/lib/trainer/loadIntelligence.ts

Tunable default

Heavy-lift weekly cap (25 sets)Tunable default

Soft weekly cap of ~25 sets ≥80% 1RM before flagging fatigue saturation.

Basis: Internal default — not a specific published number.

Caveat: Adjust per athlete and training age.

src/lib/trainer/loadIntelligence.ts

Velocity-loss match-day capsTunable default

Stop-set velocity-loss caps tighten toward a match: MD-5/-4 20%, MD-3 15%, MD-2 10%, MD-1/MD+1 5%.

Basis: Internal taper policy — uncited.

Caveat: A sensible default taper; adjust to your model.

src/lib/micropulse/trainingGraph/schema.ts

Game-model demand weightsTunable default

How each game model (high press, possession, …) taxes each physical quality, used to rank development emphases.

Basis: Internal coaching model — configurable per team.

Caveat: Calibrate with your S&C coach.

src/lib/micropulse/trainingRead/catalogue.ts

Readiness deload defaultsTunable default

A yellow day keeps 85% volume / 90% intensity; a red day becomes recovery.

Basis: Internal defaults — tunable per readiness band.

src/lib/client/readinessAdjust.ts

Movement-signature driftTunable default

Flags a personal-norm drift at ±1.5 SD while calibrating (8–13 sessions), tightening to ±1.0 SD once mature (≥14).

Basis: Method: Robertson 2017 (personal z-scoring). The SD cutoffs are tuning parameters.

src/lib/micropulse/movementSignature/index.ts

Thresholds marked “tunable default” are starting values coaches can change — they are not fixed safety limits. We update this page whenever the basis changes.