The body, the record,
and the feedback loop.

The connection is an interface between systems.

The Bishop Protocol: Foundations organizes resistance training through preparation, two top sets and a separate reduced-load finisher with the Bishop Cadence. It distinguishes experience levels and presents a four-week, three-day full-body programme. That is the physical practice the digital project needs to represent accurately.

MMG’s job is to preserve meaning as information moves between the book, the person and the trainer. A useful record identifies the source version, the selected exercise, what was prescribed, what was actually completed and who reviewed it. Computation can transform those inputs into transparent summaries. AI, when connected and selected, can help explain source material. The person and trainer remain responsible for interpreting context and making changes.

The shared interface is the training record: what was intended, what was observed, and what a person decided next.

Physiology: stimulus, response and recovery.

Physiology concerns the living processes involved in exercise: neuromuscular activation, fatigue, energy use, recovery and adaptation. The current ACSM resistance-training position stand examines prescription variables in relation to outcomes including strength, hypertrophy, power and physical function. This supports careful attention to training variables; it does not validate the entire branded Bishop method. ACSM’s 2026 research context.

A log supplies partial observations. More repetitions recorded at the same load may be useful in a review, but technique, equipment, fatigue and reporting accuracy also matter. A completed checkbox does not measure muscle growth or recovery. The interface must keep recorded performance separate from inferred biological adaptation.

Kinesiology: describe the movement precisely.

Kinesiology connects the programme to movement: exercise selection, range of motion, tempo, external load and hold position. The guide distinguishes stretched and contracted positions by exercise. Software therefore needs exercise-specific instructions rather than one generic hold cue.

A record also needs units. Thirty kilograms is not thirty pounds; a timed plank is not a set of repetitions. Warm-ups, top sets and the separate finisher must remain different set types. These are data-design decisions that protect the meaning of the instruction. Text entry cannot establish joint angles, joint moments, muscle activation or correct form. Those require appropriate observation or measurement.

Psychology: support behaviour with context.

The COM-B framework describes capability, opportunity and motivation as conditions involved in behaviour. Applied as a design lens, understandable instructions may support capability; realistic planning and access may support opportunity; reflection on personally chosen goals may support motivation. These are proposed product mechanisms to evaluate, not proof that the app changes behaviour or improves health. Michie, van Stralen and West’s original framework.

For Bishop Pro, the practical question becomes: what prevented or helped this person carry out the plan? Optional reflections and trainer conversations can add context. A missed session should not automatically become a label about motivation, and a streak should not be presented as a physiological result.

Computation: a reproducible estimate.

The working calculator adds Mifflin–St Jeor to the project at the user’s request. The original research estimated resting energy expenditure using age, height, weight and sex-specific coefficients, comparing predictions with indirect calorimetry in healthy adults aged 19–78. The equation is a prediction; the website does not measure metabolism. An existing measured resting result can instead be entered and clearly labelled. Mifflin and colleagues, 1990.

Activity is a second, explicit assumption. Multiplying resting energy by a chosen activity factor produces a rough daily expenditure estimate. NIDDK describes typical activity levels from 1.4 to 2.5, but our simple multiplication does not reproduce its dynamic Body Weight Planner. Neither output automatically establishes how much someone should eat. NIDDK activity-level context.

The book provides a general protein range of 1.6–2.2 grams per kilogram per day. The optional calculator output shows that arithmetic and its source. It does not assess individual suitability. Calorie deficits, complete macro prescriptions and food-delivery menus have not been defined by this prototype.

AI: source explanation within a reviewed process.

Guide retrieval, deterministic arithmetic and generated AI explanations are separate functions. Retrieval finds relevant text. The calculator follows a stated equation. A connected language model could summarize selected records or explain a passage, with its source available for checking. The current prototype does not have a connected AI provider.

A useful AI output should expose missing context and distinguish a quote, an estimate and a suggestion. It should not silently resolve contradictory instructions, diagnose a condition or claim to have watched a repetition. Saving a workout does not retrain the language model. The project is a digital support system, not a validated digital twin of the body.

Engineer the feedback loop.

1. Plan: identify the source version and trainer-reviewed instructions.
2. Practice: the person performs the physical work and recovers.
3. Observe: record completed work and clearly label self-report, measurements and estimates.
4. Represent: validate units, preserve provenance and show reproducible calculations.
5. Review: bring the record and any explanation to the client and trainer.
6. Adjust: save the reviewed decision and its reason for the next session.

In engineering terms, inputs are supplied information; state is the current stored context; observations are the limited signals available; outputs are calculations, explanations and records. Feedback means those observations inform the next decision. Sampling and latency matter: a record made weeks later offers different information from one made during a session. Observability is limited because many physiological processes are unmeasured. This design is not a proof of a mathematically stable biological controller.

Choose KPIs that answer a real question.

The creator-review workbench proposes three modest measures for one exercise: complete top-set records, total recorded repetitions and external load multiplied by repetitions. The last is a bookkeeping proxy, not mechanical work, caloric expenditure or muscle growth. Comparison requires consistent exercise, equipment, technique and load convention.

A future trainer view could compare prescribed and completed work, show adherence over an agreed period, retain recovery context and record the trainer’s progression decision. Definitions, missing-data rules and review thresholds need the protocol creator’s feedback. Higher numbers are not automatically better, and form remains a human review point.

What the book can and cannot define yet.

Foundations supplies experience categories, weeks and days, exercise grids, set structure, load adjustments, repetitions, cadence holds and recovery guidance. These can inform a structured intake and session record. Goals, equipment availability, optional effort/recovery reports, consent and assigned trainer are additional MMG design fields.

Before automatic programming, the creator needs to reconcile source differences. The week-one overview says one finisher exercise per workout, while some corresponding grids mark two. The core method describes one to two cadence cycles, while other sections include two to three. The experience categories also overlap at three years. The prototype leaves those prescriptions for review instead of generating an authoritative answer from conflicting text.

Scale the method while keeping responsibility clear.

The proposed network separates a versioned shared method and catalogue from each trainer’s assigned clients and schedule. Protocol ownership, trainer participation and permitted scope require agreement. MMG can operate the digital experience, while the existing Rubare merchant handles product checkout. An ebook order does not approve a trainer or reserve an appointment.

Today the original PDF is available for CAD $39. The public project has the working energy calculator, systems explorer and temporary KPI review example. The private companion has guide search and training records. Public appointment booking remains closed; access and trainer interest go through MMG email. Durable KPI tracking, approved trainer accounts, generated AI explanations, subscriptions and food delivery remain later implementation stages.

This prototype is designed to make the next conversation concrete: the creator can try the calculations, inspect the proposed measurements and revise how the digital layer represents the method.