52 terms, defined the way Base uses them.

Acute-to-chronic workload ratio

This week’s training load divided by the average of recent weeks, used to catch a load spike before it becomes an injury.

Base computes it on endurance minutes against the previous four weeks. The 0.8 to 1.3 window is a productive build; past about 1.5 the injury-risk curve turns sharply upward. Below 0.8 is reported as neutral rather than as a warning, because a taper week and a rest week both live there and both are deliberate.

Adaptive maintenance

A maintenance-calorie figure measured from what your bodyweight actually did against what you actually ate, rather than predicted from your height and weight.

Over a trailing 14-day window ending yesterday, Base takes the change in trend weight, converts it to a daily energy balance at roughly 3,500 kcal per pound, and subtracts that from your average logged intake. The result is smoothed with an exponentially weighted moving average so it does not lurch on one bad weigh-in. It needs at least ten logged days and eight weigh-ins in the window; below that it falls back to a Mifflin-St Jeor estimate and labels itself "estimated" rather than presenting a guess as a measurement.

How adaptive maintenance works

Aerobic base

The capacity to sustain low-intensity work for a long time, built almost entirely by accumulating easy volume.

It is the part of endurance fitness that responds slowest and decays slowest. The most common way to fail to build it is to run the easy days too hard, which is precisely what the session grade exists to catch.

Autoregulation

Adjusting the load or the volume of a session based on how you actually performed that day rather than on what was written down weeks earlier.

The meta-analytic evidence favours autoregulated progression over fixed-percentage progression for trained lifters. RIR is the input that makes it possible: without it there is no way to tell a rep from a rep that nearly failed.

Baseline

Your own recent history, used as the reference a reading is judged against.

Base never ranks a health number against a population. An HRV of 48 ms is good or bad only relative to your own thirty days, and a percentile against other users would rank who else happens to use the app.

BMR

Basal metabolic rate: the energy your body spends at complete rest, before any movement at all.

Distinct from resting energy expenditure, which Apple Health reports and which includes the small cost of being awake and digesting. Base uses Mifflin-St Jeor for BMR only as a fallback and as a sanity clamp on the projected resting figure — the number it actually steers by is adaptive maintenance, because that one is measured.

Carbohydrate periodisation

Varying carbohydrate intake day to day so the biggest intakes land on the hardest sessions.

The lever a hybrid athlete has that a pure lifter does not need and a pure endurance athlete uses differently. Base does not prescribe it, but the day-template system is what makes running it practical: a training-day template and a rest-day template, one tap each.

Coggan zones

The standard cycling power bands expressed as fractions of functional threshold power.

Base uses them directly for rides, because they are what every cyclist and every head unit already means by "zone 3": easy 0.55 to 0.75 of FTP, long 0.60 to 0.75, tempo 0.76 to 0.87, threshold 0.95 to 1.05, intervals 1.06 to 1.20.

Confidence

A label on an estimate saying how much weight it will bear.

Threshold estimates carry one — high, medium or low — based on how recent and how long the effort behind them was. A number without provenance is a number the athlete has no way to check, and every prescribed zone is downstream of this one estimate.

Day template

A whole day of food saved as one unit and logged in a single tap.

The core of Base’s food logging, and the reason it does not present a search box first. A serious athlete eats close to the same thing most days; a template replay turns a five-minute logging session into one tap, which is the difference between logging for six months and logging for three weeks.

Deload

A planned reduction in training stress to let accumulated fatigue clear.

The 2024 survey of strength and physique athletes has them deloading every 5.6 weeks give or take 2.3, but a reactive trigger beats a calendar one. Base deloads when two independent fatigue signals fire at once — depressed recovery score, RIR drifting up at fixed load, short sleep, flat strength trend — with week six as a backstop rather than the rule.

Double progression

Adding reps within a rep range until the top of it is reached at the target effort, then adding load and resetting to the bottom of the range.

Load progression and rep progression produce equivalent hypertrophy over eight weeks, so the choice between them is about adherence rather than physiology. What makes double progression work is the effort clause: twelve reps at 0 RIR earns the load increase, twelve reps at 4 RIR means the load was too light and should jump immediately.

e1RM

Estimated one-rep max: what a set of submaximal reps implies you could lift once.

Base uses the Epley formula, weight multiplied by one plus reps over thirty, and only for sets of twelve reps or fewer — the estimate degrades badly at high rep counts. Its value is as a trend line across a training block, not as a number to quote.

Energy availability

The energy left for basic physiological function after training has taken its share, expressed per kilogram of fat-free mass.

The concept behind RED-S. Low energy availability is the specific mechanism by which a hybrid athlete who adds endurance volume without adding food loses bone density, sleep quality, hormonal function and — eventually — the muscle they were training to build. It is the single strongest argument for tracking endurance calories and food in the same place.

RED-S and hybrid training

Fatigue

The accumulated cost of training that reduces performance without reducing fitness.

It matters here because it is the shared account that lifting and endurance both draw on. A hybrid athlete rarely fails from too little stimulus; they fail from stimulus arriving faster than fatigue clears.

Fractional set

A way of counting training volume that credits a set fully to the muscle doing most of the work and half to the muscles assisting.

Direct sets plus 0.5 times indirect sets. Pelland et al. found this model fit the hypertrophy dose-response data better than counting either only direct sets or all sets equally. It matters for a hybrid athlete because it changes what "twelve sets for quads" means once you have also done a hard hill session.

FTP

Functional threshold power: the highest power a cyclist can sustain for roughly an hour, in watts.

Estimated in Base from your best sustained effort of at least twenty minutes, scaled by how long that effort already lasted and reaching a factor of 1.0 past about fifty minutes. A recent proper test always beats an inference from training data, which is why a manual override wins.

HRV

Heart rate variability: the variation in time between consecutive heartbeats, in milliseconds.

Base reads the SDNN value Apple Watch computes overnight rather than deriving its own, and uses it as the largest single input to Recovery Score at 50% of the mix. It is only meaningful against your own baseline — comparing an HRV number between two people says more about their ages and genetics than their readiness.

Hybrid training

Training for muscular size or strength and for endurance at the same time, on purpose, with both as real goals.

Distinct from a lifter who does some cardio and from a runner who lifts twice a week for injury prevention. The defining feature is that both halves are programmed and both are expected to improve, which is what creates the recovery and energy conflicts nothing else on the market measures together.

The full pillar page

Hypertrophy

Growth in the cross-sectional area of muscle fibres.

Distinct from strength, which also depends on neural factors, technique and leverages. The two diverge often enough that Base tracks e1RM and volume separately rather than treating either as a proxy for the other.

Interference effect

The idea that endurance training blunts the strength and size gains from lifting.

Real but much smaller than folklore claims. Three recent meta-analyses found no significant compromise to strength or hypertrophy overall. What does modulate it: modality (running interferes, cycling does not at statistical significance), timing (same-session effect size around -0.66, but around -0.10 when separated by six hours or more), and volume.

What the interference research actually says

Lactate threshold

The intensity above which blood lactate accumulates faster than it clears.

The physiological event that "threshold" in a training plan is a practical proxy for. Base does not measure lactate and does not pretend to; it estimates the pace or power that corresponds to it and says how confident that estimate is.

Lean ceiling

The maximum rate at which your body can add lean tissue, which caps how much of a calorie surplus can go anywhere except fat.

Base models a pound of lean tissue as costing roughly 2,500 kcal — far less than a pound of fat at 3,500, because lean tissue is mostly water. At a 0.5 lb/week lean ceiling that is about 179 kcal a day. A surplus larger than the ceiling has nowhere to go but fat, which is the entire argument for a +250 surplus over a +500 one.

Leucine threshold

The amount of the amino acid leucine in a meal needed to maximally trigger muscle protein synthesis.

Roughly 25 to 35 grams of quality protein clears it for most people, and the anabolic response to a meal lasts two to two and a half hours. It is why distribution matters: the same 160 grams produces a different result across two meals than across four.

Macros

Protein, fat and carbohydrate, the three energy-yielding nutrients.

Base sets protein at roughly 1.0 g per pound of bodyweight, fat at 25% of the calorie target with a floor near 0.3 g per pound, and carbohydrate as the remainder — which is what makes it move with the endurance work rather than being fixed.

Micronutrient coverage

How much of the day’s logged food Base actually knows the micronutrients for.

Shown explicitly rather than assumed, because a micronutrient total computed over 40% of your food is not a micronutrient total. Base tracks eighteen nutrients selected on two bars: shortfall demonstrably costs training adaptation, or shortfall costs health and a real share of adults are short.

NEAT

Non-exercise activity thermogenesis: everything you burn moving that was not a workout.

Fidgeting, walking to the kitchen, standing. It is the most variable component of daily expenditure between people and within one person across a week, and it is the main reason a formula-based calorie target drifts out of true. Adaptive maintenance absorbs it automatically because it measures the outcome rather than predicting the inputs.

P-ratio

The proportion of weight gained or lost that is lean tissue rather than fat.

Higher body fat at the start of a surplus tends to partition worse. It is also not something you can measure at home with any precision, which is why Base steers by trend weight and waist rather than by any body-fat estimate.

Polarised training

A distribution that puts most endurance volume well below threshold and a small share well above it, with little in between.

The contrast is with threshold-heavy or pyramidal distributions. Base does not enforce a distribution — it prescribes the session you chose and grades whether it landed where it was supposed to, which is the input a distribution decision actually needs.

Progressive overload

Increasing the demand on a muscle over time so it has a reason to keep adapting.

The demand can rise through load, reps, sets, range of motion or proximity to failure. What is not optional is that it rises, and that it is recorded — an unrecorded session cannot be progressed from.

Protein target

How much protein to eat, expressed per unit of bodyweight.

Base targets 1.0 g per pound, about 2.2 g/kg. Morton et al. 2018 put the breakpoint at 1.62 g/kg with the confidence interval extending to 2.2; Base deliberately takes the top of that interval, because the effect is enhanced by training experience and in a surplus there is no cost to sitting at the upper bound.

Ramp rate

How fast endurance load is climbing, expressed as this week against the recent norm.

Total minutes is a deliberately crude measure of training stress: it needs no power meter, no heart-rate strap and no lab test, and the thing it is catching — a large sudden jump — is coarse enough that a coarse measure finds it.

Readiness

A decision, not a score: what today’s session should ask for given how recovered you are.

Base turns recovery data into two concrete outputs — how many reps to add to today’s RIR target, and whether to deload — rather than leaving a number on a screen for the athlete to interpret.

Recovery Score

A 0 to 100 figure built from your own thirty-day baselines for HRV, resting heart rate and sleep.

Weighted 50% HRV, 30% resting heart rate inverted so lower is better, and 20% sleep score, each expressed as a z-score against your own history and renormalised if a signal is missing. It needs fourteen days of baseline; below that Base says "building baseline" rather than showing a number it cannot yet mean.

Refeed

A deliberate return to maintenance calories after a period of deficit, before attempting a surplus.

Expect three to six pounds on the scale within two weeks, and none of it fat: refilling muscle glycogen binds roughly three grams of water per gram of glycogen, and the digestive tract refilling adds another pound or two. This is the single most likely moment for an athlete to conclude the plan is failing while it is working exactly as intended.

Resting heart rate

The lowest sustained heart rate the watch observes when you are genuinely still, usually overnight.

A rise of five or more beats per minute over your baseline is usually the first sign of illness, and it typically shows up a day before you feel it. Base labels which day a reading came from, because Apple often writes it hours late.

Reverse dieting

Raising calories very slowly out of a deficit over eight to twenty weeks, on the theory that it minimises fat regain.

Not well supported. Metabolic adaptation largely dissipates simply by eating at maintenance; crawling upward mostly delays the point at which anything useful starts happening. Base goes to maintenance fast and holds there instead.

Riegel formula

A way of predicting performance at one distance from a result at another, using an exponent of roughly 1.06.

Base uses it to project a shorter hard effort to a one-hour equivalent when estimating threshold pace. Without that projection a twenty-minute effort would set every zone too fast, because twenty-minute pace is genuinely faster than hour pace.

RIR

Reps in reserve: how many more reps you could have completed before failure.

The evidence puts the useful band at 0 to 4 RIR, with most work at 1 to 3. Training to failure is not superior for hypertrophy and costs disproportionate systemic fatigue on compounds. Base prescribes 2 to 3 RIR on compounds and 0 to 1 on isolation work, and raises the target by a rep when recovery is compromised.

RPE

Rate of perceived exertion, usually on a ten-point scale where 10 is momentary failure.

The inverse of RIR for practical purposes: RPE 8 is roughly 2 RIR. Base records RIR rather than RPE because the question "how many more could you have done" is answered more consistently than "how hard was that out of ten".

Session grade

A verdict on whether an endurance session landed in the band it was prescribed in.

Base grades on the best evidence available and says which it used: rep by rep against the laps the watch recorded, or time-in-zone across the measured session, or — only when neither is available — the session average. The average is last because it hides the run that went out too hard and limped home: the halves cancel and the mean lands inside a band the run was almost never in.

Show the math

The rule that every derived number in Base can open a plain-English explanation of the formula that produced it.

It is a product constraint, not a feature: a number an athlete cannot interrogate is a number they will eventually stop trusting, and the formulas here are opinionated enough that hiding them would be dishonest.

Sleep efficiency

Time asleep divided by time in bed.

One of three inputs to Sleep Score at 40% of the mix, alongside duration at 40% and bedtime consistency at 20%. It is also where awake time is judged — Base does not rate awake time directly, because there is no defensible target band for it.

Sleep Score

A 0 to 100 figure combining how long you slept, how efficiently, and how consistent your bedtime was.

Duration is measured against an eight-hour ideal, consistency against your own fourteen-day average bedtime with a ninety-minute tolerance. No percentile is offered, because the score is Base’s own construction and there is no population that computes it to rank you against.

Surplus

Eating above maintenance so there is spare energy available to build tissue.

For a hybrid athlete the surplus has to be defined against a maintenance figure that already includes the endurance work, or it is not a surplus at all — it is a rounding error on an unmeasured deficit.

TDEE

Total daily energy expenditure: everything you burn in a day, resting and moving.

The word gets used for two different things: a formula prediction from your body size and an activity multiplier, and a measured figure derived from your own intake and weight change. Base means the second. The first is only a fallback.

Threshold

The one number every endurance zone in Base is a percentage of: about the hardest effort you could hold for an hour.

Expressed as pace per mile for running and as functional threshold power in watts for cycling. Run and ride never mix — different units measuring different capacities — and a combined "endurance score" could not answer either question.

Time in zone

The share of a session’s measured duration that fell inside the prescribed intensity band.

Used to grade continuous sessions. Base counts a session "nailed" at 80% of time in band, or 70% of reps for a structured session, and treats a few percent outside as "close" because that is inside normal GPS and power-meter error.

Trend weight

A smoothed bodyweight that filters out day-to-day water, food volume and glycogen noise.

Base uses an exponentially weighted moving average with an alpha of 0.25, so each new weigh-in moves the trend by a quarter of its distance. Raw scale weight can swing four pounds across two days on identical body composition; the trend is what any decision should be made against.

VO2 max

The maximum rate at which you can consume oxygen during exercise.

A ceiling on aerobic capacity, and the number most watches estimate. Base does not build anything on it: it is estimated with wide error from wrist data, and threshold — which is trainable, measurable from an ordinary hard session, and directly prescriptive — answers every question a plan actually asks.

Volume

How much work was done, which means at least three different things.

Tonnage is sets times reps times load, in pounds or kilos. Hard sets per muscle per week is the dose that actually drives hypertrophy in the dose-response literature. Session volume is neither. Base shows tonnage for a session because it is what changed since last week, and fractional sets per muscle per week because that is what the research is denominated in.

Waist-to-height ratio

Waist circumference divided by height, used as a cheap, repeatable proxy for change in abdominal fat.

It is used here for its trend, not its absolute value. Alongside trend weight it is what separates "the surplus is going where it should" from "the surplus is too big" — a scale number alone cannot tell those apart.