How is recovery score calculated?

From your own thirty-day baselines for HRV, resting heart rate and sleep, each expressed as a z-score against your own history, weighted 50 percent HRV, 30 percent resting heart rate inverted so lower is better, and 20 percent sleep score. The weighted z-score becomes a 0 to 100 figure via 50 plus 15 times z, clamped. If a signal is missing the remaining weights are renormalised.

Why does Base need fourteen days before showing a recovery score?

Because the score is a z-score against your own baseline, and a baseline computed from four days has a standard deviation that is mostly noise. Below fourteen days it says building baseline with a countdown rather than showing a number that would be arithmetic rather than information.

When should I deload?

When two independent fatigue signals fire at once: recovery depressed against your baseline, reps-in-reserve drifting up at fixed load, three or more nights under seven hours in the last week, or a flat strength trend. One signal is noise. Two is a pattern. Week six is a backstop for when the signals get missed, not the primary rule.

Does a low recovery score mean I should skip training?

Almost never. It means training slightly further from failure — Base raises today's RIR target by a rep rather than cancelling the session. Under real fatigue the same bar is further from failure than it feels, and holding the usual target turns a productive session into a missed rep and a hole to climb out of.

The Recovery Screen: A Decision, Not a Number

Q: Is HRV worth tracking?

Against your own baseline, yes — it is the largest single input here at 50 percent. Against anyone else's, no. HRV varies enormously between people for reasons of age and genetics that no amount of training changes, so a comparison across people ranks demographics rather than readiness.

How the score is built

zHRV   = (todayHRV - mean30d) / sd30d
zRHR   = (mean30d - todayRHR) / sd30d      // inverted: lower RHR is better
zSleep = (todaySleep - mean30d) / sd30d
z      = 0.5*zHRV + 0.3*zRHR + 0.2*zSleep  // renormalised if a signal is missing
score  = round(clamp(50 + 15*z, 0, 100))

Bands: 0–33 red, 34–66 yellow, 67–100 green.

Three decisions in there are worth stating.

Every input is a z-score against your own thirty days. Not a population chart. HRV varies enormously between people for reasons of age and genetics that no training changes, so a cross-person comparison ranks demographics rather than readiness. Your 48 ms is good or bad only relative to your own record.

Resting heart rate is inverted. Lower is better, so the z-score is computed as baseline minus today rather than the reverse. A sign error here would be silent and catastrophic.

Missing signals renormalise rather than zeroing. If HRV is unavailable, the remaining weights are rescaled to sum to one. Treating a missing signal as a zero z-score would drag every score toward 50 and make the whole thing mush.

Fourteen days of baseline required. Below that the screen says "building baseline — 6 more days". A standard deviation from four days is mostly noise, and dividing by it produces numbers that look like information and are not.

Output one: today's RIR target

Under real fatigue the same bar is further from failure than it feels. A set you would normally take to 2 reps in reserve becomes a missed rep — and a missed rep on a compound is disproportionate systemic fatigue for zero extra stimulus.

So Base adds reps to the target rather than removing the session:

Readiness	RIR offset	What it means
Primed	0	The protocol band as written
Normal	0	The protocol band as written
Compromised	+1	Stop one rep further from failure

The offset is never negative. No recovery score is good enough to justify prescribing work closer to failure than the protocol band. A green morning is permission to execute the plan, not to exceed it — that asymmetry is deliberate, because the downside of one over-aggressive session is much larger than the upside.

The reason this is the output rather than "rest today" is that skipping is almost always the wrong response to a moderately low score. Training slightly further from failure preserves the stimulus, preserves the habit, and costs almost nothing.

Output two: the deload trigger

Four fatigue signals, and two firing at once is the bar.

Signal	Threshold
Recovery score depressed	Below 40 against your own baseline
RIR drifting up at fixed load	Slope above 0.5
Short sleep	3+ nights under 7 h in the last week
Flat strength	e1RM trend below 0.1% per week

One signal is noise. Two is a pattern.

Week six is a backstop, not the rule. The 2024 survey of strength and physique athletes has them deloading every 5.6 weeks give or take 2.3 — a useful average and a poor prescription, because fatigue does not arrive on a schedule and a calendar deload lands either too early or too late almost every time.

A reactive trigger is better precisely because the signals are already being measured. The calendar is there for when the signals get missed.

What this means for a hybrid athlete specifically

The signals do not know which discipline caused the fatigue, and that is a feature.

A lifting-only fatigue model misses the athlete whose problem is a spiked endurance week. An endurance-only model misses the one whose problem is three heavy lower-body sessions. Because HRV, resting heart rate and sleep are systemic, they catch both — and a hybrid athlete's actual problem is almost always that the two loads combined exceeded what either alone would have.

This is why ramp rate sits at the top of the endurance screen. A spike week is both an injury risk and a direct draw on the recovery budget your lifting is spending, and the recovery score is where you see the withdrawal.

What the screen will not do

It will not diagnose illness. A resting heart rate five or more beats above baseline is often the first sign of illness and usually shows up a day before you feel it — the screen surfaces the reading and says what it commonly means. It does not tell you that you are sick. This is a training log, not a medical device.

It will not show a number without a baseline. Fourteen days or nothing.

It will not rank you against other people. There is no percentile here, because the score is Base's own construction — nobody else computes this number, so there is no population to rank against. Why refusing a number is also showing the math.

Reading it honestly

One low reading is one bad night. Alcohol, a late meal, a warm room and a poor night's sleep all depress HRV, and none of them means your training block is failing.

What matters is the trend against your own baseline over seven to ten days, and whether a second signal is firing alongside it. A week of depressed recovery with a flat strength trend is a deload. A single red morning after a wedding is a wedding.

What to do next

The sleep screen covers the 40/40/20 split behind the sleep component, and recovery for hybrid athletes covers what to do with a compromised week beyond the RIR offset.