Rebuilding Chronic Load From Zero Without Overshooting

The first week back from a running injury is where most plans quietly set up the second injury. Not with a hard session. Not with a race. With arithmetic.

Here’s the mechanism. You ran 4 km on Monday and 4 km on Thursday. Your acute load is up, your chronic load is basically nothing, and your app calculates the ratio between them. Garmin calls the result Acute Load Ratio and will happily colour it red. Strava’s Fitness/Freshness chart shows a Form number that looks alarming. Runna’s plan builder ramps you by percentages of your current weekly volume. Every one of these tools is doing something reasonable with data that cannot support it, because the denominator is close to zero and near-zero denominators make ratios meaningless.

So the practical question for anyone rebuilding: how to rebuild running fitness after injury when the load metrics you rely on are mathematically unusable for the first month. The answer is to stop dividing and start adding. For the first six weeks back, throw out the 10% rule and the 0.8–1.3 ACWR window, and progress on absolute weekly increments instead: fixed kilometres, fixed minutes, decided in advance.

Why the ratio breaks

The 10% rule and acute:chronic workload ratio both assume a stable baseline. The maths works fine at marathon volume. Take a runner holding 60 km/week: a 10% jump is 6 km, roughly one easy run’s worth of extra work spread over seven days. Tissue-level insult, small. Signal-to-noise, high.

Now run the same maths at 8 km/week. Ten percent is 800 metres. Eight hundred metres. That is not a training progression, it’s a rounding error, and no runner on earth is going to execute it: you’d be adding four minutes of jogging per week and reaching 20 km/week sometime in December.

The ratio fails in the other direction too, and this is the dangerous one. ACWR divides your 7-day load by your 28-day rolling average. Four weeks after a layoff, that 28-day average includes three weeks of zero. Suppose you ran 0, 0, 0 and then 10 km. Your chronic load is 2.5 km/week. Your acute load is 10. That’s a ratio of 4.0, which on Garmin’s dial sits far outside the shaded optimal band and which the literature would call “danger zone.” But 10 km in a week, split into three 3.3 km jogs, is an entirely sane first week for most people cleared to run.

Conversely, and worse: run 14 km one week off a 3.5 km chronic base, then 20 km the next. Chronic has climbed enough that your ratio might read 1.25, comfortably inside the green. The absolute jump was 6 km in a week on tissue that hasn’t absorbed impact in two months. Green light, real overload. The ratio drifts into range precisely because your recent history is rising fast, which is the thing you wanted the metric to warn you about.

Garmin’s Acute Load Ratio needs a populated 28-day window before it means anything, and Strava’s Fitness number is an exponentially weighted average with a 42-day time constant. Both are honest about being smoothed. Neither is designed to be read on day 9 of a comeback. That’s not a bug in the tools, it’s you asking a long-window metric a short-window question.

The absolute-increment method

Six weeks. Fixed additions, in kilometres or minutes, chosen once and written down before you start.

Three rules govern that table.

One: pick your increment from absolute volume, not percentage. Under 20 km/week, add 2–3 km per week. Between 20 and 40 km/week, add 4–5 km. Above 40 km, you’re back in territory where percentage rules behave and you can let your app resume driving. The increment is per week of running, not per week of calendar, so a week you skip doesn’t earn you a double jump.

Two: no single run grows by more than 1 km while you’re under 25 km/week. Weekly volume growth and single-session growth are different stresses. The long run is where most comeback injuries actually happen, because that’s where form degrades and cumulative impact count spikes. Week 5 in the table adds 4 km of weekly volume by adding a fourth session, not by extending the longest run at all.

Three: hold every third week flat. Week 4 repeats week 3 exactly. This is not a taper, it’s a consolidation week: same volume, same session count, giving tendon and bone the chance to complete an adaptation cycle rather than being chased into the next one. Bone remodelling runs on a roughly 3-week early timeline, much slower than the cardiovascular adaptations that will make you feel ready sooner than you are.

Time-based works identically if you prefer it, and for anyone coming back from a calf or Achilles problem I’d argue it’s better because it stops you chasing pace. Substitute minutes: week 1 is 3 × 20 min, week 2 is 3 × 25 min, week 3 is 3 × 30 min, week 4 repeats, week 5 adds a fourth 25-minute run. Your increment is +10 to +15 min/week under 120 min/week.

Making your app cooperate

The plans are not built for this, so you have to override them, and the overrides differ by tool.

If you’re on Runna, its plans start from a target race and a current volume you declare at signup. Declare it honestly low. If you tell it 25 km/week because that’s what you were doing in June, week 1 arrives with a 7 km long run and a tempo block. Set your starting volume at your actual week-1 number (9 km, in the example above), pick the longest plan duration available, and skip or shorten the quality sessions for the first three weeks regardless of what the app prescribes. Runna’s “easy run” distances are the ones to keep.

Garmin Coach (and Daily Suggested Workouts) reacts to your last seven days, which during a rebuild means it chases you upward. It sees you complete 5 km comfortably, your HR was fine, so it suggests 7 km with a bit of threshold. Two things help: turn off Daily Suggested Workouts entirely for six weeks, or build the block as a manual structured workout schedule in Garmin Connect so the calendar has your numbers rather than the algorithm’s. Also ignore Training Readiness and Training Status during this period. Both will read “Unproductive” or “Detraining” for weeks because your VO2max estimate is stale, and neither says anything about tissue tolerance.

A ChatGPT-built plan is the most flexible and the most likely to overshoot, because it will produce whatever shape you asked for and its default ramp is aggressive. Give it the constraints explicitly, in absolute terms. Something like: “Build six weeks. Week 1 total 9 km across 3 runs. Add exactly 3 km per week through week 3. Week 4 identical to week 3. Weeks 5 and 6 add 4 km each, adding a fourth session rather than lengthening the long run. No run exceeds 6 km. No tempo, threshold or interval work.” Then check its output against the numbers rather than trusting the prose, because plans generated this way frequently include a week that violates the constraint you just stated. Add up each week yourself.

For the full clinical picture on what should happen before week 1 (loading tests, pain rules during and 24 hours after, walk-run structure), the return-to-running after injury guide covers the ground this post assumes you’ve already cleared.

Reading your data when the metrics are useless

You still want feedback during those six weeks. It just has to come from somewhere other than load ratios.

Watch pace at a fixed heart rate, not pace. Pick your easy-run HR ceiling (many people use roughly 70–75% of max, or the top of zone 2), keep every run under it, and record the average pace. A typical comeback trajectory looks like this in Strava’s activity list:

Wk 1  3.0 km  6:52/km  avg HR 141
Wk 2  4.0 km  6:44/km  avg HR 139
Wk 3  5.0 km  6:38/km  avg HR 142
Wk 4  5.0 km  6:29/km  avg HR 140
Wk 5  5.0 km  6:31/km  avg HR 141
Wk 6  6.0 km  6:25/km  avg HR 142

That’s the shape you want: pace improving 5–10 sec/km per week at constant HR, with the odd flat or backwards week. Roughly 20 seconds gained over six weeks at the same cardiac cost is a healthy rebuild. If pace is deteriorating at constant HR across two consecutive weeks, you’re accumulating fatigue faster than you’re adapting, and the fix is to repeat a week rather than push through.

The second signal is morning symptom check on the injured tissue: a simple 0–10 stiffness score, logged before you get out of bed. Two consecutive mornings above your baseline means you repeat the current week instead of incrementing. This is the only metric in the whole method that can stop a progression, and it beats every number your watch produces because it measures the actual tissue you injured.

Cadence deserves a look too. If your comeback cadence is 8+ steps/min below your pre-injury figure at the same pace, you’re likely running with a longer, more braked stride than you used to, which raises impact peak per step. It usually self-corrects by week 4. If it hasn’t by week 6, that’s worth a physio conversation.

Where week 7 goes

At the end of six weeks the runner in the example table is at 23 km/week off four sessions, with a 6 km long run. Chronic load now has real data behind it: 28 days of continuous running, no zeros in the window. Strava’s Fitness line has a genuine slope. Garmin’s Acute Load Ratio is computing against a populated baseline instead of three weeks of nothing.

That’s the handover point. From week 7 you can let your app’s ratio logic take over, because the denominator finally means something, and you can reintroduce one quality session per week (strides first, then a short threshold block). The absolute-increment phase exists to build the baseline the ratios need. It isn’t a better training philosophy; it’s the scaffolding you take down once the structure can hold itself up.

What tends to happen if you skip it is not a dramatic reinjury in week 2. It’s a slow accumulation: week 3 feels great so you run 24 km instead of 15, week 4 feels fine too, and the same tendon goes at week 6 when your app was still showing a comfortable 1.1 ratio the whole way. The numbers were never wrong. They were just answering a different question than the one you needed asked.