The Walk-Run Protocol for Coming Back From Injury
Most injured runners don’t get hurt again because they came back too early. They get hurt again because they came back too continuously. The tissue was ready for eight minutes of running. They gave it thirty-five, because thirty-five is what the plan said, and the plan didn’t know they’d spent six weeks not running.
The walk-run protocol solves this by breaking the run into doses. You run for a set interval, walk for a set interval, and repeat. Total volume goes up week to week, but the length of any single continuous running bout goes up separately and more slowly. That separation is the whole point, and it’s the thing almost no training app does automatically.
This page covers how to build and progress a walk-run protocol after injury, how to load it into Garmin, Coros or a Runna plan that wasn’t designed for it, and how to read your own data to decide whether to progress, hold or drop back. For the wider picture of return-to-running decision-making (when to start at all, pain rules, strength work, how to talk to a physio about your plan) see the return-to-running after injury overview.
Why Intervals Beat Continuous Running Early On
Running tissue injuries are almost always cumulative. Bone stress, tendinopathy, plantar fascia problems: they develop because load arrived faster than the tissue’s repair rate, not because one stride was too hard. Walking breaks interrupt the accumulation. They let calf muscles unload for 60 to 90 seconds, which changes the shock-absorption picture for the next running bout, and they cap the number of consecutive loading cycles any single structure sees.
There’s a mechanical argument too. When you fatigue, your form degrades in predictable ways: cadence drops, ground contact time lengthens, vertical oscillation often climbs. If you have a Garmin with a running dynamics pod or an HRM-Pro strap, you can watch this happen in the lap-by-lap data. A runner returning from Achilles tendinopathy will frequently show ground contact time creeping from 250ms in lap one to 275ms by lap six of a continuous 30-minute run. Those later laps are where the reinjury lives. Walk breaks keep you out of them.
The counterargument you’ll hear is that walk-run “isn’t real running” and delays fitness. In the first four to six weeks back, that trade is worth making every time. Aerobic fitness comes back fast (it’s mostly still there). Tendon and bone adaptation takes 8 to 12 weeks minimum. The limiting factor isn’t your heart.
Setting Your Starting Point
Two numbers define the start: the run interval and the number of repeats.
Run interval should be short enough that it feels almost silly. For most runners coming off two to eight weeks of no running, 1 minute is right. For longer layoffs, or a bone stress injury, 30 seconds. Walk interval is typically 1 to 2 minutes, and it should be a genuine walk, not a shuffle.
Repeats should produce a total running time of roughly 8 to 12 minutes in that first session. Here’s what that looks like:
| Layoff length | Injury type | Session 1 |
|---|---|---|
| 2–3 weeks | Soft tissue, resolved | 8 × (2 min run / 1 min walk) |
| 4–6 weeks | Tendinopathy | 10 × (1 min run / 1 min walk) |
| 6–10 weeks | Tibial bone stress | 12 × (30 sec run / 90 sec walk) |
| 10+ weeks | Femoral/pelvic bone stress | 8 × (30 sec run / 2 min walk), physio-led |
Run these every other day. Three sessions per week, never two days consecutively, for at least the first three weeks. The rest day is when adaptation actually happens, and it’s also your diagnostic window: pain that shows up 24 hours after a session is far more informative than pain during it.
Progressing: Two Dials, Not One
This is where most self-built and AI-built plans go wrong. They increase everything at once.
You have two dials. Total running volume (minutes of running per session) and continuous bout length (the run interval itself). Move one per week, not both.
A worked 8-week progression for a runner returning from mid-portion Achilles tendinopathy after five weeks off, three sessions a week:
Week 1 10 × (1:00 run / 1:00 walk) 10 min running
Week 2 14 × (1:00 run / 1:00 walk) 14 min running
Week 3 7 × (2:00 run / 1:00 walk) 14 min running <- bout up, volume held
Week 4 9 × (2:00 run / 1:00 walk) 18 min running
Week 5 6 × (3:00 run / 1:00 walk) 18 min running <- bout up, volume held
Week 6 8 × (3:00 run / 1:00 walk) 24 min running
Week 7 5 × (5:00 run / 1:00 walk) 25 min running <- bout up, volume held
Week 8 6 × (5:00 run / 1:00 walk) 30 min running
Notice the pattern: volume week, bout week, volume week, bout week. Weeks 3, 5 and 7 look like non-progression on a Strava weekly-mileage graph, and that’s exactly why relying on that graph will mislead you. Total running time is flat, but you’ve doubled, then tripled, then quintupled the continuous loading bout. That’s a large increase in demand that a mileage number cannot see.
By week 8 you’re doing 30 minutes of running in a 35-minute session. From there, the standard exit is to reduce the walk to 30 seconds, then take it out entirely on one session per week while keeping walks in the other two.
Getting This Onto Your Watch
Doing walk-run by glancing at your wrist and guessing works badly. You will run 90 seconds when you meant 60, every time, and you’ll skip walk breaks when you feel good. Build the workout properly.
Garmin. Garmin Connect, Training > Workouts > Create Workout > Run. Add a Run step with duration 1:00, then a Recover step with duration 1:00, select both, and choose Repeat with your rep count. Save it, send it to your watch, and it will buzz at you. Set the run step target to “no target” rather than a pace or HR range, because chasing pace on a 60-second interval defeats the purpose. Note that if you’re on a Garmin Coach plan, a manually pushed workout won’t replace the scheduled one, so you’ll want to skip or delete the coach’s session for that day or your training readiness numbers get noisy.
Coros. Workout builder in the app, same structure, though Coros handles nested repeats more gracefully and lets you set an interval alert without building a full workout (useful for one-offs).
Apple Watch. Custom workouts in the Workout app support work/recovery intervals with repeats as of watchOS 9. Set the work interval to time-based, not distance.
Runna. As of late 2025 Runna has a return-to-running option and will generate walk-run sessions, but it progresses on a fixed template and won’t hold volume flat while extending bouts. What works: set your plan start date four to six weeks out, do your own walk-run build in the gap, and let the plan pick you up when you’re running 25 to 30 continuous minutes. Runna’s easy runs are usually the right length by then.
ChatGPT-built plans. Ask for the two-dial structure explicitly. A prompt that works: “Build me an 8-week walk-run return from [injury], 3 sessions/week, alternating weeks where total running time increases and weeks where run interval length increases while total time stays flat. Output as a table with run interval, walk interval, reps, and total running minutes per session.” Without that instruction you’ll get something that goes 1min → 2min → 5min → 10min while also adding volume, which is two big jumps stacked.
Reading Your Own Data To Decide
Three signals, in order of usefulness.
Pain the next morning. This outranks everything. The working rule most sports physios use: pain during running up to 3/10 is acceptable if it settles within 24 hours and isn’t worse session to session. Morning stiffness in a tendon that takes more than 10 minutes to walk off means the last session was too much. Drop back one week in the progression and repeat it.
Cadence stability across the session. Open the session in Garmin Connect or Strava and look at cadence by lap. On a well-dosed walk-run session, your running laps should sit within about 3 spm of each other. A runner whose first three running laps average 168 spm and whose last three average 160 spm is fatiguing inside the session, which means the volume or the bout length is ahead of the tissue. Hold the week.
Strava’s lap view does this fine if you’ve built the workout properly on the watch, since each run and walk step becomes its own lap. If you’re doing it freestyle with no laps, you get one averaged number and lose the information entirely, which is another reason to build the workout.
Heart rate drift within running bouts. Less reliable early on (short bouts don’t give HR time to stabilise, and the walk breaks reset it), but from week 5 onward, with 3-minute bouts or longer, it starts to mean something. If your first 3-minute bout tops out at 142bpm and your sixth tops out at 158bpm at the same pace, you’re accumulating fatigue faster than the walk breaks can clear. Garmin’s Performance Condition metric captures something similar and will typically read -3 or worse by the end of such a session.
What to ignore: training load, VO2max estimates, Training Status, and anything Garmin says about your “productive” or “unproductive” trend. These are built on continuous-running assumptions and walk-run sessions confuse them. Your Garmin will likely tell you you’re detraining through the entire protocol. It’s wrong, and arguing with it costs you nothing but it does tempt people into adding volume they shouldn’t.
The Two Mistakes That Undo Everything
First: running the walk breaks out early because you feel fine. You will feel fine. Aerobic fitness returns in two to three weeks and tissue tolerance takes two to three months, so there’s a window of six weeks or more where you feel capable of far more than you can safely absorb. The protocol exists to protect you from your own legs feeling good.
Second: one big weekend session. Someone does 6 × 3 minutes on Tuesday, 6 × 3 on Thursday, and then a friend suggests a parkrun on Saturday, and 25 continuous minutes of running at 5K effort lands on tissue that has never seen more than three minutes in a row. This is the single most common reinjury pattern in return-to-running. Keep the three sessions symmetrical until you’re through the protocol. Race pace comes back last, not first.
If you want a specific rule for when walk-run is genuinely behind you: three consecutive weeks of 30+ continuous minutes, no next-morning symptoms, cadence stable across the session, and one of those weeks including a session with some faster running in it. At that point you’re not a returning runner with a walk-run plan. You’re a runner, and you can hand the plan back to Runna or your coach and let it do what it’s good at.