Training science,
without the textbook.
PaceLab uses evidence to set the direction and labels the parts that are conservative coaching heuristics rather than proven universal rules.
Click a concept. See what it changes in your plan.
Each topic gives you the physiology, the practical training rule and the evidence strength. The animation is deliberately simple: it is there to make the idea memorable, not to pretend biology is that tidy.
STRONG DIRECTIONWhy so much easy running?
Accumulate aerobic work with less fatigue so you can train consistently.
+
Why so much easy running?
Accumulate aerobic work with less fatigue so you can train consistently.
What is changing?
Low-intensity endurance work supports central and peripheral aerobic adaptations, including oxygen delivery and muscular oxidative capacity. The exact response varies by athlete and training history.
What PaceLab does
Most plans aim for roughly 75–85% of weekly running time at easy intensity. Long runs are predominantly easy, while quality work remains a minority of total training.
How it should feel
RPE 2–3/10. Comfortable conversation. If conditions make your usual pace feel harder, slow down rather than chasing the number.
Research supports low-intensity-dominant endurance training, but not a universal law that exactly 80/20 is best for every recreational runner.
Oliveira et al., 2024 →
WELL SUPPORTEDWhy RPE first?
Because the same physiological effort does not always produce the same pace.
+
Why RPE first?
Because the same physiological effort does not always produce the same pace.
What is RPE?
Rating of Perceived Exertion is a simple 1–10 description of how hard the session feels. It captures internal load without needing a laboratory or wearable.
Why pace can mislead
Heat, hills, sleep, accumulated fatigue and terrain can move pace while the intended effort stays the same. Heart rate can also drift or vary between individuals.
PaceLab zones
2–3: easy. 4–5: steady. 6–7: controlled hard / threshold-oriented. 8–9: hard intervals used sparingly.
Session-RPE is widely used as a practical measure of training load; the talk test is also useful for controlling aerobic intensity.
Haddad et al., 2017 →
RACE-RELEVANTWhy controlled tempo work?
Build the ability to hold a faster aerobic pace without making every workout maximal.
+
Why controlled tempo work?
Build the ability to hold a faster aerobic pace without making every workout maximal.
The idea
As intensity rises, lactate production and clearance eventually stop remaining in balance. Training around this region can improve sustainable endurance performance.
What PaceLab does
Most beginner plans use one quality session per week. During build and specific phases, that may include controlled tempo, cruise intervals, hills or race-specific blocks.
How it should feel
Usually around RPE 6–7/10: clearly working, but controlled enough that the final repetition does not become a sprint.
For beginner half-marathon and marathon users, PaceLab prioritises aerobic volume, long-run development and sustainable quality work. Very hard intervals are optional rather than weekly by default.
OBSERVATIONAL + PRACTICALWhy the long run?
Build endurance gradually enough that race distance stops feeling abstract.
+
Why the long run?
Build endurance gradually enough that race distance stops feeling abstract.
What it contributes
The long run extends the amount of continuous aerobic work you can tolerate and rehearses fatigue management, pacing and — later — race fueling.
What PaceLab does
Half-marathon long runs can progress toward roughly 18–20 km. Marathon plans can progress toward roughly 26–30 km when the starting base and available session time support it.
What it does not do
It does not force everyone to the same peak distance. Cutback weeks shorten the run and the plan caps long-run progression when weekly volume or available time do not support more.
In recreational runners, higher weekly volume and longer endurance runs are associated with faster marathon performance, although observational data cannot prove one exact optimal long-run length.
Fokkema et al., 2020 →
CONSERVATIVE HEURISTICWhy not simply add 10% every week?
The famous 10% rule is not a proven universal injury threshold.
+
Why not simply add 10% every week?
The famous 10% rule is not a proven universal injury threshold.
What the research does not give us
Reviews do not establish one exact week-to-week percentage that guarantees safety. Sudden large changes may matter, but the biology is more complicated than one number.
PaceLab's conservative rule
Normal build weeks rise by roughly 5–8% from the previous high week, with a cutback about every fourth week. The plan never forces a predetermined peak if your starting point and timeframe do not support it.
Why label this honestly?
This is a cautious product rule informed by the evidence gap, not something PaceLab presents as a scientifically proven law.
Systematic reviews find limited and inconsistent evidence for exact load-change thresholds and running injury.
Damsted et al. review →
META-ANALYSISWhy taper?
Reduce accumulated fatigue without switching the engine off before race day.
+
Why taper?
Reduce accumulated fatigue without switching the engine off before race day.
What changes
Total volume falls meaningfully while some training frequency and brief intensity exposure are retained.
What PaceLab does
Most half-marathon plans use a two-week taper. Marathon plans use two to three weeks depending on plan duration, with volume progressively reduced.
Why not just stop?
The aim is to shed fatigue while maintaining readiness and familiarity with running rhythm.
A meta-analysis found performance benefits with approximately 41–60% volume reduction over up to three weeks while retaining intensity and frequency.
Wang et al., 2023 →Evidence-informed does not mean clinically validated.
PaceLab is an educational training tool for generally healthy adults. The plan-generation rules have not been prospectively validated as a medical or injury-prevention algorithm. Current pain that changes running mechanics stops a normal plan from being generated.