The load numbers in your training app were built for cyclists: athletes with power meters bolted to their cranks, measuring work directly. Runners inherited the system secondhand, acronyms and all: CTL, ATL, TSB, ACWR. The ideas underneath are sound and worth having. They just need translating into running, and into words. Start with the term itself: training load is your running scored for stress and summed over time, one number for how much you’ve asked of your body lately.
One number per run: session stress
Everything starts by scoring each run’s training stress: one number built from how long you ran and how hard, where an hour at your threshold scores about 100. Short and easy scores low; long or hard scores high; long and hard scores highest.
The honest way to get “how hard” for a runner is pace corrected for the terrain. debrief scores stress from normalized grade-adjusted pace against your threshold, so a hilly hour counts as the effort it was, not the splits the hills allowed. When GPS can’t be trusted, on a treadmill or a trail under canopy, it falls back to a heart-rate-based score, and when even that thins out, to the oldest currency in the sport, Banister’s TRIMP: minutes weighted by how high your heart rate ran. Best available data wins; the run always gets scored.
Long-term base (CTL)
Average those daily scores with a long memory, weighted toward recent weeks and stretching back about six, and you get your long-term base, the number the acronym crowd calls CTL. It answers one question: how much training is your body currently adapted to absorb?
The base moves slowly, and that slowness is the point. One big week barely nudges it; a consistent quarter raises it for real. It is the closest thing the system has to a number for “fitness,” and the only way it goes up is the way fitness goes up: patiently.
Short-term load (ATL)
The same average with a one-week memory is your short-term load (ATL): what you’ve asked of yourself lately. It spikes with a hard block and falls within days of backing off. On its own it says little; its job is the comparison that comes next.
Training balance (TSB)
Subtract recent load from long-term base and you get training balance (TSB), the one number worth reading daily. A debrief might say: your training balance is −13.9 (TSB), meaning you’re carrying more short-term load than your base currently absorbs.
Negative is not bad. Negative is what productive training weeks look like; the stress that builds the base lives there. Positive means freshness: where a taper should land you, and where fitness quietly leaks away if you stay too long. What deserves attention is the deep end: a balance far below zero, held for weeks, is how overreaching announces itself before your legs do.
The acute-to-chronic load ratio, and why it’s contested
The acute-to-chronic load ratio (ACWR) divides last week by the month behind it: above 1.0 and you’re ramping, below and you’re absorbing. debrief computes it both standard ways: the rolling 7-day-on-28-day version and the exponentially weighted variant proposed by Williams and colleagues in 2017. The literature uses both.
You may have seen this ratio sold with a “sweet spot” and an injury-risk red zone. Read the fine print: Impellizzeri and colleagues took the method apart in 2020, ratio arithmetic and causal claims alike, and the injury-prediction story doesn’t stand as advertised. That doesn’t make the ratio worthless; a month of context on a week of work is a sensible thing to glance at. It makes it context, not a verdict. debrief labels it a contested metric wherever it appears and never hangs a conclusion on it alone. Any tool that gives you a red-alert injury forecast from this one number is selling certainty the science doesn’t have.
Your numbers only mean something against your own history
A training balance of −15 is a heavy week for one runner and a cutback for another. Population norms have nothing useful to say about which you are. debrief compares every run against your own recent history (the same kind of session, a similar duration, the last twelve weeks) because a baseline pooled from strangers would be noise. Whether today’s effort was ordinary or remarkable is a question only your history can answer. Load says how much you’ve trained; whether it’s holding up shows in markers like aerobic decoupling.
How debrief reads these for you
You could track four acronyms across three charts. Or the numbers could arrive already read: a debrief opens with what the load picture means, whether building, absorbing, or overreaching, and the figures ride along in parentheses where they earn the space. That’s the working style of the AI running coach behind this site: meaning first, numbers second, silence when nothing needs saying.