Training Load
How much stress your training is putting on your body, the general concept behind specific scores like TSS, CTL, and ATL.
Training load is the general idea that every workout places some amount of stress on your body, and that stress can be estimated, tracked, and managed over time. It's the umbrella concept underneath more specific tools like TSS, CTL, and ATL, which are all particular ways of turning "how hard was this training" into a number you can actually work with.
Where the idea comes from
The underlying concept predates any single scoring formula. Sports scientists have modeled training as a dose that produces a response, an approach often called an impulse-response model, since exercise physiologist Eric Banister proposed one in 1975 while working with collegiate swimmers. The specific formulas came later and multiplied: some use power output, some use heart rate, some use nothing more than how hard a session felt combined with how long it lasted. All of them are trying to answer the same underlying question, just with different inputs.
How it's usually measured
The most common approach in endurance sport combines two things: how long a session lasted, and how intense it was. A power-based score like TSS does this with cycling power data. Running and swimming often use pace-based or heart-rate-based variants for the same purpose. Where none of that instrumentation exists, a simpler method called session-RPE multiplies a workout's duration by a 1-10 RPE rating, and it correlates surprisingly well with the more data-heavy scores. None of these methods is "the" correct one. They're different rulers measuring the same rough thing, and consistency in which ruler you use matters more than which one you pick.
Why one session's number isn't the point
A single workout's load score, on its own, tells you almost nothing useful. A 150-point session could be a huge day for a beginner or a routine day for someone deep into an Ironman build. Load only becomes useful information once it's tracked across days and weeks, which is exactly what CTL and ATL do: CTL as a slow-moving average that behaves like accumulated fitness, ATL as a fast-moving average that behaves like current fatigue. A rapid, sustained jump in load compared to what an athlete is used to is one of the more consistently observed risk factors for injury across sports-science research on training workload, which is part of why the shape of load over time matters more than any individual day's total.
Why this belongs in the plan, not just the log
Tracking load isn't just a record-keeping exercise after the fact. A well-built plan uses it to decide, in advance and in real time, when to add volume and when to back off, including the deliberate drop in load that makes up a taper before a race. That's also the core mechanical difference between software that just logs your training and software that actually adapts your plan around it: for more on what that distinction does and doesn't mean in practice, see what adaptive training can and can't replace.