Training Load Explained: ATL, CTL, TSB and What Simma Does With Them

Simma turns a squad's sessions into numbers, and three of those numbers, fitness, fatigue and form, are the ones coaches most often ask us to explain. This piece says what they are, exactly how we compute them, and where they stop being useful. What to do about them is the coach's call.

Where the model comes from

The three metrics belong to the performance management model developed by Andrew Coggan for cycling power data and since used across endurance sport. It models three things: how much training an athlete has absorbed over a long window, how much over a short one, and the difference between the two. That is all it does. The names vary by platform. Chronic training load, acute training load and training stress balance are the originals; Simma calls them fitness, fatigue and form, because those are the words the numbers are standing in for.

The input is a stress score per session. For rides with a power meter it is the power-based score; for rides without one Simma estimates from speed and gradient and marks the result as approximate; for runs it is pace-based; for swims it is a swim-specific score. Each is relative to that athlete's own threshold, which is why two athletes doing the same session get different scores.

Fitness (CTL)

An exponentially weighted average of daily stress with a 42-day span. It moves slowly. A week of heavy training nudges it; a month of consistent training lifts it; a fortnight off lets it sag.

It is called fitness because sustained load is the best single proxy the model has for adaptation. It is not a measure of speed, skill or race readiness. A high fitness number from a month of unstructured volume and a high fitness number from a month of purposeful sessions look identical here. The model counts stress. It does not grade it.

Fatigue (ATL)

The same calculation with a 7-day span. It moves fast, because it is mostly this week. A race on Saturday and a long ride on Sunday show up on Monday.

Fatigue is the model's early signal. When it rises well above fitness the athlete is absorbing more than they have adapted to. Whether that was a planned overreach block or three unplanned hard days in a row is not something the number can tell. It is the same spike either way.

Form (TSB)

Fitness minus fatigue. Positive means fresher than their recent average, negative means carrying more fatigue than they have adapted to.

Simma labels form on each athlete's page so it can be read at a glance:

• Transition: above +15
• Fresh: +5 to +15
• Optimal: -10 to +5
• Tired: -25 to -10
• Overreaching: below -25

Those boundaries are fixed and the same for everyone. They are a starting point, not a verdict. Two athletes who race well at very different form values are both normal, and the fixed labels cannot know which is which. That is one reason the training-status classification, described below, calibrates its zones to each athlete rather than using these.

What the athlete is never shown

Simma never quotes a form number to an athlete, in a draft or anywhere else, and never opens a message with fitness, fatigue or form. The numbers are a scale the athlete did not choose and does not use. What an athlete gets instead is the reading beside their own recent normal: "your easy runs were six beats higher at the same pace as a fortnight ago". The model runs underneath. The coach and the athlete talk about sessions.

What the model cannot see

This is the part worth reading twice.

Session quality. Same stress score, very different sessions: intervals hit at the prescribed intensity, or an easy day ridden too hard. The model cannot tell them apart.

Distribution. Two athletes with identical fitness can have opposite intensity distributions. The model does not know.

Cross-sport fatigue. Simma computes the three numbers per sport, because the inputs differ per sport. Fatigue does not respect that boundary. A big bike week affects Thursday's run, and the running model does not see the bike. Simma's training status checks total sessions across all sports in the last seven days and flags a high count, but that is a count, not a model.

Everything that is not a session. Sleep, illness, work, travel, a bad night, a good one. These move performance and injury risk at least as much as any load number, and none of them has a data channel.

A form number is a point on a chart. What made it, and what it will become, is in the sessions and in the person.

From a number to a pattern

A single form value tells you where an athlete is today. It does not tell you how they got there or where they are heading. A form of -12 after a week of stress, recovery, stress is one thing. A form of -12 at the end of four consecutive sessions of decline with no recovery between them is another. Same number.

That is why Simma classifies each athlete's recent pattern per sport rather than showing you a chart to read: whether form is oscillating, falling without recovery, climbing back, or flat, and whether it is sitting in the athlete's own deep zone or their normal range. Training Status Explained describes how, and How to Read Your Squad View says what each status looks like on your dashboard.

What should your assistant be able to do here?

The 42-day and 7-day spans and the form labels are the model's conventional values. If you want your assistant to work with something else, form bands set per athlete, a combined load across sports, a projection of form on race day from the plan, email coaching@getsimma.com and tell us. We build tools from those.

Simma computes fitness, fatigue and form for every athlete and every sport from the sessions they record, reads the pattern, and hands you a status and a reason rather than a chart, in the web app or inside the assistant you already use. Create your coaching account.