Skip to content
BestTurboTrainers

Guides

How Do Training Apps Track Progress? (Coach Explains)

Matt Hargreaves Level 2 British Cycling Coach · BSc Sport & Exercise Science Updated 19 April 2026

The short answer

  • Training apps track progress by recording every watt and heart rate you produce, then turning that raw data into FTP, power curves and load metrics.
  • FTP (Functional Threshold Power) is the anchor number: most plans and zones are calculated as a percentage of it.
  • Power curves show your best output across durations, from a 5-second sprint to a 20-minute effort, so you can see your strengths and weaknesses.
  • CTL, ATL and TSB (fitness, fatigue and freshness) model whether you are building, recovering or overreaching.
  • The numbers are only as good as your trainer's accuracy, so cross-check power against a known source before you trust the trend.

Training apps track progress by capturing every watt, heartbeat and second of your rides. They then convert that raw data into a handful of metrics: your FTP, your power curve, and your fitness and freshness scores (CTL, ATL and TSB). The app learns your best efforts across different durations. It estimates the hardest power you can sustain. It models how training load is building or draining your form over weeks. Cleaner power data means more honest numbers. Your trainer’s accuracy matters as much as the software.

After years of coaching riders through their first indoor seasons, I have watched the same few metrics either motivate people or quietly mislead them. Below I walk through what each one actually means, how the apps build it, and where the numbers quietly lie to you.

What data do training apps actually record?

Every indoor ride sends a stream of numbers to the app, usually several times a second:

  1. Power (watts) from your smart trainer or power meter. This is the headline figure and the most reliable progress signal.
  2. Cadence (rpm) from the trainer or a separate sensor.
  3. Heart rate from a chest strap or optical band.
  4. Speed and distance, which on a turbo are simulated from power and a virtual course rather than measured.
  5. Time, which underpins every duration-based calculation.

The app derives everything else from that raw stream. None of the clever metrics exist in the data itself: FTP, power curves and fitness scores are all calculated afterwards. That matters because a bad input ruins every output downstream.

FTP: the number everything hangs on

FTP (Functional Threshold Power) is roughly the highest average power you can hold for about an hour. It is the anchor that training apps use to set your zones and prescribe workouts. When a session says “ride at 88 percent” it means 88 percent of your FTP.

Apps work out your FTP in a few ways:

  • A dedicated test, usually a 20-minute all-out effort with your average multiplied by about 0.95, or the newer ramp test where resistance climbs until you fail.
  • A modelled estimate (TrainerRoad’s AI FTP Detection, Zwift’s ramp result, or similar) that infers your threshold from recent hard rides without a formal test.
  • A manual entry, if you already know it from outdoor data.

Whichever route you use, the rule I drill into every rider is to keep the conditions consistent so an FTP change reflects real fitness rather than measurement drift. If a trainer reads higher than the power meter you ride outdoors, a sudden “FTP gain” can simply be the trainer warming up or losing calibration, not new form.

If you want the full picture of how watts are measured and where errors creep in, I cover it in detail in my guide on understanding power readings on your turbo trainer.

Power curves: your fitness fingerprint

An FTP is one number. A power curve is the far richer view: it plots your best-ever power for every duration, from a 5-second sprint through 1-minute, 5-minute and 20-minute efforts, all the way out to multi-hour rides. The shape of that curve is effectively your physiological fingerprint.

A short spiky curve that is huge at 5 seconds but sags by 5 minutes says “sprinter”. A flat high curve that holds power for 20-plus minutes says “diesel engine, good for time trials and long climbs”. Most of us sit somewhere in between. The curve tells you exactly where your training is paying off and where the gaps are.

Apps build the curve by scanning all your historical rides and keeping the single best effort at each duration. Do a hard 3-minute effort today that beats your previous best, and that point on the curve nudges up. It is a record of peak ability, not average effort.

I have written a dedicated breakdown of how to read and train these in what are power curves in training apps. For now, the key idea is that the power curve answers a different question from FTP: not “how strong am I overall” but “how strong am I at this specific duration”.

MetricWhat it answersHow the app builds itBest use
FTPYour sustainable threshold power20-min test, ramp test, or modelled estimateSetting zones and scaling workouts
Power curveYour best power at every durationKeeps the peak effort per duration across all ridesSpotting strengths, weaknesses and PBs
CTL (Fitness)How much load you have banked42-day weighted average of daily TSSTracking long-term build
ATL (Fatigue)How tired you are right now7-day weighted average of daily TSSSpotting overreaching
TSB (Freshness)How ready you are to performCTL minus ATLTiming rest and event days

Fitness, fatigue and freshness: CTL, ATL and TSB

CTL, ATL and TSB model fitness, fatigue and freshness from your training load. This trio confuses most riders. Here is the plain-English version.

TSS: the building block

Every ride earns a Training Stress Score (TSS). A one-hour ride exactly at FTP scores 100 by definition. Ride harder or longer and the number climbs; spin easy and it stays low. TSS combines intensity and duration into one figure, and it is the raw material for everything below.

CTL is your fitness

Chronic Training Load (CTL) is a rolling 42-day weighted average of your daily TSS. Apps usually label it “Fitness”. Because it averages six weeks of work, it moves slowly: one massive session barely shifts it, but a consistent block pushes it up steadily. A rising CTL over a season is the cleanest evidence that your training is actually accumulating.

ATL is your fatigue

Acute Training Load (ATL), labelled “Fatigue”, is the same idea over a much shorter 7-day window. It spikes after a hard weekend and drops quickly when you rest. ATL is the metric that reacts to what you did yesterday and the day before.

TSB is your freshness

Training Stress Balance (TSB) is simply CTL minus ATL, and apps call it “Form” or “Freshness”. When TSB is negative you are carrying fatigue (normal during a build). When it swings positive after a taper, you are fresh and ready to perform. The art of peaking is letting fitness stay high while fatigue falls, so freshness rises on the day that matters.

Building these numbers properly depends on a sensible week of training rather than random hard rides. If your weekly structure is haphazard, the load metrics will be noisy and misleading. My guide on creating a weekly turbo training plan shows how to spread load so the graphs actually tell you something.

How the trainer and workout builder feed the data

The quality of every metric above traces straight back to the watts your trainer reports. A direct-drive smart trainer with tight accuracy gives clean data; a basic wheel-on with estimated power gives a noisier signal that wobbles your FTP and curves. If you are deciding which to buy with progress tracking in mind, my direct drive vs wheel-on comparison lays out the trade-offs.

Structured workouts are where the data gets created in the first place. When you build or follow an interval session, the app prescribes targets as a percentage of FTP, then records how closely you hit them, feeding TSS and your power curve. If you want to understand the tools that shape those sessions, see understanding workout builder features, which covers ERG mode, target zones and how custom intervals translate into trackable load.

Do the numbers actually matter for a beginner?

Honestly, less than the apps would have you believe. Consistency beats analytics in your first season. Turn up three or four times a week and follow a sensible plan and your FTP will climb whether or not you understand TSB. The fitness and freshness graphs become genuinely useful later when you are juggling a build towards an event and need to time your rest.

So my coach’s advice: trust FTP and the power curve early on because they directly reflect what you can do. Treat CTL, ATL and TSB as a longer-term steering wheel once you have a few months of consistent data behind you. Above all make sure the watts going in are accurate, because no clever metric can rescue bad data.

Verdict

Training apps track progress by recording your power, heart rate and time, then distilling it into FTP (your threshold), power curves (your strengths by duration) and the CTL/ATL/TSB triad (fitness, fatigue and freshness). Used well, they turn a winter of garage sessions into a clear, honest trend line. Just remember the golden rule I have learned the hard way: the graphs only ever reflect the data you feed them, so calibrate the trainer, test under consistent conditions, and let the numbers earn your trust over weeks, not days.

Frequently asked questions

How do training apps calculate my fitness?
They use Training Stress Score (TSS) from each ride to build a long-term load average called CTL (Chronic Training Load), which most apps label as 'fitness'. The higher and more consistent your training load over six weeks, the higher your CTL climbs. It is a rolling 42-day weighted average, so a single big session barely moves it, but a steady block does.
What is the difference between FTP and a power curve?
FTP is a single number, roughly the highest power you can hold for about an hour. A power curve is a chart of your best power for every duration, from a few seconds to several hours. FTP tells you your threshold; the power curve tells you whether you are a sprinter, a puncheur or a steady diesel.
Do I need a smart trainer to track progress?
No, but it helps. A power meter or smart trainer gives you accurate watts, which is the cleanest progress signal. You can track progress with heart rate and perceived effort alone, but the numbers drift with heat, fatigue and caffeine, so the trend is noisier.
How often should my FTP go up?
For a beginner, every four to six weeks of consistent training is realistic in the first season. After that, gains slow to a few watts per block. If your FTP jumps 20 percent overnight, suspect a trainer calibration change rather than a fitness miracle.
Why does my fitness score drop when I take a rest week?
Because CTL is a rolling average of recent training load. When you train less, the average falls. That is normal and even desirable before an event: your fitness dips slightly while your freshness (TSB) rises, which is what lets you ride strong on the day.