Form (TSB)

Form (TSB), or Training Stress Balance, is the difference between Fitness (CTL) and Fatigue (ATL). It shows where fast-moving recent load sits relative to the slower long-term load model.

“Form” is a readable UI label. TSB is derived from two modeled load curves; it is not a direct readiness, recovery, health, or performance measurement.

Example values are for interface preview only.

Where the value comes from

TSB has no separate activity input. Trainingload.ai first calculates same-day CTL with a 42-day time constant and same-day ATL with a seven-day time constant, using the same daily-load series. Any activity, threshold, or load-source change that affects either curve can also change TSB.

TSB uses the same modeled load scale as CTL and ATL. It is not a percentage or physiological unit.

Formula

Trainingload.ai follows the same-day convention used by Intervals.icu:

TSB_today = CTL_today − ATL_today

For example, if today’s CTL is 50 and ATL is 65, TSB is 50 − 65 = −15.

Platforms do not all use the same date convention. Intervals.icu uses same-day CTL minus ATL, while TrainingPeaks publicly describes today’s Form using yesterday’s Fitness minus yesterday’s Fatigue. Before comparing platforms, verify the date convention, time constants, and load source.

Reading Form

• Negative TSB: ATL is above CTL; recent modeled load is above the slower long-term level.
• Near-zero TSB: the two load curves are close; this does not prove physiological balance.
• Positive TSB: ATL is below CTL; recent modeled load is below the slower long-term level.

When load increases, ATL usually rises faster than CTL and TSB falls. During rest or tapering, ATL usually decays faster and TSB rises. The sign describes a load relationship—not how the athlete must feel.

Use and limits

TSB can help you check whether a build, recovery week, or taper changed the relationship between recent and long-term load as planned. It is most useful when compared with the athlete’s own similar historical periods.

TSB alone cannot establish race readiness, label a negative value as dangerous, label a positive value as optimal, prescribe rest days, or explain pain and illness.

Avoid chasing universal “peak,” “optimal,” or “high-risk” bands. Such bands depend on the load method, sport, training history, individual response, and data quality. Use Fitness (CTL), Fatigue (ATL), source activities, performance trends, and subjective state together.

References

• Intervals.icu developer explanation of the same-day Form convention
• RUNALYZE glossary: Training Stress Balance (TSB)
• TrainingPeaks: Performance Management Chart
• GoldenCheetah source: PMCData.cpp
• Calvert et al. (1976): A Systems Model of the Effects of Training on Physical Performance
• Clarke and Skiba (2013): Mathematical modeling of athletic training and performance

Related pages

• Fitness (CTL)
• Fatigue (ATL)
• Using PMC to adjust marathon training
• Training analysis workflow