Training Stress Score.

1 hour at your threshold = 100 TSS. Everything scales around that. A 30-minute easy spin might be 25 TSS. A 3-hour hilly ride at upper-Z2 could be 220. A race-pace half-marathon lands around 150.

The point is to compare a 90-minute ride to a 45-minute run to a 3 km swim on one axis — “how much did my body get worked” — instead of three separate scales.

Andy Coggan’s formula, published in Training and Racing with a Power Meter:

Formula

TSS = (duration_seconds × NP × IF) / (FTP × 3600) × 100

where:

• NP (Normalized Power) — variability-weighted power. Not the arithmetic average. Strava computes this.
• IF (Intensity Factor) — NP divided by FTP. Your 20-minute flat-out is IF ≈ 1.05; your easy ride is IF ≈ 0.65.
• FTP — your threshold power, roughly what you can hold for 1 hour all-out.

Quick check

1 hour at FTP (IF=1): TSS = 1 × 3600 × FTP × 1.0 / (FTP × 3600) × 100 = 100. Exactly as the definition requires.

No power meter?

Fall back to HR-based TRIMP — a simpler formula that uses HR ratio squared. Less accurate (HR drifts with heat, sleep, hydration) but directionally correct.

Joe Friel adapted the concept for running. Same 100-at-threshold anchor, but intensity comes from pace.

Formula

rTSS = (duration_hours × IF²) × 100

where IF = threshold_pace / actual_pace (faster pace = higher IF).

Example

If your threshold pace is 4:00 min/km and you run 5 km at 4:12 pace (21 min total):

• IF = 240 / 252 = 0.952
• Duration = 21/60 = 0.35 hours
• rTSS = 0.35 × 0.952² × 100 ≈ 31.7

Short, intense effort = modest rTSS. A 90-minute easy run at 5:00 pace (IF = 0.80) is rTSS ≈ 96 — nearly 3× the stress of the 5 km even at much lower intensity. Duration matters.

Swim stress uses a cubic of pace ratio because water drag grows cubically with speed. Anchor still stays at 100 per threshold hour.

sTSS = duration_hours × (threshold_100_pace / actual_100_pace)³ × 100

If you don’t have a threshold swim pace set, we can’t compute sTSS. Strava swim entries with distance but no stroke-rate detail still give us enough to estimate pace.

TSS alone is just a number. Its value comes from what you do with it:

• CTL (fitness) is the 42-day exponentially weighted average of daily TSS.
• ATL (fatigue) is the 7-day exponentially weighted average.
• TSB (form) is CTL − ATL.

Once you have TSS, you have the whole model. Without it you just have distance and duration — numbers that can’t distinguish a recovery spin from threshold intervals of the same length.

Meet an age-group triathlete training for a 70.3 in August. FTP is 240 W, run threshold pace is 4:15 min/km, swim threshold 100-pace is 1:35. Here’s a mid-block week in March:

Monday — Rest

No activity. TSS = 0. Real zero matters — it pulls monotony denominator up later in the week.

Tuesday — 45-min run, 8 × 400 m at 5 km pace

Avg pace 4:45, warmup+cooldown at 5:30. Duration 0.75 h, average IF ≈ 0.89. rTSS ≈ 59.

Wednesday — 90-min endurance ride at IF 0.68

NP 163 W, duration 1.5 h. TSS = 1.5 × 0.68² × 100 ≈ 69.

Thursday — 3 km swim, mixed intervals

Average 100-pace 1:42, 58 min total. sTSS ≈ 0.97 × (1.35 / 1.70)³ × 100 ≈ 49.

Friday — Easy 30-min spin

NP 130 W, IF 0.54. 0.5 × 0.54² × 100 ≈ 15.

Saturday — 3-hour ride with 2 × 20 min at FTP

Overall NP 210 W, IF 0.875. 3 × 0.875² × 100 ≈ 230. The long day dominates the week.

Sunday — 14 km long run at 5:00 pace

IF ≈ 0.85, duration 1.17 h. rTSS ≈ 84.

Weekly totals

• Total TSS: 0 + 59 + 69 + 49 + 15 + 230 + 84 = 506
• Mean daily TSS: 72
• Daily stdev: ~78 (big gap between 0 and 230)
• Monotony: 72 / 78 ≈ 0.92 — healthy
• Strain: 506 × 0.92 ≈ 465 — manageable

This week looks busy but the variation is honest — two real quality days, two low days, a rest day. The numbers agree with how the week felt: productive, not crushing. That agreement is the whole point of TSS.

• Using HR-based TSS for a power-based workout and expecting agreement. HR drifts with heat, dehydration, sleep, caffeine. A threshold interval session can score 20–30% lower on HR-TSS than on power-TSS on a bad day. Pick one source of truth per sport — ideally power for cycling, pace for running — and stay consistent.
• Treating rTSS from a capped treadmill as comparable to road rTSS.If your treadmill tops out at 17 km/h and your threshold pace is faster, your IF is artificially clipped and rTSS understates the session. Flag treadmill sessions in Strava and don’t compare them to outdoor rTSS at the same perceived effort.
• Adding swim TSS and cycling TSS as if they’re the same unit.The 100-at-threshold anchor is the same, but the strain on tissue, tendons, and the cardiovascular system isn’t interchangeable. Adding sTSS + TSS gives you a combined load number that matters for CTL and monotony — but don’t expect 50 sTSS to feel the same as 50 cycling TSS.
• Chasing a weekly TSS target regardless of context. “I need to hit 500 TSS this week” is the wrong framing. You need to build CTL at a safe ramp, with appropriate monotony. Some weeks that’s 600 TSS, some weeks it’s 300. Let the plan prescribe the number.
• Ignoring TSS after strength or cross-training. A hard gym session leaves cardiovascular capacity intact but the legs heavy. CTL won’t reflect that, so you can show up for Tuesday intervals feeling wrecked. TSS-based readiness is a useful signal, not a complete one — listen to your legs too.
• Re-testing FTP only when it goes up. FTP falls in the off-season and during injury blocks. Refusing to lower it keeps TSS numbers looking impressive while your training quality silently drops. Re-test every 6–8 weeks in build phases; re-estimate after any 3+ week break.

CTL, ATL, and TSB

TSS is the daily brick. CTL, ATL, and TSB are what you build with it. CTL is your 42-day EWMA of daily TSS — fitness. ATL is the 7-day version — fatigue. TSB is the gap — form. If TSS is the raw input, these are the curves you actually act on. See the CTL, ATL, TSB guide for bands and ramp rates.

Monotony and strain

TSS totals can be honest or sneaky. 500 TSS over a week with two big days and two rest days is very different from 500 TSS spread 71-71-71-71-71-71-74 — same total, but the second version has monotony above 3.0 and a real illness risk. Foster monotony and strain catch this, and they’re built from the same daily TSS numbers. See training monotony for the math.

Training readiness

Our 0–100 daily score combines TSB, monotony, and recovery since last quality. TSS doesn’t appear directly — but everything downstream of it does. When readiness drops below 50, the answer is almost always visible in the previous seven days of TSS. See the readiness guide.