Category: Triathlon

The question comes back every time the season changes: how to split time across training zones to get the most out of 10 hours a week. The polarised model has the appeal of simplicity, easy or very hard, nothing in between. The pyramidal model sounds more reasonable, more at the bottom, less in the middle, a sharp peak at the top. In 2026, a fresh research summary from Arturo Casado, PhD, puts the matter to rest: neither model wins everywhere. That is the most important conclusion of the season.

What actually separates the polarised model from the pyramidal one

Both models describe what percentage of training time is spent in three intensity zones:

• Zone 1 (Z1): below the first lactate threshold (VT1/LT1). Conversational pace, easy effort.
• Zone 2 (Z2): between the first and second thresholds (LT1 and LT2). A pace that cannot be held for long, but is not all-out either.
• Zone 3 (Z3): above the second threshold. Lactate tolerance work, VO2max repeats.

Polarised (POL): lots of Z1, almost nothing in Z2, a decent amount of Z3. Pyramidal (PYR): lots of Z1, moderate Z2, little Z3.

Why these are not the “zones” shown on a Garmin, an important note

Anyone training with a Garmin, Wahoo or TrainingPeaks likely sees five zones based on the model developed by Dr Andrew Coggan, the researcher behind the FTP concept and co-author of Training and Racing with a Power Meter from the early 2000s. That model splits intensity by percentages of FTP and maximum heart rate, treating Zone 4 as the lactate threshold.

The three-zone model used in scientific research (Seiler, Casado, Filipas) is built on two physiological ventilatory thresholds, VT1 and VT2. It is a different cut of the same physiology:

When research says to avoid Z2 in the polarised model, it means zones 3 and 4 on a watch: half-marathon pace, threshold intervals, “sweet spot” work. It does not mean avoiding easy aerobic base, which sits inside Z1 of the research model and is exactly what belongs there. Seiler himself uses the three-zone model in research for simplicity, but works with athletes using five zones, because they give more precision when planning sessions.

Figures from a study by Sellés-Pérez et al. (Journal of Sports Science and Medicine, 2019) on age-group athletes racing the half-iron distance, IRONMAN 70.3 (1.9 km swim, 90 km bike, 21.1 km run), commonly called a “half”, make this concrete:

The Z2 gap is stark: nearly a fifth of training time in the pyramidal group, barely a mention in the polarised one.

What the research shows at the middle distance

The same study found an interesting relationship: time spent in Zone 2 correlated with a better race result among age-groupers at the half-iron distance. In other words, in the 70.3 group, the pyramidal distribution beat the polarised one on race outcome.

That is counterintuitive for anyone who listens to podcasts about the “Norwegian model” and the Iden and Blummenfelt brothers. But the logic is simple: a 70.3 puts an athlete at near-threshold intensity (Z2) for four to five hours. Training that includes plenty of Z2 prepares an athlete better for exactly that.

Casado, on episode 695 of the That Triathlon Show podcast, discusses a study by Filipas et al. (Scandinavian Journal of Medicine & Science in Sports, 2022), in which a 16-week periodised plan, the first 8 weeks pyramidal and the next 8 weeks polarised, produced better running results than either model held for the full cycle. The improvement was mainly in 5k time and speed at lactate threshold; VO2max changes did not differ significantly between groups.

What this means in practice, periodisation instead of doctrine

The research synthesis available through 2026 points to a phased approach:

General phase (base period, November to January). Pyramidal distribution. Plenty of Z1 (around 75 to 80%), a meaningful chunk of Z2 (15 to 20%), little Z3 (3 to 5%). This builds aerobic capacity, movement economy and threshold.

Specific phase (February to April). Shift toward polarised. Keep Z1, reduce Z2 to roughly 5 to 8%, add deliberate Z3 blocks, Norwegian threshold sessions, VO2max repeats. This is when 30/30 intervals start to make sense.

Race phase (May to September, depending on the calendar). Mixed, maintaining all systems while the specific demands of the race take over. For a half, return to more Z2 in the week before the start. For a full-distance event, hold polarisation with long Z1 sessions and short Z3 sparks.

Recovery phase / off-season. Everything in Z1 with one sharp stimulus a week. Not the time for a plan.

What about the 80/20 model

The popular 80/20 model (80% easy, 20% hard) comes from Stephen Seiler’s research into intensity distribution among elites, and is a simplified version of the polarised model. It works well for runners training at very high volume, 15-plus hours a week. For an age-group athlete training 8 to 10 hours a week, the problem is different: the 80% “easy” training often is not easy at all. It ends up in a grey zone, too hard for recovery, too weak for progress.

For an age-grouper, what matters more than choosing between the polarised and pyramidal model is this:

• Keeping Z1 genuinely in Z1. The test: it should be possible to hold a conversation in full sentences.
• Controlling Z3. A real VO2max repeat is one you would cut off your own arm to end.
• Not dragging every session up to a perceived effort of 7 out of 10. That is the road to overtraining without progress.

What to use when building toward a 70.3

A concrete outline to work from:

• November to January: pyramidal. Plenty of long Z1 (bike 2 to 3 hours, run 60 to 90 minutes), one Z2 session a week (for example 4×8′ at threshold), one short Z3 session (for example 6×3′ VO2max).
• February to April: shift toward polarised. Two Z3 sessions a week (Norwegian double threshold, classic 4×4′ VO2max), Z2 limited to brick sessions and specific tests. Everything else in Z1.
• May to September: race specificity. Two weeks before a 70.3, bring back Z2 (race-pace repeats: on the bike 2×30′ at 85% FTP, on the run 3×3 km at half-marathon pace).
• October to November: everything in Z1 plus one sharp stimulus a week.

There is no sacred percentage. There is a phase, a race, and an athlete’s own recovery. The 2026 scientific consensus: be periodised, not dogmatic.

Verdict

Polarised versus pyramidal is a false argument if the choice has to be made once and for all. The best adaptations come from a plan that shifts the intensity distribution across the season: pyramidal in the base period, polarised in the specific period, mixed at the race. Casado and the 2026 research say the same thing good coaches have said for years, just with numbers behind it now.

Anyone who has asked a coach how much they should train has heard some version of “it depends.” A weekly plan for a half distance should look different from one for an Ironman, that much is known. What has been missing is a concrete benchmark for how much the average age-grouper who actually finishes the race trains in a week. Until now. In May 2026, TrainingPeaks published a rundown on its blog of an independent scientific study in the peer-reviewed journal Frontiers in Sports and Active Living, giving the first real picture of that answer.

What was studied, and how

This was not a survey. It was objective session data exported from TrainingPeaks over a six-month period. Every participant had duration, distance, heart rate, Training Stress Score (TSS) and power zones logged. In total: 2,177 weeks of training from 95 age-groupers across 25 countries, 34,731 individual sessions.

The sample is small in athlete count (95) but enormous in logged sessions. The data is objective, not raw Strava pulls where one athlete’s GPS catches a signal and another’s does not. The study was carried out by researchers at Deakin University in Australia and published in a peer-reviewed journal.

Athletes were split into two groups:

• Short-course: sprint, Olympic and middle distance, up to and including 70.3.
• Long-course: full Ironman (140.6).

Numbers worth pinning above your desk

Weekly averages, taken across the full season:

The claim that finishing an Ironman requires 15 hours a week of training still circulates in age-group circles, and these numbers are a reality check. The average is 10h15 for the full distance. Not 15, not 20. Ten, on average.

For a half-Ironman, the short-course average is 8h27 a week. Also less than popular plans suggest.

The second finding: training phase outranks age or sex

The most interesting observation in the study: age and sex were weaker predictors of training load than the phase of the season. In other words, whether an athlete is 30 or 45, a woman or a man, the strongest influence on their weekly load is which phase of the season they are in.

In the specific phase, the weeks closest to race day, TSS load runs about 21% higher than the general phase. In the taper/race/post-race phase it drops about 14% below the general phase. In the off-season it hits its low point: the study puts it around 340 min/week, roughly 38% below the general phase.

That sounds obvious, but it has real consequences:

Generic “12-week to Ironman” plans deserve suspicion. If a plan does not vary load between base, specific and taper phases, discard it.

Comparing yourself with others on Strava only makes sense within the same phase. If a training partner logged 620 to 650 TSS in a week, check whether they are in the specific phase while you are just entering base.

“Only 8 hours a week, so I won’t finish?” is the wrong question. Eight hours in the specific phase with well-distributed TSS can be more effective than 12 hours of chaotic hard work.

What this means for the everyday age-grouper

For an amateur working a 40+ hour week, this is good news:

1. Your goals are more realistic than you think. Giving 8 to 10 hours a week puts an athlete in the international average for age-groupers finishing a half-Ironman. Fifteen hours a week is not required to finish a 70.3.

2. TSS counts, not time. The 452 TSS/week short-course average works out to roughly 65 TSS a day. That could be an easy 90-minute ride (about 50 to 60 TSS) or a 60-minute threshold run (80 TSS). What matters is that the weekly total adds up.

3. Periodisation beats volume. The same training time, spread through a 3:1 progression (three weeks building, one week recovering) with distinct phases, produces dramatically better results than training 8 flat hours a week all year with no variation.

4. Compare against the average, not the elite. A pro trains 25 to 30 hours a week at 1,200+ TSS; that is not the right benchmark. The benchmark is 615 minutes and 574 TSS for an Ironman.

What not to do with this data

Two interpretive traps worth watching for:

Trap 1: “If the average is 10 hours, 10 hours is enough for me.” That figure is averaged across the whole season and the whole sample. What counts as “enough” depends on the finishing time being chased. Finishing an Ironman in 14 hours is a different task from qualifying for Kona.

Trap 2: “If TSS is 574, every week should be 574.” No. That is a seasonal average across different phases: base weeks run around 400, the specific phase 700 to 750, taper below 500. Week-to-week fluctuation is correct and expected.

The verdict

The first study of this scale shows that finishing an Ironman as an age-grouper does not require living like a monk. It requires consistency, periodisation and a defined TSS load, distributed sensibly through the season. The average is 615 minutes a week. Not 1,000.

You’re on the start line. Goggles on for the hundredth time, but the strap never sits the way it does at the pool. The bike: you built it out of the travel box and did a couple of loops round the block, but you’ve never done that in a wet wetsuit, with the PA blaring and a few hundred other athletes around you. A bottle mounted between the aerobars, except how much did you actually drink from that setup in training?

This is exactly the problem race-simulation training is meant to solve.

What a Simulation Session Is, and Why It Matters

A simulation session is a workout where you deliberately recreate the conditions and sequence of race day: the swim exit, T1, the bike, T2, the run. It’s not about covering the full race distance, that’s what your A-race is for. It’s about reprogramming your nervous system so that nothing on race day feels new.

There are three benefits:

• You test tactics and pacing before there are real consequences
• You test gear and nutrition under real conditions, not just in your head
• You build confidence: you’ve done this before, and you know you can handle it

When to Schedule a Simulation Session

The ideal is twice per training block: one session about 8 weeks before the race, a second about 4 weeks before. The first is reconnaissance, you find out what isn’t working and what needs fixing. The second confirms you’re ready.

What not to do: don’t schedule a simulation the week before the race. You need 5 to 7 days to recover from a session like this, and taper week is not the time for new experiments.

Structuring a Simulation Session (Olympic Distance)

Swim: 500 to 750 m

Instead of just swimming the distance, recreate the effort pattern of the race:

• 100 m at start pace (the opening metres of a mass start can get chaotic)
• 200 m at settling pace (moderate effort, finding your rhythm)
• 100 m hard (simulating a sprint to a buoy or the finish)
• Short rest, repeat

One key drill: stripping off your wetsuit. Do it under time pressure, twice, with people watching if you can arrange it.

Bike: about 30 to 40 km

A format TrainingPeaks recommends for Olympic distance:

• 4 x 2 km at race pace-plus (slightly above your target FTP%)
• 8 km at base pace between intervals
• No breaks between blocks, ride it continuously, the way you would on race day

Non-negotiable: eat and drink exactly what you plan to use on race day. Not a simulation of your nutrition, the real thing: the same gel, at the same point on the course, that you’ll use on race day.

Run: 5 to 10 km

Run it straight off the bike. That’s the whole point of the exercise.

• 5 x 1600 m at race pace, with 400 m jog recoveries
• Run the first 800 m purely by feel, without looking at your watch. Your legs lie about pace after the bike.

What to watch: how far off your target pace is that first kilometre. If you’re starting too fast, that’s something to fix before race day.

What to Test During the Simulation

Some things seem obvious and only fail on race day:

Goggles. Do new ones slip on a run-in start? If you haven’t been in the water with them 10-plus times, you’re carrying a risk.

Cleats. Jumping into your shoes with cleats already clipped to the pedals, or putting shoes on standing beside the bike. Do it at least five times before the simulation.

Race number. How are you fixing it? Pins, elastic, a belt? Does it stay straight through 10 km of running?

Running cap. Does it stay on during the run.

The Mental Side: The Invisible Piece

A simulation session is also a mental exercise. Start the clock, set yourself a race target, and go. Watch what happens in your head once the pace gets hard.

Feel the pressure? Good, that’s where the learning happens. Every hard 200 metres of a simulation session is a rehearsal for the real race.

Don’t avoid discomfort in these sessions. You need to learn that you can handle it, and that means being in a hard place at least once before race day.

Two Sessions per Block

Eight weeks out: reconnaissance, diagnosing weak points. Four weeks out: confirmation that you’re ready, the sharp point of the season.

Each session costs 2.5 to 4 hours. In return: a calm start, gear you’ve already tested, and a head that knows it’s done this before.

The season is in full swing, and triathlon groups are debating: one athlete trains 15 hours a week and posts it on Strava, another swears four hours is enough. Both are right, and neither is.

The right training volume depends on your goal, your level and how much recovery you can actually get. But the science points to a few concrete thresholds worth knowing.

The Minimum That Actually Moves the Needle

Recommendations from Tri Coaching Finland and TrainingPeaks analysis point to a few tiers:

4 to 6 hours a week: enough to feel fit and improve your base endurance. A realistic ceiling for someone with a job and a family. Split: one swim, two bike sessions (one longer), two runs. Progress is slow but consistent.

6 to 9 hours a week: this is where serious sprint and Olympic-distance racing starts. Roughly 75% of volume in zone 2 (aerobic), 25% above threshold. At this level you can compete for decent times at local and regional races.

10 to 15 hours a week: full-distance ambitions, a half-distance IRONMAN as the season goal. Requires planned recovery and sensible periodisation, or you end up injured in August.

Zone 2 Is the Foundation, Not the Boring Option

If you train 8 hours a week and spend 6 of them in zone 3 to 4 (suffering, but not quite hard, not quite easy), you’re doing it wrong. Research consistently confirms that most of your volume should sit in zone 2, conversational pace, where you can talk in full sentences.

It’s mentally uncomfortable because zone 2 looks too easy. But that’s exactly where mitochondrial density, movement economy and fat-burning capacity get built, and those adaptations move the needle more than another interval session.

A practical rule: for every hour of intense training, put in at least three hours of aerobic work.

Why Consistency Beats Intensity

This is one of the few points where training science agrees across the board: regular, smaller-volume sessions outperform occasional, oversized training blocks.

Three 45-minute sessions spread evenly through the week beat one 3-hour session once every 10 days, with a much lower risk of injury, illness and burnout. Your muscles, tendons and hormonal system need regular training signals, not sudden overload.

In practice: if you have less time this week, don’t try to make it up the next. Keep the pattern and shorten the sessions instead.

When Volume Starts to Hurt

The trap in the middle of race season is stacking high volume onto a body still tired from racing. After every race you need 1 to 2 days of easy training for every 10 km of running distance covered (a rough rule of thumb). After an IRONMAN 70.3, a minimum of a week of lighter training before you start building volume again.

Signs you’re training too much: resting heart rate rising more than 5 to 7 beats per minute above normal, HRV clearly below your baseline, worse sleep quality, chronic leg fatigue that persists after three days. A watch with HRV monitoring (Garmin, COROS, Polar) helps you catch these signals earlier.

What 7 to 8 Hours a Week Looks Like in Practice

A sample week’s skeleton in the middle of an Olympic-distance season:

• Monday: rest or 30 minutes of mobility work
• Tuesday: 45-minute threshold run (1 x 20 minutes in zone 3 to 4)
• Wednesday: 45 to 60-minute swim (technique plus 2 x 400 m at threshold)
• Thursday: 60 to 75-minute bike (zone 2, steady rhythm)
• Friday: rest or a short 30-minute zone 1 run
• Saturday: long bike, 2 to 2.5 hours, zone 2
• Sunday: brick session, 45-minute bike plus 20-minute run off the bike

Total: about 7.5 hours. There’s no room here for heroic sessions, just a plan.

You do a 20-minute FTP test. It comes out at 250 W. All your zones flow into the app: Z2 up to 165 W, Z3 up to 205 W, Z4 up to 250 W. And you train that way for eight weeks. Meanwhile your anaerobic work, 30-second intervals, sprints on a descent, attacks on climbs, lives in a vacuum. FTP tells you nothing about it, because FTP is one number for one point on the curve.

The power curve changes that. Critical Power (CP) plus W’ give you two points instead of one. Suddenly you see the whole physiology of your legs, not just that one 20-minute segment.

What the power curve is

The power curve (the power-duration curve) is a graph showing the maximum power you can hold for a given length of time. Watts on the vertical axis, seconds or minutes on the horizontal.

• 5 seconds: a neuromuscular sprint. Trained age-grouper: 800-1100 W. Pro: 1500-1800 W.
• 1 minute: max VO2max effort. Age-grouper: 350-450 W. Pro: 600-700 W.
• 5 minutes: the classic VO2max test. Age-grouper: 280-340 W. Pro: 450-500 W.
• 20 minutes: the classic FTP test. Age-grouper: 220-280 W.
• 60 minutes: true threshold power (CP).

The longer the duration, the lower the power. The curve decays exponentially. Two numbers describe it: CP and W’.

Critical Power: your true aerobic limit

Critical Power (CP) is the highest power you can hold in physiological steady state. Below CP, your lactate, VO2 and pH stay stable. Above CP, all three climb exponentially and you move into non-steady state: fatigue accumulates and sooner or later you blow up.

CP is not the same as FTP, though the two sit close together. CP typically runs 4 to 5% higher than your true 60-minute maximum power. So if your real one-hour power is 240 W, CP comes out around 250-252 W. FTP usually lands somewhere in between, around 245 W.

The difference sounds marginal. But it is a meaningfully more precise metric, because CP is not “60 minutes that felt hard”. It is a value calculated mathematically from two or three tests, grounded in physiology rather than how you felt.

W’: your anaerobic tank

W’ (read “W prime”) is the work you can do above CP before you are exhausted. Measured in kilojoules (kJ) or joules.

It is a finite store of anaerobic energy. Picture a fuel tank: you do not burn it while idling below CP. But push the throttle above CP and the tank empties. When it is gone, you are at zero.

Typical values:

• Age-group triathlete: W’ = 10-15 kJ
• Road cyclist with a strong anaerobic engine: W’ = 18-22 kJ
• Track sprinter: W’ = 25-30 kJ

With CP and W’, you can calculate exactly how many seconds you will hold a given power above CP. The formula is simple: time = W’ / (P – CP), where P is the power you are holding and CP is your critical power.

Example: CP = 250 W, W’ = 15 kJ. You hold 350 W (an attack on a climb). How long can you sustain it? 15,000 J / (350 – 250) = 150 s = 2.5 minutes. After that, you blow up.

How to test CP and W’

Two protocols are most common:

1. The 3 to 12-minute test (classic):

• Day 1: 20-minute warm-up, then a 3-minute maximal all-out effort. Average power is P3min.
• Day 2 (or after 48 hours of recovery): 20-minute warm-up, then a 12-minute maximal effort. Average power is P12min.
• Calculation: CP = P12min – ((P3min – P12min) x 3) / 9 (an approximation). W’ = (P3min – CP) x 180 seconds.

In practice, apps such as TrainingPeaks, Golden Cheetah and WKO5 calculate this for you once you upload the files.

2. The 3-minute all-out test (Vanhatalo/Burnley, 2007):

• 20-minute warm-up.
• 3 minutes of maximal effort: maximum power from the first second, holding whatever you can to the end.
• CP = average power over the last 30 seconds.
• W’ = the work done above CP in the first 150 seconds.

A brutal test. But faster: you get both metrics in 25 minutes.

CP vs FTP: when to use which

• Training for IRONMAN 70.3 or full-distance starts: FTP is enough, simpler and well proven over long training blocks.
• Training for T100 or fast short-course racing: CP + W’, for precision in the anaerobic zones.
• Age-group triathlon, building an aerobic base: FTP is enough.
• Advanced triathlon, fighting for seconds in a 70.3 race: CP + W’, for better pacing analysis.
• No power meter: neither, use heart rate and RPE.

For most age-groupers, FTP is still the practical standard. But for anyone already using WKO5 or Golden Cheetah, analysing every interval, moving to CP and W’ is the natural next step.

Training zones based on CP

The classic Coggan model has seven zones based on FTP. CP opens up an extra, seventh zone for anaerobic training:

• Z1 (recovery): below 55% CP
• Z2 (aerobic base): 56-75% CP
• Z3 (tempo): 76-90% CP
• Z4 (threshold): 91-105% CP
• Z5 (VO2max): 106-120% CP
• Z6 (anaerobic capacity): 121-150% CP, where you train W’
• Z7 (neuromuscular): above 150% CP, sprints of 10-30 s

For a triathlete, Z2 and Z4 matter most. Most training, 64-74% of FTP or roughly 65-75% of CP, is aerobic base. Z4 is threshold intervals. Z6 and Z7 are seasonal, done in blocks, not daily.

Practical takeaways

1. If there is no power meter, the power curve does not matter. Go back to heart rate and pace.
2. If FTP alone drives the app, that still works well for 80% of sessions. No need to change.
3. If every session gets analysed in TrainingPeaks or WKO, test CP and W’. See how the zones shift and whether the intervals are actually hitting their targets.
4. If racing in virtual events, W’ is critical. Knowing there is, say, 15 kJ available for attacks means planning when to spend it and when to save something for the finish.

CP and W’ are not magic. They are mathematics, mathematics that shows real physiology, not the round marketing numbers from a 20-minute FTP test.