A world-record team pursuit lasts about as long as a masters 1 km. In 2024 somebody finally published what that squad’s training actually looked like. It is not what you would guess.
The man with the bike
Every club car park contains at least one. He arrives on a road bike, wearing the latest kit, and within four minutes has told you his FTP. He refers to watts per kilogram unprompted. He holds views on tyre widths and pressure that he will share regardless of whether a question has been asked. His shoes make a noise on the boathouse floor like a small horse.
I spent years being politely dismissive of this man, on the grounds that cycling is a sport for people who could not get into a boat, like shaving their legs, and that anything performed sitting down while being assisted by rolling along is not, strictly speaking, an effort.
That was pure snobbery, and it cost me about a decade of useful information.
Why the pursuit, of all things
The men’s team pursuit is 4,000 m on a track. World-class squads cover it in a little under four minutes. A masters 1 km takes most of us three to four.
Same time domain. Same physiological problem: a hard start, a long middle that has to be held against rising acidosis, and a finish decided entirely by what you have left. Same tolerance for muscle mass — like rowing, and unlike running, cycling lets you carry the bulk required to make force without paying for it on every stride. Both sports reward an athlete who is strong and aerobically enormous at the same time, which is a rare combination that these two disciplines happen to share.
So if you wanted to know how to train a well-trained athlete for a three-to-four-minute maximal effort, the pursuit squad is the closest thing to a controlled experiment that exists — because unlike masters rowing, it is performed by full-time professionals whose data eventually gets published.
In 2024, Stadnyk, Stanley, Decker and Slattery published two papers in the International Journal of Sports Physiology and Performance doing exactly that: the training of an elite team pursuit squad across the 36 weeks preceding a world-record ride, and a comparison of the preparation phases before two consecutive world records.
I read them expecting to find a great deal of racing.
What the data actually shows
Five male riders, average age twenty-two. 4.4 W/kg at anaerobic threshold. VO₂max of 68.7 ml/kg/min. The reference performance is 3:49.804 for 4 km.
| The world-record programme | Figure |
|---|---|
| Total training, 36 weeks | 543 hours |
| Sessions completed | 436 |
| Sessions on the bike | 69.9% |
| Sessions in the gym | 12.7% |
| Training in the low-intensity zone (by power) | over 70% |
| Training above anaerobic threshold (by power) | 7.7% |
| Weekly volume in the 3-month prep phases | 805 and 725 min |
| Volume at or above world-record lead torque (55 N·m) | 4.4% |
| Volume above world-record lead power | 2.4% and 0.9% |
The distribution is pyramidal rather than polarised: a very large base, a moderate middle, a sliver at the top. Stephen Seiler has been making a version of this argument for twenty years, and the pursuit squad is about as clean a confirmation of it as you will find in a sport where the outcome is measured to the thousandth of a second.
But it is the last two rows of that table that stopped me.
A squad that broke a world record over four minutes spent somewhere between one and four per cent of its training at or above the intensity it was going to race at.
My resistance, in full
My first response was that this could not possibly be correct for a four-minute event. Four minutes is a sprint. It hurts from the second lap. It should therefore be trained by rehearsing the hurt, repeatedly, until the hurt becomes normal. Everybody knows this. It is the entire folk methodology of masters rowing.
My second response was that the pyramid must be an artefact of base-building in the early weeks, and that it would surely invert as the record approached.
It does not. The 81%-below-first-threshold figure comes from the twelve weeks immediately preceding the record ride. That is the sharp end of the programme, and it is still four-fifths easy.
My third response, which lasted considerably longer than I would like to admit, was that cycling is simply different.
It is different, in ways I will come to. But not on this. A 2,000 m row draws roughly 67–88% of its energy from aerobic metabolism, depending on which review you read and how the measurement was made. A 1 km at three to four minutes is more anaerobic than that — my working figure is somewhere around 55–75% aerobic, which I would flag as a reasoned estimate rather than a consensus value. Either way, the majority of the energy in the race you are training for comes from the system you build with easy work. The pain is anaerobic. The engine underneath it is not.
The bit where I did the arithmetic on myself
So I went back through my own training year and worked out what proportion of it was spent at or above 1 km race pace.
I am not going to publish the number. I will say that it was not one to four per cent, that it was not close to one to four per cent, and that a fair proportion of the work I had filed under “steady” was in reality sitting just below threshold, tiring me without training anything in particular.
Masters rowers do race-pace work as a lifestyle. Every third piece becomes a test. Every head race becomes a trial. Somebody goes hard early off the start and the outing quietly converts itself into a 500 m flat out. Proportionally, we do far more specific work than a world-record squad — and our easy work is not easy, which means we have arranged our training so that nothing in it is meaningfully distinct from anything else.
Their easy riding is genuinely easy because their hard work is genuinely hard; and the hard work is genuinely hard because there is so little of it that they arrive fresh enough to do it properly. Ours is a low-grade version of everything, all year round, and it produces precisely what you would expect.
Three things worth stealing
1. Torque, not just power. This is the one nobody in a boathouse is doing. The squad tracked force at the pedal — newton-metres — as a variable separate from watts, and reported how much of the year was spent at or above race torque: 55 N·m in the lead position, with 4.4% of total volume at or above it.
That is a strength-endurance target expressed as force per stroke rather than as average power. The rowing equivalent is peak handle force, and the number of strokes per week you spend above your race force. Your PM5 and RPS displays will draw you a force curve. Almost nobody uses it as a training variable. If your race is decided by the ability to keep applying a given force per stroke as fatigue arrives, then force per stroke is the thing to overload — and it can be overloaded at low rate and low aerobic cost, which is exactly what a masters athlete with a limited recovery budget should want.
2. Rate of force development is trained separately from maximal strength. During the preparation phase, one athlete’s mean peak rate of force development in a single-leg press was 71% higher at mid-phase and 46% higher at late-phase than at the start. Not maximal strength — the rate. How quickly the force arrives.
A 1 km begins from a standing start. The first ten strokes are almost entirely a rate-of-force problem, and they establish the position you will spend the next three minutes defending. Maximal strength work — 3 × 5 at 80% of a current one-rep maximum, which if you are already a Faster member, you have as a protocol in the member library — builds the ceiling. It does not train the speed of arrival. That requires short, explosive, fully recovered work, and it is the piece most masters programmes leave out altogether.
3. Specificity is a seasoning, applied late. They accumulated more volume above race power and above race torque in the season that produced the better ride. So specificity does matter — but it was a larger slice of a very small number, applied close to the event, sitting on top of a base built almost entirely out of easy riding. Not smeared thinly across the whole year as a permanent state of mild racing.
What does not transfer, and I would rather say so
They draft, and we don’t. This is the big one. In a team pursuit, riders sit in the wheel for three-quarters of the race at substantially reduced power, partially reconstituting their finite work capacity while they are there. Their race is intermittent. Our 1 km is continuous, unshielded, and has no wheel to hide behind. Any modelling of work-capacity balance built for the pursuit does not describe your race, and I would be wary of anyone selling it to you as though it did.
They are twenty-two, and professional. Twelve to thirteen and a half hours a week, more than four hundred sessions a year, with a physiologist attached. You cannot copy the volume. You can copy the shape, which is the more valuable thing in any case.
Cycling has no catch. Force arrives smoothly, through a crank. Rowing force arrives as a spike, against a stretcher, through a spine, twenty-eight to thirty-four times a minute. Force targets that are safe on a bicycle are not automatically safe on an ergometer or in a boat, and the cost of getting that wrong is a season.
What I changed
- The easy work got easier, and I stopped negotiating with the definition. Below the first threshold means below it, not slightly above it with a serious expression.
- Race-pace work became a budget rather than a habit. I now allocate it as a percentage of the year — a few per cent above threshold, and something in the low single digits at or above race power — instead of as sessions per week.
- I started counting race-force strokes, not only metres. Force per stroke is now a tracked variable with a weekly target above race force.
- I added rate-of-force work as a quality separate from maximal strength: short, explosive, fully recovered, early in the session, nowhere near failure.
And the fifth thing, which is really the whole article: I stopped treating “specific” as a synonym for “good”. Specificity is the last two per cent. Everything underneath it is built out of work that feels, at the time, like nothing much at all.
Which brings me back to the car park, and the man with the with the road bike. He was right and I was wrong, though not in the way either of us imagined. His sport taught me nothing about how to be fast for four minutes by going fast. It taught me how to be fast for four minutes by almost never doing it.
Faster membership includes every training plan, protocol and update on this site — so if you are already a member, go and open the sprint plan. It applies the budget above: the distribution, the force targets, and where the small quantity of specific work sits across the twelve weeks before a regatta. The worst outcome is that you disagree with me in an informed way, which is most of what this membership is for.