Cycling team tactics explained: why nobody wins alone

A rider crosses the line alone with his arms in the air and the result sheet records one name. Seven other men in the same kit roll in later, some of them a long way later, and between them they did most of the work that made the win possible. Nothing in the results says so.

That mismatch is the subject. Road cycling is scored one rider at a time and raced in squads, and the whole tactical apparatus of the sport exists in the gap between those two things.

Most attempts at getting cycling team tactics explained start with a list of job titles. Leader, domestique, climber, sprinter, and on down the roster. The titles are real but they are consequences rather than causes. Every one of them exists because of a single physical fact about riding a bicycle in company, and once that fact is understood properly the roles, the breakaways, the echelons and the sprint trains all fall out of it in order.

Cycling team tactics explained by one physical fact

Riding directly behind another rider at racing speed costs a great deal less than riding at the front. Not slightly less. So much less that a rider who spends five hours sheltered arrives at the finish with a usable reserve, while the rider who sheltered him arrives empty.

The reason is that on flat road at racing speed, almost everything slowing a bicycle down is air. Rolling resistance is there, and it matters, but it grows roughly in proportion to speed. Air resistance does not behave nearly so politely. The drag force rises with the square of the speed through the air, and because power is force multiplied by speed, the power needed to overcome that drag rises with the cube. Go ten per cent faster and the aerodynamic power demand goes up by about a third. Go twenty per cent faster and it roughly doubles.

That cube is the reason bike racing looks the way it does. It means the cost of leading is not merely high, it accelerates viciously as the race speeds up, so the faster the racing gets the more valuable shelter becomes. It also means the saving available to the rider behind is largest exactly when the race is hardest.

A rider moving through still air drags a volume of disturbed air along behind him. The pressure in that wake is lower than the pressure in front, the air inside it is already moving in the direction of travel, and a second rider who places his front wheel into it meets a much reduced apparent headwind. He is not being sucked along, whatever the commentary implies. He is simply being asked to push a smaller amount of air out of the way.

Alone on the front100%

Directly behind one rider74%

Fifth wheel in a line66%

Deep inside a large bunch58%

These numbers are invented and used only to show the shape of the effect. Real savings depend on speed, spacing, group size, rider position and wind direction, and are measured in wind tunnels and field tests rather than assumed. Nothing here describes a specific rider or race.

Show the numbers

Invented illustrative index: what it costs to hold the same speed

Item	Value
Alone on the front	100%
Directly behind one rider	74%
Fifth wheel in a line	66%
Deep inside a large bunch	58%

Three things follow from the shape of that, and each of them writes a chapter of the tactics.

The saving grows as the group grows, up to a point, because a big bunch disturbs a big volume of air and the riders in the middle of it are sheltered from several directions at once. That is why a peloton of a hundred and fifty riders can sit at a speed that would destroy a group of five.

The saving collapses when speed collapses. On a steep climb at low speed, most of the resistance is gravity, and gravity does not care who is in front. This is the single most important qualification to everything below, and it is why mountain stages look so different from flat ones.

And the saving depends entirely on where the wind is coming from. Directly behind is only the correct place to sit when the wind is directly ahead. Move the wind round to the side and the sheltered position moves with it, which is the beginning of the most brutal tactic in the sport. The mechanics of how a bunch actually shelters its riders, and how far that shelter reaches, deserve a closer look at the group itself.

The job titles are a description of who spends what, and when

A team arrives at a race with a defined result in mind and a budget of human effort to buy it with. The roles are the budget lines.

The leader is the rider the result is for. His job is to arrive at the decisive part of the race with more left than his rivals, which mostly means doing as little as possible for as long as possible. A leader who is seen on the front of the bunch on a flat stage is either in trouble or making a point.

Domestiques are the riders who spend themselves on his behalf. The work is unglamorous and constant: riding in the wind at the front, dropping back to the team car for bottles and food and ferrying them forward, closing gaps when a rival attacks, shepherding the leader through the last few hundred metres before a narrow section, and handing over a wheel or an entire bicycle when the leader has a mechanical failure. A domestique can be a rider capable of winning races and still finish twenty minutes down, because his own day ended before the race began to be decided.

The road captain is the one making decisions at speed. Directors talk to their riders, but a director is in a car behind a convoy, watching a screen and listening to race radio, and he cannot feel that the bunch has gone quiet or see that a rival's climber has just moved up three places on the left. The road captain can. He is usually an older rider whose own results are behind him, and his value is judgement rather than watts: which move to cover, which to ignore, when to move the leader forward, when the team is being drawn into doing somebody else's work.

Climbers are riders whose power relative to their body weight is high enough to hold a hard pace on a long gradient. A team's best climbing domestique is the one whose day matters most, because he is the rider still with the leader when the race actually happens.

The sprinter is a different animal entirely, protected all day in order to produce an effort measured in seconds. He is usually heavier than the climbers, which costs him in the mountains and is exactly what makes him fast on flat road.

Lead-out riders exist to deliver him. They are strong enough to ride very fast for a kilometre or two and are content to be passed in the final two hundred metres by the man they have just towed there.

Rouleurs are the engine room, riders built to hold a high steady speed on flat and rolling roads for a long time. They do the tempo work, the chasing and the front-line duty in crosswinds. Neither the lightest nor the fastest, and the reason anything gets controlled at all.

52%21%14%13%

• Sheltered in the bunch, doing nothing52%
• Riding on the front setting tempo21%
• Ferrying bottles and food from the car14%
• Closing gaps and covering attacks13%

An invented breakdown of a single rider's day, by share of the stage distance, used to show where the work sits. It is a constructed illustration and not a measurement of any real rider.

Show the numbers

Invented worked example: how one domestique's flat stage divides up

Item	Value
Sheltered in the bunch, doing nothing	52%
Riding on the front setting tempo	21%
Ferrying bottles and food from the car	14%
Closing gaps and covering attacks	13%

Two points about that shape. Most of a domestique's day really is spent doing nothing, because effort is a resource and the whole discipline is about not wasting it. And the bottle-carrying is not a joke role. A rider dropping back through a convoy, loading up, and riding back to the front through a moving peloton is doing a substantial effort on his own in the wind, several times a day, and it is one of the reasons a team leaves that job to specific riders rather than sharing it about.

Which riders a team can afford to carry for which jobs is settled long before the start, in the argument about who gets the eight places on a three-week roster.

Letting a breakaway go is a calculation, not a gift

The most interesting decision in road cycling happens in the first hour of a stage and is almost never explained on television.

A stage begins with a period of attacking. Riders go up the road in ones and threes, the bunch chases them back, somebody counter-attacks immediately, and this repeats. It is the fastest and hardest part of many stages, which surprises people who assume the finish is where the effort is. Eventually a group establishes itself, the bunch sits up, and the gap begins to grow.

What happened in that moment was not fatigue and it was not generosity. It was a decision, taken in several team cars at once, that this particular group of riders is acceptable.

The test is composition, not size. A director looks at the names in the move and asks a short list of questions. Is anybody in there close enough on overall time to threaten our position if he gains ten minutes? Is anybody in there a threat in the points competition we are trying to win? Is one of our own riders in it, which would mean we have no obligation to chase and can sit behind while others do? If the answers are all comfortable, the break is allowed to go and the bunch stops racing.

If a dangerous rider is in the move, the bunch does not stop, the move comes back, and the attacking starts again. This is why a breakaway sometimes takes an hour to form and sometimes forms in three minutes. It has almost nothing to do with how hard the riders attacked and everything to do with who they were. The reason a rider's overall position or points total changes how the whole bunch behaves comes straight from what each of the classifications actually counts.

There is a second, quieter reason to allow a break. A race with a breakaway up the road is a controlled race. Everybody knows where the danger is and how far away it is, the gap is a number that can be watched, and rival teams have somewhere to put their attacking instincts. A race with no breakaway is a race where every rival is still in the bunch with fresh legs and an itch. Many teams would rather manage a known gap than police two hundred nervous riders.

The arithmetic of the catch, and why the bunch is usually right

Once the break is up the road, the gap becomes a number in a spreadsheet, and the spreadsheet is not complicated.

A time gap is really a distance gap wearing a disguise. Take an invented example with clean numbers: a break riding at 40 kilometres an hour with a six-minute lead is four kilometres up the road, and every kilometre of that road is worth ninety seconds. If the bunch also rides at 40, the gap never changes, no matter how impressive the speed looks on screen. The gap only moves when the two speeds differ.

Now the useful part. If the bunch rides two kilometres an hour faster than the break, it takes back two kilometres of road every hour, which is three minutes an hour. Six minutes therefore takes two hours to close. If the bunch rides four kilometres an hour faster, the same six minutes takes one hour. Doubling the effort halves the time, which is the ordinary arithmetic of chasing, and the reason a chase either starts early and gently or late and violently.

• Bunch matches the break
• Bunch rides 2 kph faster
• Bunch rides 4 kph faster

A constructed example with round numbers. The break is assumed to hold a steady 40 kilometres an hour throughout and the gap is shown in seconds. No real race, rider or speed is being described.

Show the numbers

Invented worked example: what a six-minute gap does under three chase speeds

Item	Bunch matches the break	Bunch rides 2 kph faster	Bunch rides 4 kph faster
Gap established	360s	360s	360s
After 30 minutes	360s	270s	180s
After 1 hour	360s	180s	0s
After 90 minutes	360s	90s	0s
After 2 hours	360s	0s	0s

The real constraint is not time, it is road. A director does not ask how long the chase would take, he asks how many kilometres it needs, and compares that with the kilometres left. Chasing at four kilometres an hour faster for an hour means covering roughly forty-four kilometres of road while doing it. If there are thirty kilometres to the finish, that chase is arithmetically impossible however committed the riders are, and the sensible move is to accept the day is gone rather than empty the team for nothing.

This produces the pattern that makes stages readable. The gap grows for an hour or two while nobody cares, plateaus at whatever ceiling the controlling teams have decided on, sits there for a long time, and then falls steeply in the last fifty kilometres as the teams that need a bunch finish take the front and start converting seconds into kilometres. When the gap falls faster than that curve suggests it should, somebody has panicked. When it falls slower, teams that were expected to chase have quietly refused.

Two things break the arithmetic, and both of them are why breakaways ever survive.

The break is not a constant. Riders in it are cooperating only as long as cooperation serves them. As the finish approaches, the strongest riders in the move stop taking full turns because they are saving themselves to attack each other, and the group's speed drops at precisely the moment the chase is speeding up. A break of eight that turns into a break of three fighting each other is often slower than a break of eight rolling smoothly.

And the chase is not a single organism. It is several teams, each hoping another does the work. If two sprinters' teams commit and a third sits behind, the chase happens at two teams' pace, not three, and the arithmetic quietly changes underneath everybody.

1. 1The attacks start at kilometre zeroRiders go clear in small groups and are pulled back, repeatedly, often for an hour. This is usually the hardest racing of the day. Nobody is trying to win a stage yet; they are trying to be in the group that gets permission to leave.
2. 2Decision point, the composition testDirectors read the names, not the number. If nobody in the move threatens the overall standings or the points competition, and ideally if one of our own riders is in it, the move is acceptable. If a dangerous rider is present, the bunch keeps chasing and the attacking begins again.
3. 3The bunch sits up and the gap opensOnce nobody has an interest in chasing, the gap grows quickly, because the break is racing and the bunch is not. The riders in the break settle into a rotation, sharing the wind in short turns so no one of them pays the full cost of leading.
4. 4Decision point, how big a gap is allowedThe gap has a ceiling, chosen by the teams with something to protect. Too small and the break gives up, which returns the race to a bunch full of fresh attackers. Too large and the day is lost. A team rides tempo on the front to hold the number where it wants it.
5. 5The chase starts where the arithmetic saysA director converts the gap into the kilometres needed to close it and compares that with the kilometres remaining. Chasing earlier than necessary wastes riders. Chasing later than necessary loses the race. The teams that need a bunch finish take the front.
6. 6The break stops cooperatingInside the break, the strongest riders begin sitting on and saving themselves for the attack. The rotation slows exactly as the chase accelerates. This is often the moment a break that looked safe becomes doomed.
7. 7Decision point, does the terrain help the front or the backA late climb, a narrow technical run-in or a crosswind section favours the small group. A wide, straight, flat road favours the organised chase. The same gap with the same riders is safe on one course and hopeless on the other.
8. 8Caught, or survivesMost breaks are caught, usually with a handful of kilometres left, because the chase is designed to arrive exactly on time and not a minute early. A break survives when the chase started too late, when the teams chasing were fewer than expected, or when the terrain broke the bunch apart before it could finish the job.

The decision points are the parts that matter. Each one is a choice made by teams with competing interests, not an inevitability.

Riding on the front is expensive, so why would anyone do it

If shelter is so valuable, the obvious question is why any team volunteers to give it away for five hours.

Teams take the front for three reasons, and none of them is politeness.

The first is control of the gap. A team defending an overall lead cannot allow the break's advantage to exceed the margin it holds, so somebody has to ride at a speed that keeps the number where it needs to be. That work is a tap, not a switch. Ride slightly too hard and the break dies early, which hands the race back to a bunch full of attackers. Ride slightly too easy and the number climbs past the point of recovery.

The second is safety, and it is underrated. The front of a bunch is calmer, wider and much less likely to contain a crash. Riders further back are reacting to movements they cannot see, braking on the brakes of the man ahead, and taking corners at whatever speed the accordion effect dictates. A team burns two riders all day simply to keep its leader in the first fifteen wheels, because a broken collarbone costs more than two domestiques ever will.

The third is that a hard, even tempo is itself a weapon. Attacking out of a bunch riding at a steady high speed is enormously more expensive than attacking out of a bunch riding easy, because the attacker must generate the speed difference from a starting point that is already close to his limit. A team that rides a relentless tempo is not trying to drop anybody. It is making attacks unattractive, which keeps the race predictable, which is what a team with the strongest leader wants.

The price is that everybody else is resting. A team that spends four riders controlling a stage has four fewer riders for tomorrow, and its rivals, who did nothing all day, have all of theirs. This is the permanent tension in the sport: the team best placed to win is also the team obliged to pay for the racing, and over three weeks that bill compounds. It is why teams will sometimes engineer things so that a rival inherits the race lead, and with it the duty to control, for a week in the middle of a race.

Echelons are the cheapest way to end somebody's race

Change the wind direction and the geometry of shelter changes with it, which turns an ordinary flat road into the most destructive terrain in the sport.

With a headwind, the sheltered spot is directly behind the rider in front, and the group forms a line. With a crosswind, the apparent wind a rider experiences is a combination of the wind and his own motion, so it arrives at an angle. The sheltered spot moves round to the downwind side, behind and off to one shoulder. A group riding efficiently in a crosswind is therefore a diagonal line stretching from one edge of the road to the other. That diagonal is an echelon.

The vicious part is the road. A diagonal has a fixed length before it runs out of tarmac, so an echelon can only accommodate a certain number of riders, and that number is often far smaller than the bunch. Everybody who does not fit is left behind the last rider in the diagonal, which is not a sheltered position at all. They are in the gutter, riding in the full force of the wind, in single file, at a speed set by a group that is sharing the work.

That is not survivable for long. The line stretches, somebody's legs go, a gap appears, and the bunch is now two groups. And once separated, they rarely rejoin, because the front group has organised riders sharing the wind while the second group is disorganised, demoralised and chasing into the same wind with fewer usable riders.

The tactic follows directly. A team that knows a crosswind section is coming moves its whole squad to the front before it arrives, so that when the road turns, its riders are in the first echelon by default. Then it rides hard, immediately, to make the split happen while its rivals are still moving up. Minutes can be taken in a quarter of an hour on flat, unremarkable road, which is a strange thing to say about a sport people associate with mountains.

Defending against it costs a fortune. It means keeping the leader in the first twenty riders for an entire windy afternoon, which means several domestiques riding in the wind for hours just to hold position. Teams that get this wrong lose races on days that were not supposed to matter, and the honest ones admit afterwards that they lost it in the fifteen kilometres before the wind arrived, not during.

A lead-out train is a queue built in reverse

A bunch sprint looks like chaos and is closer to a scheduled operation. The lead-out train is the mechanism, and its logic is entirely about the order.

A sprinter can produce an enormous effort for a short time. The problem is that the last few kilometres of a flat stage are ridden very fast by everybody, so a sprinter who has to fight for his own position arrives at the line having spent his effort on getting there. The train's purpose is to let him spend nothing until the last possible moment.

The riders line up in reverse order of how long each can hold a hard effort. The first man in the train takes over the front with several kilometres to go and rides a long, hard, steady turn, fast enough that nobody can easily come past. He swings off. The next rider does the same over a shorter distance and at a higher speed. Then the last lead-out man, who is often a very fast rider in his own right, takes the final few hundred metres flat out. The sprinter sits on his wheel until he judges the moment, comes past, and finishes.

Each stage is shorter and faster than the one before, which is the whole design. The train is an accelerating structure, and the acceleration is what stops rival sprinters simply drafting the train and coming past at the end, because coming round a line that is still speeding up means riding in the wind for longer than you can afford.

Trains fail in a small number of well-understood ways. A rider takes his turn too early and blows before he has delivered anybody, leaving the sprinter exposed with a kilometre still to go. The train loses a man in a crash and the sequence collapses. Two trains arrive at the same corner at the same time and the tidy plan becomes a fight for one line. Or a rival sprinter with no train at all simply follows the strongest train and uses it for free, which is the standing complaint of every team that has ever built one. The finer detail of that sequence, and why a wheel is worth more than a metre of road, is worth taking apart on its own.

The other thing a train does is remove decisions. In the last three kilometres, a sprinter following a wheel he trusts is not making choices about lines, gaps or timing. He is doing one thing, and doing one thing at that speed is a large part of why it works.

In the mountains, the shelter goes and the pacing takes over

Everything above assumes speed high enough that air resistance dominates. On a long steep climb it does not, and the tactics change shape accordingly.

At climbing speeds, most of the effort goes into lifting rider and bicycle against gravity. A rider on a wheel still saves something, but not remotely what he saves on the flat, and no amount of teamwork can carry a rider up a mountain he cannot climb.

So the mountain domestique's job is not to shelter his leader. It is to set a pace.

The mechanism is worth stating plainly, because commentary tends to describe it as controlling the race without ever saying what is being controlled. A very hard steady tempo is an efficient way to hurt people. Riders who attack, get caught, attack again and get caught are paying a surcharge for every acceleration, because each one takes them briefly above the effort they can sustain and the recovery afterwards is never complete. A leader sitting behind a domestique riding a relentless even pace is doing none of that. He is riding a single, hard, steady effort, which is by some distance the cheapest way to cover the same ground at the same speed.

That is the actual trade. The team on the front is not preventing attacks, it is making them expensive, and it is making them expensive for everybody except its own leader, who is the one man on the climb not having to respond to anything.

The pace has to be chosen carefully, and the choice is a rider's own physiology rather than a number on a screen. Too easy and rivals sit comfortably, arrive fresh and attack at the top. Too hard and the domestique blows early and leaves his leader alone with three kilometres of climb still to come. A director tries to set that pace using power data, which is a large part of what those numbers on a power meter are for, though the useful figure is not the leader's ceiling. It is the pace the domestique can hold to a particular bend.

This is why the last domestique standing is the most valuable rider on a grand tour team. Anybody can ride tempo on the lower slopes. Very few riders can still be pacing a leader at the front on the final climb of a hard day, and a team that has one is a team whose leader spends the last twenty minutes doing the easiest hard ride available to anybody in the race.

The other mountain job is smaller and equally decisive: waiting. If a leader punctures or crashes, a teammate stops, gives him a wheel or a bicycle, and then paces him back to the group. Pacing back is drafting used as a rescue, and it is the one moment on a climb where the shelter still matters, because a rider chasing at high speed on a shallower gradient is back in a regime where the wind is the enemy again.

The time trial is where the teamwork stops

There is one discipline where all of it is switched off, and looking at it clarifies everything else.

In an individual time trial, riders start alone at intervals and are forbidden to shelter behind each other or behind a vehicle. The regulations require riders to keep clear of the rider they are overtaking and prohibit pacing, and the commissaires enforce it with time penalties or disqualification. The point of the rule is precise: the event only measures one rider if nobody is allowed to give another an aerodynamic advantage.

What remains for the team is real but narrow. The mechanics build the bike and the position. The director drives behind with time checks and pacing instructions, which is information rather than assistance. The team decides the gearing and the tyre choice and the warm-up. Nobody rides in front.

The result is that time trials are where a rider's own ability is measured with the team subtracted, and it is why a rider who is protected all week can still lose a stage race in an afternoon. The gaps are usually larger than a flat road stage produces, because on a road stage everybody gets the same time and here nobody does. Riding alone against the clock is a different sport wearing the same clothes, and it has its own set of problems and equipment rules.

The team time trial is the exception that proves the point. There, the whole squad rides together in a rotation, drafting is the entire event, and the finishing time is taken on a specified rider rather than the first one across the line, with the relevant position set out in that race's regulations. It is the only format where the cooperation that decides every road stage is also the thing being scored.

Race radios turned tactics into an argument about who is racing

Modern teams ride with earpieces. The director sees the television feed and the timing data, talks to the riders, and the riders talk back. Race radio also carries neutral information to everybody: time gaps, crashes, hazards, road furniture.

Nobody disputes that radios make racing safer. A warning about a traffic island in a bunch travelling at speed is worth having, and the alternative is a chalkboard held out of a car window.

The argument is about the tactical half. Critics say radios have flattened the racing, because a director watching a screen can identify a threat within seconds and instruct four riders to close it before anybody in the bunch has even noticed the move. Attacks that used to succeed because nobody reacted in time now fail because the reaction is instant and centrally organised. On that reading, the sport lost something when the decisions moved from the road to the car.

The defenders make two replies. Removing radios does not remove information, it just makes it worse: riders still get time gaps from the chalkboard, from the motorbikes and from shouting at team cars, so the effect is a slower, patchier version of the same thing rather than a return to instinct. And the riders with the best instincts, the experienced road captains, would simply become more valuable, which advantages the same big teams the ban was supposed to constrain.

Both sides have a point, and races have been run under both regimes, which is the sensible way to settle it. What is not in dispute is that intelligence is now a tactical resource in its own right. Teams model the route in advance, know which fifteen kilometres of a stage face a crosswind, know where the road narrows, and plan the day around three specific places on it. The instruction on the radio is usually the last step in a decision that was made the night before.

Teams that are trying to beat each other still cooperate, briefly

Alliances in cycling are not friendships. They are overlapping interests that last exactly as long as the interest does.

The common case is a chase. Three teams have fast finishers and want a bunch sprint. None of them can close a large gap alone. So their riders rotate on the front together, taking turns, dragging the whole peloton along, and the moment the break is caught they go back to trying to beat each other in a straight line. Nothing was agreed beyond the obvious, and nothing outlasts the catch.

The interesting cases are the refusals. A team that thinks it cannot win a sprint has no reason to help close a gap, and can simply decline to send anybody forward, which forces the others to do more work and arrive at the finish with weaker lead-out riders. That is not passivity. It is a tactic, and a good one.

Then there is the position nobody wants, which is being the only team with something to defend. A squad holding the race lead can find itself doing all the chasing every day while every rival sits behind and lets it. There is no rule requiring anybody to help, and no team helps unless helping serves its own result. Refusing to work while a rival empties itself is a completely legitimate way to win a bike race, and over three weeks it is one of the most effective.

Cooperation inside a breakaway is the same logic at a smaller scale. A break works together because a break that stops working gets caught and nobody in it wins anything. As soon as one rider decides he is stronger than the others, the cooperation ends, the speed drops, and quite often everybody loses to the bunch behind. It is the clearest example in the sport of a group that only survives while every member believes cooperating is better than defecting.

Why the strongest rider in the race so often loses it

This is the thing the physics has been building to.

Suppose two riders have identical ability. One spends the day sheltered by teammates, positioned near the front without doing any work, handed bottles instead of fetching them, and paced up the first two-thirds of every climb by a rider who then rides himself out of the race. The other spends the day fighting for his own position, riding in the wind whenever the bunch strings out, dropping back to the car himself, and chasing his own way back after a puncture.

By the time the race is actually decided, they are not identical any more. Nothing dramatic happened to the second rider. He simply paid a hundred small aerodynamic bills that the first rider did not, and the arithmetic of the cube law means those bills were larger than they looked.

Give the second rider an advantage in raw ability and the sums change but the structure does not. He now has to be enough better to cover the difference in what he has spent, and on a hard day that gap is not small. This is why a strong rider on a weak team can look outstanding for two hours and be ordinary in the final thirty minutes, and why the same rider moving to a squad that can protect him appears to improve overnight without anything about his physiology changing at all.

The same effect shows up in positioning, which is where teams add value in a way that never appears in any data. The last three kilometres before a climb, a cobbled sector or a narrow corner are a fight, and a rider with three teammates to move him forward arrives at the front while a rider on his own arrives fortieth. Arriving fortieth at the bottom of a climb means starting the climb in the wind, chasing, having already lost several seconds to an accordion effect that the front of the bunch never felt. It is the least visible advantage in the sport and one of the biggest, which is also why the aerodynamic side of the sport has become an engineering problem teams take seriously.

None of which means support beats ability. A leader who cannot climb with the best will lose whatever his team does, because on the final climb the domestiques have gone and there is nothing left to hide behind. Team tactics do not create winners. They decide who arrives at the decisive moment in a condition to use what they have, and in a sport where the decisive moment is often twenty minutes at the end of a five-hour day, that is most of the outcome.

Cycling team tactics explained by watching the front of the race

Turn on a stage in progress and the coverage will show you the break, the gap and the scenery. The tactics are visible in a different place: the first fifteen riders of the peloton.

Four questions read almost any race.

Whose jerseys are on the front, and how many of them. One team with two riders is managing something. One team with five riders is committed and will be paying for it tomorrow. Three teams rotating together are a chase with a deadline. Nobody on the front means nobody has decided the break is a problem yet.

Where is the leader sitting. A protected rider should be around fifth to fifteenth wheel, surrounded by his own kit, doing nothing. If he is on the front, or drifting back through the bunch, something has gone wrong or is about to.

What is the gap doing relative to the road remaining, rather than in isolation. Convert it: a gap in minutes is only meaningful next to the kilometres left and the number of teams willing to work. Two minutes with sixty kilometres to go and four teams chasing is a break that is already caught. The same two minutes with fifteen kilometres to go and one team chasing is a break that has won.

And which way is the wind blowing. If the bunch is riding on one side of the road in a diagonal, the race is already happening, whatever the commentary is talking about.

Read those four and the sport stops being a five-hour procession with fifteen interesting minutes at the end. It becomes a long argument about who is spending what, conducted at fifty kilometres an hour, and settled by the rider who spent least. More on how the racing works, and on the rules that shape it, sits in the cycling archive.