Skip to content
Coaching18 min read

RACING WITH A POWER METER: HOW TO USE YOUR NUMBERS WHEN IT ACTUALLY MATTERS

By anthony-walsh

WHO THIS IS FOR

IS THIS YOU?

  • Riders who train with power but race entirely by feel and want to close that gap
  • Anyone who overcooks the first 20 minutes of every race because adrenaline masks the effort
  • Cyclists unsure what to display on their head unit during a race versus training
  • Riders who want to use post-race file analysis to find where they left performance on the table

THE ROADMAN VIEW

The Roadman View

  • I see so many riders train with power all winter and then race entirely by feel. Your power meter does not stop being useful when you pin on a number — but you need to strip your race screen down to three fields: three-second power, lap normalised power, and heart rate.
  • The first 20 minutes is where I see riders blow their race. 280 watts feels like 240 when the adrenaline is up. That small overshoot compounds across the hours that follow, and by then it is too late to claw it back.
  • I have spoken to coaches who say the same thing: power meters are decision-support tools, not decision-making tools. If the winning move goes up the road and you have the legs, follow it. The screen should inform your racing, not replace it.

You bought the power meter. You tested your FTP. You built a season of structured training around zones that Andrew Coggan and Hunter Allen defined decades ago. Every interval session is prescribed by watts, executed by watts, reviewed by watts.

Then the race starts and you stare at your stem for ninety minutes, riding on feel, reacting to wheels, and hoping it all works out.

This is the gap that costs amateur cyclists the most. Not fitness. Not equipment. The disconnect between having precise data in training and ignoring it entirely when the numbers actually matter. Your power meter does not stop being useful when you pin on a number. It becomes more useful — if you know what to look at, what to ignore, and when to override the screen entirely.

The Training-Racing Disconnect

The problem is cultural as much as it is technical. Power meters arrived in cycling through the training door. Coggan's power zones, Allen's training protocols, the entire TrainingPeaks ecosystem — all of it was built to structure preparation. Racing was left to instinct, experience, and whatever you could read from the legs around you.

And instinct works. It got riders through races long before anyone strapped a strain gauge to a crank arm. But instinct has a well-documented flaw: it lies to you in the first twenty minutes.

When the adrenaline is up, your legs are fresh, and the bunch is rolling at speed, 280 watts feels like 240. Your body hasn't caught up to the cost yet. Heart rate lags. Lactate is still being cleared. The signals that usually tell you "this is too hard" are delayed by excitement and glycogen. So you overcook it. Not dramatically. Just five or ten per cent above where you should be. And that small overshoot compounds across the hours that follow.

This is fixable. Not with more fitness — with better use of the data you are already collecting.

Setting Up Your Race Screen

Your training head unit screen is wrong for racing. Most riders run four to eight data fields in training: power, heart rate, cadence, speed, distance, elapsed time, normalised power, TSS. That density is fine when you are doing 2 x 20 threshold intervals on a quiet road with nothing to think about except holding 270 watts.

In a race, you are holding a wheel at 45 kph in a group of forty riders through corners you have never seen before. You get half-second glances at your screen. If you cannot absorb the information in that half second, the data field is noise.

The race screen: three fields.

  1. Three-second power. Not instantaneous power, which jumps wildly and tells you nothing useful. Three-second smoothing gives you a real-time read on current effort without the spikes.
  2. Lap normalised power. This is your running account of cumulative cost. Hit the lap button at the start and this number tells you, at any point during the race, how much you have spent relative to your capacity. If your FTP is 270 watts and your lap NP is sitting at 260 after twenty minutes of a two-hour race, you are deep in the red whether it feels like it or not.
  3. Heart rate. Not as a target — as a cross-reference. If power and heart rate are both climbing, you are working harder. If heart rate is climbing but power is dropping, you are fatiguing. If heart rate is low but power is high, conditions are good and you are riding within yourself.

That is it. Remove speed — it is meaningless in a bunch. Remove distance — you are not pacing by kilometres. Remove cadence — you already know how to pedal. Remove elapsed time unless you are riding a time trial. Every field you add dilutes the three that matter.

Set this up as a dedicated race profile on your head unit. Name it something obvious. Do not hunt for it in the settings while you are warming up on the start line.

The First 20 Minutes: The Most Expensive Mistake in Amateur Racing

If there is a single piece of power-meter advice that would improve every amateur race result overnight, it is this: ride the first twenty minutes to a number, not to a feeling.

Coggan identified this pattern years ago. The physiological cost of the opening phase of a race is almost always higher than riders perceive. Fresh legs, full glycogen, and adrenaline create a sensation of effortlessness that does not reflect the actual metabolic cost. Riders burn through matches they will need later, and the bill comes due at seventy or eighty per cent of the way through.

Here is what this looks like with real numbers. Take a rider with a 260-watt FTP entering a 100-kilometre road race. The sustainable normalised power for a three-hour effort is somewhere around 210-220 watts — roughly 80 to 85 per cent of FTP. If that rider starts the race at 240-250 watts normalised (which feels comfortable in the first twenty minutes), they have spent the opening phase at 92-96 per cent of FTP.

That is not sustainable for three hours. It is barely sustainable for one. But by the time the rider realises, they are forty-five minutes in, the damage is done, and the second half becomes a survival exercise.

The fix: Use the first twenty minutes as a power-capped warm-up. Set a hard ceiling on your three-second power: no more than 85 per cent of FTP unless the bunch forces you above it. Let people ride off the front. Let the early surges go. Sit in the wheels, read the three-second number, and refuse to chase feelings.

This requires discipline, especially in amateur races where the start is chaotic and every rider who has watched a pro race thinks they should be attacking from the gun. You are not those riders. You are managing a finite energy budget across a known duration, and the first twenty minutes is the most expensive place to overspend.

For sportive riders, this is even more critical. A road race might reward an aggressive start if it puts you in a tactical position. A sportive never does. Pacing an event is about even distribution, and the opening kilometres are where most sportive riders destroy their second half.

Reading the Numbers Mid-Race: Normalised Power vs Average Power

Your power file after a race will show two key numbers: average power and normalised power. Understanding the difference between them mid-race — not just in post-ride analysis — is where power-literate racers gain an edge.

Average power is the arithmetic mean. If you ride at 200 watts for thirty minutes and 300 watts for thirty minutes, your average is 250 watts. Simple.

Normalised power is the metabolically weighted average. It accounts for the fact that surges cost your body disproportionately more than steady-state riding. Andrew Coggan developed the algorithm specifically because average power was misleading in variable-effort events — which is what every road race and most sportives are.

Here is where it gets really interesting. Two riders can finish a race with identical average power and completely different levels of fatigue. Rider A holds 230 watts steady for two hours. Rider B alternates between 350-watt surges and 150-watt recoveries, averaging 230 watts. Rider B's normalised power will be significantly higher — possibly 260-270 watts — because those surges carry a metabolic penalty that simple averaging ignores.

In a race, this means your lap normalised power is the number that tells the truth about how you are racing. Average power lies in variable conditions. When you glance down and see your NP climbing, you are spending energy faster than the average figure suggests.

Practical mid-race use:

  • If your lap NP is above 90 per cent of FTP in the first hour of a long race, you need to ease back. Sit deeper in the wheels. Skip a few surges. Let the group do more of the work.
  • If your lap NP is at 75-80 per cent of FTP with an hour to go, you have matches left. You can follow moves, bridge gaps, or increase your effort on the final climb.
  • If your NP is climbing but average power is stable, the race is getting more variable — more surges, more accelerations out of corners, more tactical riding. This costs you more than a steady tempo even if the average looks manageable.

The variability index — NP divided by average power — quantifies this. A VI of 1.05 means the race has been relatively steady. A VI above 1.15 means heavy surging. In a criterium, you might see 1.20 or higher. The higher the VI, the faster you are burning through your reserves relative to what the average number implies. For a deeper read on what these metrics reveal, the power file analysis guide walks through a worked example.

When to Follow an Attack: The Power Maths

This is the tactical application that separates riders who own a power meter from riders who use one. The question is never simply "can I follow that attack?" It is "what will it cost me to follow, and is the potential return worth the expenditure?"

Bridging to a breakaway or following an acceleration requires a sustained effort well above threshold. The typical cost is 120-150 per cent of FTP for anywhere from thirty seconds to two minutes, depending on the gap and the speed differential.

Your power meter gives you a framework for making this decision in real time.

The 130 per cent rule (a useful heuristic, not a law):

If bridging to a move will cost you more than 130 per cent of FTP for longer than 90 seconds, the physiological cost is high enough that you need a very good reason to go. Either you are racing for the win and this is the decisive move, or you let it go. The maths says the recovery cost of that effort will compromise your performance for the next ten to fifteen minutes, during which the race could move again.

Below that threshold — say, 120 per cent of FTP for forty-five seconds — the cost is manageable. You will recover within a few minutes if you can sit in the wheels once you bridge across. This is the kind of move worth making when the group is strong, the gap to the bunch is growing, and you can see the race splitting.

In practice:

When someone attacks, you have three to five seconds to decide. Glance at your three-second power as you start to respond. If you are sitting at 340 watts and your FTP is 260, you are at 130 per cent. Now the question is: how long will this effort last? If the rider ahead is ten metres up the road and accelerating, thirty seconds of that effort bridges the gap. Manageable. If the rider is a hundred metres up and you are into a headwind, that is two minutes at threshold-plus. The return needs to justify the cost.

This is not about being conservative. It is about being selective. Races are decided by the last hard effort, not the third or fourth one. Every premature bridge burns a match that could have decided the result in the closing kilometres. Check your training load trends before the race — knowing your current form helps you gauge how many matches you actually have.

Sportive Pacing vs Road Race Pacing

The same power meter, mounted on the same bike, requires completely different strategies depending on whether you are racing against other riders or racing against the clock.

Sportive pacing is an exercise in energy management. The goal is to spread a known quantity of work as evenly as possible across a known distance. This is closest to what Coggan and Allen designed power training for — it is essentially a long time trial with hills. Your target is a normalised power you can sustain for the full duration, and every deviation from that target costs you more than it returns.

For a 4-5 hour sportive, most riders should target 70-75 per cent of FTP as their normalised power. Climbs will push you above that — perhaps to 85-90 per cent — and descents will bring you below it. The key is that the overall NP across the event stays in that sustainable range. Start at 80 per cent because it feels easy, and you will pay for it on the climbs in the second half. The sportive pacing guide covers the even-effort model in full.

Road race pacing throws all of that out the window — partially. The race is stochastic. You do not control the pace. The bunch surges, attacks, slows, accelerates through corners, strings out on climbs. Your power trace after a road race looks nothing like your power trace after a well-paced sportive. It looks like a heart-rate trace from someone watching a horror film.

But the underlying physiology does not change. You still have a finite capacity. Normalised power still reflects the true cost. The difference is that in a road race, you choose which surges to follow and which to let go. That is where the power meter becomes a tactical tool rather than a pacing tool.

The practical distinction:

In a sportive, the power meter is the boss. You ride to the number. If the group ahead is riding above your target, you let them go. If the climb demands 20 watts more than your target, you accept a slower pace. Even effort wins.

In a road race, the power meter is an adviser. You ride to the situation and check the number to assess the cost. You follow attacks that matter regardless of what the screen says. But between those decisive moments — which might total ten minutes in a two-hour race — you use the data to sit as low in the energy budget as possible. Soft-pedal when the bunch eases. Sit sixth or seventh wheel instead of first or second. Let your three-second power drop to 150 watts in the lulls so that your NP stays manageable for the efforts that decide the race.

Most amateur racers make the mistake of riding a road race like a sportive — constant high tempo, always near the front, steady power output. That is a recipe for being strong with thirty kilometres to go and empty with three. The riders who win amateur races are the ones who spend the least energy between the moves that matter.

Post-Race File Analysis: What to Look For

The race is where you use the power meter. The post-race file is where the power meter earns its long-term value.

Most riders glance at the summary — average power, normalised power, max power, TSS — and close the file. That is leaving the most useful information untouched.

Five things to look for in every race file:

1. Power-Duration Curve vs Training Bests

Overlay your race-day power-duration curve against your all-time or recent training bests. Where did you meet or exceed your training numbers? Where did you fall short? If your 5-minute power in the race was 95 per cent of your training best but your 20-minute power was only 80 per cent, you know exactly where the gap is. You were willing to go deep on short efforts but paced the sustained efforts too conservatively — or you were already too fatigued from early overcooking to hold threshold.

2. First-Half vs Second-Half Power

Split the race file at the midpoint. Compare normalised power for the first half against the second half. A drop of 5-8 per cent is normal in a long race. A drop of 15 per cent or more is a red flag — you went too hard early, you underfuelled, or both. This is the single most diagnostic split for pacing errors and it takes thirty seconds to check. The post-race analysis guide walks through more of these splits in detail.

3. Match Count

Look at the number of efforts above 120 per cent of FTP that lasted longer than thirty seconds. These are your "matches" — the hard accelerations that draw from your anaerobic reserves. Count them. Then look at when they occurred. If you burned six matches in the first third of the race and only two in the final third, you over-committed early. If you had matches left at the finish, you raced too conservatively. The ideal distribution puts the majority of your hard efforts in the decisive phase of the race.

4. Normalised Power by Section

If your head unit supports laps or if you can manually split the file by road section (climbs, flats, technical sections), compare your normalised power across each. Climbs should show your highest NP. Flat sections in a bunch should show the lowest. If your NP on flat group sections is nearly as high as your climbing NP, you were working too hard in the wheels — either poor positioning, constant surging, or braking and accelerating through corners.

5. Heart Rate Drift

Compare heart rate vs power across the race. Early in the race, a given power output should correspond to a predictable heart rate based on your training data. If heart rate begins climbing while power stays constant or drops, you are fatiguing — either from accumulated effort, dehydration, or thermal stress. The rate of that drift tells you about your endurance fitness and your fuelling strategy. A steep drift suggests the engine is running out of fuel.

Use the race predictor before your next event to set realistic target ranges based on your current power profile, so your post-race analysis has a benchmark to measure against.

Common Mistakes and How to Fix Them

Staring at the screen instead of the road. A power meter is a glance tool in a race, not a focus tool. Half a second, one number, eyes back up. If you are reading data for two or three seconds at a time, you are losing wheels, missing moves, and creating a safety risk. Practise the glance in training until it is automatic.

Using training screens in a race. Already covered above, but it bears repeating. Your eight-field training screen is a liability when you are riding in a bunch. Set up a dedicated race profile with three fields and use it for every race until it becomes second nature.

Chasing average power instead of normalised power. Average power is a misleading target in any race with variable effort — which is every race. A rider who averages 220 watts in a punchy road race might have a normalised power of 260 watts and a TSS that tells a very different story from what the average implies. Monitor NP. Forget the average until post-ride review.

Refusing to exceed threshold because "the numbers say no." If you are in a race and the decisive move is happening, go with it. A power meter is a decision-support tool, not a governor. The worst thing you can do with power data is use it as an excuse to sit in the bunch when the race is being won up the road. The data tells you the cost of the decision. You make the decision.

Never reviewing race files. The marginal gains from a power meter are not in the race itself — they are in the accumulated learning from reviewing every race file against your training data over months and seasons. One file teaches you a small lesson. Twenty files across a season teach you your racing pattern. You start to see where you consistently overspend, where you leave watts on the table, where your fitness is the real limiter and where your tactics are the bottleneck.

Ignoring the data between key efforts. Most amateur racers pay attention to power during the hard moments — the climbs, the attacks, the crosswind sections. The opportunity is in the other eighty per cent of the race. Can you sit at 140 watts instead of 180 watts when the bunch is cruising? Can you coast through corners instead of pedalling through them? The cumulative saving from riding at the bottom of the bunch's speed range during non-critical moments is substantial across a two-hour race. That is energy you can spend when it counts.

The Power Meter Is Not the Racer

Everything in this article is about using data better. None of it replaces the things that actually win races: positioning, reading the bunch, timing your effort, having the courage to follow a move or the patience to let one go.

The power meter gives you information. Information reduces uncertainty. In a race full of uncertain decisions — follow or sit, push or wait, commit or conserve — reducing uncertainty by even a small margin compounds across the hours.

Train with power. Race with power. But race first. The numbers are there to make better riders into smarter racers. They are not there to turn racers into accountants.

If you want to work through your own race files with other riders who are asking the same questions, the Roadman community on Skool is where those conversations happen. Bring the file. Bring the questions. We will help you find the answers in the data.

FAQ

FREQUENTLY ASKED QUESTIONS

What power should I aim for during a road race?
There is no single target power for a road race because the effort is stochastic — constant surges, recoveries, and tactical accelerations. Instead, monitor your normalised power across laps and compare it against your training benchmarks. If your normalised power is already at 95 per cent of FTP in the first third of the race, you are overcooking it. Save your matches for the moments that decide the result.
Should I look at my power meter during a sprint?
No. Sprints last 10-20 seconds and are won by timing, positioning, and commitment, not by hitting a number. Looking down costs you a bike length and breaks your focus. Power data from sprints is useful in post-race analysis — not in the moment.
How do I set up my head unit for racing?
Create a dedicated race screen with three fields: three-second power, lap normalised power, and heart rate. Remove speed, distance, cadence, and any other fields you use in training. In a race, you need to know current effort, cumulative cost, and cardiac response. Everything else is a distraction at 40 kph in a bunch.
Is normalised power more useful than average power in a race?
Yes. Average power underestimates the physiological cost of variable-effort racing. Normalised power accounts for the disproportionate cost of surges — a rider averaging 200W with constant 400W spikes is working much harder than a rider averaging 200W at a steady state. In any race with climbs, corners, or crosswinds, normalised power gives you the real picture.
When should I ignore my power meter during a race?
When the race situation demands it. If the winning move is going up the road and you have the legs, follow it regardless of what the screen says. Power meters are decision-support tools, not decision-making tools. The data helps you choose between two roughly equal options. It should never stop you from racing.

KEEP READING — THE SATURDAY SPIN

The week's training takeaways, pro insights, and what to do about them. 30,000+ serious cyclists open it every Saturday.

AW

ANTHONY WALSH

Host of the Roadman Cycling Podcast

RELATED CONTENT

Episode
Coaching18:14

WHY YOUR ZONE 2 TRAINING ISN'T WORKING & HOW TO FIX IT | ROADMAN CYCLING PODCAST

Zone 2 works brilliantly for some riders and leaves others stuck, tired and going nowhere. Anthony unpacks the missing context behind endurance training advice — drawing on insights from Mattias Reck, coach to Mads Pedersen — and explains why the zone is relative but the cost is absolute, and what time-crunched riders should actually do.

Episode
Coaching57:04

MADS PEDERSEN'S COACH ON THE TRAINING BEHIND A WORLD CHAMPION | ROADMAN CYCLING PODCAST

Mattias Reck has coached Mads Pedersen since 2017 — from a raw young rider into a world champion, Classics winner and Grand Tour stage hunter. He sits down with Anthony to talk about how you actually build a rider who can turn brutal racing into controlled chaos: resilience, recovery capacity, and the years it takes to learn how much training one athlete can really absorb.

with Mattias Reck

Episode
Coaching28:44

DOES AERO MATTER FOR AMATEURS? WHAT THE SCIENCE ACTUALLY SAYS | ROADMAN CYCLING PODCAST

Aero gets sold as free speed for pros, time triallists and triathletes — but does it matter for the everyday rider doing club runs and sportives? This solo episode cuts through the marketing. Your body, not your bike, makes most of the drag, which means the biggest gains are the cheapest ones: how you sit, how you hold the bars, and whether your kit actually fits. Here's when aero is real free speed, when it's expensive nonsense, and the $0 changes worth making first.