Gravel or Road: How Surface Choice Changes What You Should Train

Gravel is not road riding that happens to be harder. It is a different shape of effort, and that changes which sessions actually prepare you for it.

Riders planning a first gravel trip usually train for it the way they train for everything else, then find the ride demands something their preparation never asked for. The distance was right. The elevation was right. The problem was that the power file of a gravel day looks nothing like the power file of a road day, even when the averages match.

Vermont and upstate New York are the example, because that region genuinely offers both in the same place and forces the choice. I have not ridden there. What follows is about the physics of the surface and the training that follows from it, which travels perfectly well.

What the surface actually costs you

Start with the part that can be calculated rather than asserted. Rolling resistance is a force that depends on the coefficient of rolling resistance, the mass, and gravity:

Forty-six watts is a large number. For a rider with an FTP of 250 W it is nearly a fifth of threshold, spent purely on the surface. This is why a gravel ride at the same average speed as a road ride is a substantially harder day, and why gravel riders talk about distance so differently.

Those coefficients are typical published values rather than measurements taken here, and real gravel varies enormously. Hardpack is close to asphalt. Deep loose stuff is far worse than 0.012. The point is the order of magnitude: the surface is not a detail, it is one of the largest terms in the equation.

Three other things that change with the surface

Power becomes spiky whether you like it or not

On tarmac you choose your power. On gravel the surface chooses for you. A loose section, a steep pitch where you cannot afford to lose momentum, a rut that has to be powered out of: each one is a short involuntary surge. Over five hours those surges add up to a meaningful amount of work done above the pace you intended.

This is the same physiological problem described in the article on riding in cities, arriving from a completely different direction. Average power looks reasonable. The distribution is what tires you out.

Cadence drops and torque goes up

Climbing a loose surface means you often cannot spin, because spinning breaks traction. You end up grinding at 60 to 70 rpm on gradients where you would happily sit at 85 on tarmac. Same power, more force per pedal stroke, and a noticeably different kind of fatigue in the legs afterwards.

Your upper body does actual work

Six hours of vibration and steering correction is genuinely fatiguing in the hands, forearms, shoulders and neck. Road riding asks for almost none of this, and neither does indoor riding, which is the single biggest gap in preparing for gravel on a trainer.

Reproducing gravel demands on Zwift

Zwift does model surfaces. Dirt and gravel sections carry a higher rolling resistance in-game, and the effect on your speed at a given power is real and noticeable, particularly if you are on a road bike in the game rather than a gravel one. That helps with the feel and with equipment choice, and there is more on the in-game equipment model in the equipment article.

What it does not reproduce is the vibration or the involuntary surging. For those you have to build the sessions deliberately.

• Over-unders. 4 × 8 minutes alternating 2 minutes at 95% of FTP and 1 minute at 110%. This is the closest structured equivalent to what a rough surface does to your power file, and it trains the specific ability to recover while still working.
• Low-cadence torque work. 5 × 5 minutes at 60 rpm at tempo power, on a climb, in simulation mode rather than ERG. Builds the force-per-stroke tolerance that a loose climb demands.
• Standing repeats. 10 × 1 minute out of the saddle at threshold. Gravel climbing involves far more standing than road climbing, and standing is a skill that decays.
• Long endurance rides with surges. Every 10 minutes, 30 seconds hard, then straight back to endurance pace without a recovery period. Unpleasant, and closer to a gravel day than any steady ride.

Do not use ERG mode for the low-cadence work. ERG holds power by adjusting resistance, so as your cadence drops it adds resistance, which drops your cadence further. It fights precisely what you are trying to practise.

Reproducing road demands

The road version is more straightforward, which is part of why road riders are often better trained for their event than gravel riders are for theirs.

• Sustained tempo and threshold. Long blocks at a steady effort, because that is what smooth tarmac permits and rewards.
• Long climbs ridden whole. The Alpe or Ven-Top, at a genuinely steady power.
• Aerodynamic position work. Above about 30 km/h air resistance dominates everything else, so time spent getting comfortable low is time well spent. The physics is in the drafting article.

Why tyres matter more on gravel than position does

There is one number that explains most of the equipment advice you will ever read about either surface, and it is the speed at which air resistance overtakes rolling resistance.

Rolling resistance grows in proportion to speed. Air resistance grows with the cube of it. So there is a crossover speed below which the surface is your main enemy and above which the air is. Setting the two powers equal and solving gives:

On tarmac you are above the crossover almost all the time, which is why road cycling is obsessed with position, clothing and wheels. Everything that reduces drag pays, continuously, from 15 km/h upward.

On gravel you spend a large fraction of the day below 24 km/h, which flips the priority completely. Below the crossover, tyre choice and pressure do more for your speed than any aerodynamic change, and getting low on the bars buys you very little. This is the whole reason gravel riders talk endlessly about tyres and hardly at all about aero bars.

The practical version: on gravel, spend your attention on tyre width and pressure, and run less pressure than instinct suggests, because on a rough surface a harder tyre bounces rather than rolls and the bouncing costs more than the deformation saves. On road, the same attention is better spent on position.

Choosing between them

The honest framing is not which is harder. It is which kind of day you want, because they fail in different ways.

If you only have the fitness for one, take gravel at two thirds the distance you would plan on the road. That is a rule of thumb from the rolling resistance arithmetic above rather than a measured conversion, but it is a great deal closer than planning the same distance for both.

Where indoor preparation genuinely runs out

• Handling on a loose surface. Not simulated, not simulatable, and the thing most likely to end your day badly.
• Vibration tolerance. Hands and forearms need actual exposure. There is no indoor substitute.
• Tyre pressure and choice. Arguably the highest-leverage decision in gravel riding, and entirely absent from indoor training.
• Roadside repairs. Practise a tubeless plug and a tyre boot at home rather than in the rain, an hour from anywhere.

The example itself

Vermont and upstate New York are unusual in offering a dense network of both quiet paved roads and maintained dirt roads in the same area, which is why the region keeps coming up in this comparison. The dirt roads there are generally described as hardpack rather than loose gravel, which puts them nearer the better end of the rolling resistance range, and the terrain is consistently rolling rather than flat.

That is the limit of what I can tell you honestly. I have not ridden there, and the editorial policy on this site exists partly to stop me writing local knowledge I do not have.

The short version

• At 25 km/h and 85 kg, loose gravel costs roughly 69 W to asphalt's 23 W. The surface is one of the biggest terms in the equation, not a detail.
• Gravel also makes your power spiky, drops your cadence, and puts real work through your upper body.
• Train gravel with over-unders, low-cadence torque work in simulation mode, standing repeats and surge-y endurance rides.
• Train road with sustained blocks and long climbs ridden whole.
• Plan gravel at about two thirds the distance you would ride on tarmac for the same day.
• Air resistance overtakes rolling resistance at about 15 km/h on tarmac and about 24 km/h on gravel, which is why tyres matter more than position on gravel and the reverse on road.
• Handling, vibration and tyre choice do not transfer from indoors at all.