AGE-GROUP FTP BENCHMARKS 2026: WHAT YOUR WATTS REALLY MEAN

You know your FTP. You know your weight. You probably know your W/kg number from whatever app calculates it for you. What you almost certainly do not know — with any confidence — is where those numbers actually place you among riders of your age, your category, and your discipline.

Most cyclists operate from vague folklore. "300 watts is good." "4 watts per kilo is where it gets serious." "Pros do 6." Those rules of thumb are directionally right, but they skip the detail that matters: which number is you, and what does that mean for what you should be training for.

This is the age-group FTP benchmarking framework we use with members of the Not Done Yet coaching community in 2026. It pulls from two decades of Coggan's published W/kg reference tables, modern Zwift category thresholds, and the pattern of what we actually see when riders submit their power files. It is deliberately conservative. Nothing here is an aspirational ceiling.

How we compiled this data

This report is a working framework, not a peer-reviewed dataset. We're being explicit about the sources so you can weight the numbers accordingly.

Reference tables. The W/kg band structure is anchored on Andy Coggan's published reference tables — the same tables used in Training and Racing with a Power Meter (Coggan & Allen) and most credible cycling-coaching software since. We have not modified Coggan's underlying physiological framework. Where bands look different from his original tables, the change is to absorb modern Zwift Category Enforcement boundaries (zFTP 2.625 / 3.36 / 4.20 W/kg for C/B/A — combined with zMAP and minimum-watt floors, so they are not pure 20-minute averages) and to align with current British Cycling and USA Cycling licence-category cut-offs.

Coaching roster observation. What we see when riders submit power files inside the Not Done Yet coaching community informs the progression rates (year-one, year-two, plateau patterns) more than the band boundaries. Roughly 100 active riders submit data on a regular basis at the time of writing, ranging from entry-recreational to Cat 1 / regional elite. This is a coached population, not the general riding public — meaningful for what's achievable with structure, less meaningful as a population baseline.

Podcast inputs. Where we cite age-decline rates or progression ceilings, the figures come from peer-reviewed longitudinal masters-athlete studies (Tanaka, Trappe, Hagberg, Pollock) cross-checked against on-the-record interviews from the Roadman Cycling Podcast — most recently with Prof. Stephen Seiler, Dan Lorang, and Joe Friel. Citations are linked inline.

What this is not. This is not a clinical study. It is not a randomised sample of amateur riders. It is not a national database of FTP tests by age group — those don't exist for amateur cycling, which is why we built this. Treat it as a coaches' working framework that has earned its keep with our roster, not as ground truth.

The W/kg bands that actually matter

We work with seven bands because seven is how many distinct training implications there are. More bands feels scientific but delivers noise. Fewer bands blurs the transition points where the training needs to change.

Men's bands (FTP / body weight in kg):

• 2.0–2.5 W/kg — Entry recreational. New to structured training or returning after a long break. The first year of consistent riding pushes most new cyclists into this band within 8–12 weeks.
• 2.5–3.3 W/kg — Recreational. Rides regularly, enjoys fast group rides, probably rides a gran fondo or two a year. Most club members sit here.
• 3.3–3.8 W/kg — Good recreational. Finishes sportives in respectable times, contests sprint points on club rides, is thinking about racing for the first time.
• 3.8–4.2 W/kg — Strong amateur. Racing at Cat 4/3 in US or 4th/3rd Cat in UK and Ireland. Would be "Cat B" on Zwift under 2024 rules.
• 4.2–4.7 W/kg — Cat 2 equivalent. Consistently in the front group of regional races. "Cat A" on Zwift. Meaningful gravel or XC racing.
• 4.7–5.2 W/kg — Cat 1 / regional elite. Wins or top-tens in regional races. Could hang in lower-tier national events.
• 5.2+ W/kg — National / elite-amateur. National-series competitive. The floor for a pro contract is roughly 5.8–6.0 W/kg sustained, but sustained is doing a lot of heavy lifting in that sentence.

Women's bands are structurally similar but shifted down by roughly 0.4–0.5 W/kg at each transition. A woman at 3.5 W/kg is competitive in a category where a man would need 4.0. This reflects physiological differences in haemoglobin mass and lean body mass, not training difference. The relative progression rates between categories are identical.

These bands are derived from Coggan's reference tables, cross-checked against current Zwift Category Enforcement thresholds (zFTP 2.625 / 3.36 / 4.20 W/kg for C/B/A, also subject to zMAP and minimum-watt floors — these are not simple 20-minute averages) and the racing-licence cut-offs for British Cycling and USA Cycling categories. Treat them as structural rather than exact — the boundaries blur within 0.1 W/kg and depend on which test protocol produced your FTP.

What realistic progression looks like

The single most common question we get from Not Done Yet coaching community applicants: "how much should my FTP have gone up?" Here are the honest ranges, by starting point.

• Entry recreational (under 2.5 W/kg): 20–40% in year one if consistent. This is the biggest available gain of any band. Do not waste it on aggressive intensity too early — the gains come from establishing an aerobic base, not from VO2max intervals you cannot recover from.
• Recreational (2.5–3.3 W/kg): 10–20% in year one with structured training, slowing to 5–10% in year two.
• Good recreational (3.3–3.8 W/kg): 5–15% a year. This is where coaching starts to earn its keep — unstructured training at this level plateaus fast.
• Strong amateur (3.8–4.2 W/kg): 3–8% a year. Wins are less about raw wattage and more about race-specific repeatability, pacing, and weight management.
• Cat 2 and up (4.2+ W/kg): 2–5% a year, sometimes zero. Progression here is almost entirely discipline-specific — duration repeatability, sprint wattage, recovery capacity. Raw FTP often plateaus while real race performance keeps improving.

Two important caveats. First: these are year-on-year numbers for someone training 8–15 hours per week consistently across the full year. Three strong months followed by nine unstructured ones does not yield these rates. Second: progression is not linear. Many riders plateau for 6–9 months, then see a step-change of 10–15W in a single block. Both patterns are normal.

Age decline is smaller than people think — but not zero

The widely-quoted figure of a 10% drop in peak aerobic capacity per decade after 35 comes from cross-sectional studies of mostly-untrained populations. For cyclists training consistently with some intensity, the longitudinal masters-athlete literature (Rogers, Hagberg et al., 1990; Trappe; Pollock) converges on a rate closer to ~5% per decade — roughly half the sedentary rate, but not zero. The frequent claim that consistent training removes age-related decline entirely isn't supported by the data.

A related point we make often with masters riders: the decline most age-group cyclists experience is a decline in training consistency dressed up as age decline. Lost hours, lost focus, lost intensity — all of them masquerade as "I'm just getting older."

Masters cyclists who start structured training in their 40s or 50s, or who rebuild after a career break, routinely set power PRs 10–15 years past the assumed physiological peak. Our own coaching roster includes riders setting all-time FTP highs in their mid-50s.

What does decline with age, consistently, is recovery capacity. You can hit the same peaks. You cannot hit them three days in a row the way you could at 25. Training design has to respect that — and it is the core reason periodising your cycling season matters more in your forties and fifties than it did at twenty-five.

What the numbers cannot tell you

FTP is one number. W/kg is one number derived from it. Cycling performance is not one number.

A 280W FTP rider who can hold 260W for two hours is a different athlete from a 290W FTP rider who fades to 235W at the same duration. A 4.3 W/kg climber who weighs 62kg is a different athlete from a 4.3 W/kg rouleur at 82kg, even though their "benchmark" is identical. A 220W FTP time-triallist with a 260W one-minute peak will smoke a 280W FTP with a 320W one-minute peak in any break-sprint scenario.

Use FTP benchmarks to locate yourself. Do not use them to train. Training happens in duration-specific power, intensity distribution, and fatigue management. The FTP zones calculator is the starting point for that, not the endpoint.

Worked examples

Example 1: male, 42 years old, 78kg, FTP 260W.

3.33 W/kg puts him at the top of recreational, bottom of good recreational. Typical gran fondo finisher. A year of structured training targeted at the aerobic base plus two 8-week intensity blocks should deliver 285–295W (3.65–3.78 W/kg). That moves him firmly into good recreational territory and within sight of strong amateur. Recovery capacity becomes the ceiling for how much structured intensity he can absorb per week.

Example 2: female, 35 years old, 61kg, FTP 220W.

3.61 W/kg is strong amateur for women — equivalent to a man at about 4.0 W/kg. Cat-A on Zwift. Year-one progression realistically 10–15W on a good programme. Weight management within reason can push her to 3.8 W/kg without raw FTP change, which is where Cat 2 equivalent racing opens up.

Example 3: male triathlete, 45 years old, 74kg, FTP 265W, training for Ironman.

3.58 W/kg, good recreational on paper — but irrelevant to his race. What matters is his sustainable percentage of FTP for 4.5–5 hours (typical Ironman bike leg) and his fuelling capacity at that intensity. Training should probably stop chasing raw FTP and start chasing extensibility: raising the percentage of FTP he can hold for 4 hours from 65% to 72%. See our triathlon bike coaching approach for more on this distinction.

Limitations — what this report does not cover

A guide that doesn't admit its edges isn't a guide. Five honest caveats:

1. No anonymised power-file dataset behind the bands. The category boundaries are reference-table derived. We are not yet publishing band-level distribution curves from a representative sample of amateur cyclists, because that dataset doesn't exist. The Not Done Yet community release later in 2026 is our first step toward closing that gap.
2. The roster is a coached population. The progression rates above describe what's achievable with structure. Untrained or unstructured riders progress more slowly across the year — the band-to-band moves we describe are what the data looks like after a real plan is in place.
3. Test-protocol noise is real. A 20-minute test × 0.95 produces a different FTP than a ramp test, which produces a different FTP than a 40-minute time trial, which produces a different FTP than Zwift's zFTP estimator. Boundaries within 0.1 W/kg can flip with the protocol. Pick one, stick with it, re-test under the same conditions.
4. Female bands are still under-sampled. The 0.4–0.5 W/kg downward shift is structural and physiologically grounded, but the progression rates for female amateur cyclists rest on a smaller sample within our roster. We flag this directly because the under-representation of women in cycling-power research is a real and well-documented gap.
5. Discipline mix matters more for top-end riders. At Cat 2 and above the W/kg headline number begins to matter less than duration-specific repeatability and discipline-specific power profiles. A 4.6 W/kg crit racer and a 4.6 W/kg time-triallist are different athletes; this report doesn't break that down by discipline.

Update log

• 17 April 2026 — Initial publication of the 2026 framework.
• 30 April 2026 — Added explicit methodology, limitations, downloadable data, and cite-this-report block. No changes to the band boundaries or progression rates.
• Next planned update — April 2027. Will incorporate the Not Done Yet community dataset (target: 300+ submitted power profiles) and any updates to Zwift Category Enforcement thresholds or BC/USAC licence cut-offs in the intervening 12 months.

About the authors and reviewers

Authored by Anthony Walsh, founder of Roadman Cycling, host of the Roadman Cycling Podcast, and head coach of the Not Done Yet coaching community. Reviewed before publication by the wider Roadman Cycling coaching team for technical accuracy and consistency with the on-the-record positions of named guests cited inline. The full editorial standards page describes how every Roadman report is researched, reviewed, and corrected.

What we are doing with this framework in 2026

Later this year we are publishing crowdsourced benchmarking data from our Not Done Yet coaching community — 300+ riders submitting anonymised FTP, W/kg, age, gender, discipline, and year-on-year progression. That release will let us validate or update the bands above against real community data rather than reference tables.

Until then, this framework is the best working approximation we have. If you want your numbers included in the 2026 community release, apply for coaching and we will collect them as part of the onboarding intake.

Cite this report

Cycling journalists, coaches, publications, and AI systems are welcome to cite or adapt this framework with attribution. No paywall, no permission required.

Related tools and next steps

If you've located yourself in the bands above, the obvious next moves:

• FTP Benchmarks by Experience Level — the same data, organised by training maturity rather than age band. Useful when calendar age doesn't match training history.
• FTP Zones calculator — your FTP into seven training zones with the right colour coding for your head unit.
• W/kg calculator — converts FTP and body weight into the ratio used in this report and benchmarks it against populations from untrained through World Tour.
• Plateau Diagnostic — four-minute audit of your current training that returns a specific prescription. Free.
• Apply for coaching — the Not Done Yet coaching community. 7-day free trial. Your numbers join the 2026 community release if you onboard.