The Science and Rationale

Most cyclists pedal at 80–100 rpm. Dropping to 50–70 rpm at the same power output forces each leg muscle contraction to produce much more force per stroke — that's what "high torque" means.

The same prime movers (quads, glutes, hamstrings) are active at all cadences, but at lower cadence each stroke preferentially recruits high-threshold (Type II) motor units that stay quieter at higher cadences [Ahlquist 1992; Sarre & Lepers 2006]. The hypothesis: combining this recruitment shift with high-intensity intervals drives stronger aerobic adaptations — particularly VO2max and maximal aerobic power — than the same intervals at freely chosen cadence.

The same mechanism that makes the stimulus effective also makes it demanding on connective tissue. More force per stroke means more load on knees, patellar and quadriceps tendons, and the muscles that stabilize the joint. Tendons and connective tissue adapt slower than the cardiovascular system, which is why this protocol begins with a 3-week adaptation phase at low cadence and easy power before any high-intensity work is layered on top.

The key study (Hebisz & Hebisz, 2024)

Two groups of well-trained female cyclists followed an identical 8-week polarized training program (~8 hours/week, 4-day microcycles). The only difference: one group did all high-intensity intervals at freely chosen cadence (>80 rpm), the other at low cadence (50–70 rpm).[Hebisz 2024]

Both programs included:

• Sprint Interval Training (SIT): 8–12 × 30-second all-out efforts
• High-Intensity Interval Training (HIIT): 4–6 × 4-minute efforts at 90–100% of max aerobic power
• Low-Intensity Endurance (LIT): long Zone 2 rides
• Active Recovery day

The low-cadence group also improved their second ventilatory threshold (VT2), which the free-cadence group did not.

Why it might work: At any given power output, pedaling slower means each pedal stroke requires more force. Research has shown that moving from the first threshold (VT1) up to VO2max effort nearly doubles the percentage of maximum dynamic force expressed on the pedals. Combining high force and high intensity likely drives greater muscle fiber recruitment — both slow-twitch fibers (already active) and additional fast-twitch fibers — leading to stronger aerobic adaptations.[Hebisz 2024]

Important caveat: The study tested a program where all high-intensity work was done at low cadence for 8 weeks. It did not test the use case of adding 1–2 low-cadence sessions to an otherwise normal-cadence training week. The ongoing training approach in this document is based on coaching practice, not on this study directly.

Coaching consensus on long-term integration

The study shows the mechanism works. The coaching sources below show how practitioners integrate it into ongoing training. This combination — study mechanism + coaching practice — is the basis for this document.

Neal Henderson (Apex Coaching, coached Rohan Dennis, Taylor Phinney): "Generally I do about one of these sessions a week, at most two." He has never programmed back-to-back big gear sessions and stresses the musculoskeletal load as the main limiter. Henderson's progression for new athletes: start with tempo (3–5 min at 75–85% FTP, 50–60 rpm), then threshold (5×5 min at ~95% FTP, 50–60 rpm), then VO2max (30–90s at ~5-min power).[Henderson]

EVOQ.BIKE: Start with 1 session per week, work up to 2 per week during base season. During race season: 1–2× per month for maintenance. They position low-cadence work as an ongoing training tool, not a block to complete.[EVOQ.BIKE]

Peter Schep / EF Pro Cycling: Progressive approach — start at 60 rpm, 80–85% FTP for 15 minutes, then advance to threshold efforts, and only then to VO2max torque work (e.g. 3×5 min at 106–120% FTP, 50–60 rpm — the Noemi Rüegg workout). Recreational riders should proceed cautiously. The VO2max torque workout is a post-adaptation session, not a starting point.[EF Pro]

Honest statement: No controlled trial has tested 1–2 low-cadence sessions per week mixed with normal-cadence training against a control group. The ongoing training recommendations in this document are coaching consensus, not experimentally validated. We are using the study to understand why low-cadence training may work, and coaching practice to guide how to integrate it.

Intensity floor: Zone 2 torque work is not effective

Both the research and coaching sources converge on this: low-cadence work must be at tempo intensity or above to be effective as a torque training stimulus.

• EVOQ.BIKE explicitly states: "If you go below tempo into zone 2, you probably won't be producing enough torque to make a meaningful difference." They recommend tempo or threshold power as the minimum.[EVOQ.BIKE]
• The null-result study (Muñoz et al., 2014) used moderate-intensity low-cadence work and found no benefit over freely chosen cadence. The positive results in Hebisz came specifically from combining low cadence with high intensity (90–100% Pmax for HIIT, maximal for SIT).[Muñoz 2014]
• Henderson: His entry-level torque work starts at 75–85% FTP (low tempo to sweet spot), not Zone 2.[Henderson]

This is the main practical concern. Low cadence cycling significantly increases patellofemoral (kneecap) joint load.

Biomechanical research shows that patellofemoral compressive force increases with knee flexion angle: at 15° flexion it is approximately 1× bodyweight, at 45° it rises to 3× bodyweight, and at 75° it reaches 6× bodyweight. Cycling seats you in a position of sustained knee flexion, and grinding at low cadence maximises force through each cycle.[Dye et al.]

Clinically: "Pushing hard gears at low revolutions puts a high load through the patella, whereas lower gears at high cadence (85–90 rpm) will put less load through the patellofemoral joint with each stroke."[Physio-pedia]

All low-cadence interval work must be done seated. Standing eliminates the training stimulus — you can use body weight to press the pedal down, bypassing the forced muscular contraction that makes this training work.

Who should avoid low-cadence training entirely

• Any history of knee overuse injury, patellar tendinopathy, or patellofemoral pain syndrome (PFPS)[EVOQ.BIKE]
• Coming off a rest period or injury
• Cyclists who already naturally grind at low cadences (the stimulus is reduced and the knee stress is higher)

Knee protection rules

These apply to every low-cadence session, always:

1. Always warm up at normal cadence first (minimum 15 minutes). Don't start a low-cadence interval cold.
2. If your knees ache during a set, stop the set. Don't push through. End the session if it continues.
3. Don't go below 50 rpm unless you have months of established low-cadence work behind you.
4. Never do low-cadence sessions on back-to-back days. The joint needs recovery time.
5. All intervals are seated. Standing removes the training stimulus and changes the load pattern.
6. If you have had patellofemoral pain, patellar tendinopathy, or any knee overuse injury: skip low-cadence training entirely.[EVOQ.BIKE][Physio-pedia]

Goal: Introduce the mechanical stimulus before adding intensity. Tendons and connective tissue adapt more slowly than muscles; this phase is non-negotiable.

The cadence, session structure, and progressive volume here is adapted from standard coach guidance on torque training progressions, not from the study (the study did not include an adaptation phase — participants were already highly trained).[EVOQ.BIKE]

Why Zone 2 here, despite the intensity floor? The adaptation phase is the only exception to the "tempo or above" rule. It uses Zone 2 power intentionally, to introduce the mechanical stress to joints and tendons before adding intensity. It is not meant to drive performance gains.

Frequency: 1 low-cadence session per week. All other sessions: normal cadence.

Ready to move on?

All of these must be true before starting ongoing training:

• All 3 adaptation sessions completed
• No knee pain during or after any session
• No lingering knee discomfort 24+ hours after any session
• Cadence targets felt achievable (not struggling to stay above target)
• RPE for the intervals was no higher than ~6/10 (it's Zone 2 power — the effort should come from the legs, not the lungs)

When to stop or go back

• Any knee pain during or after a session: Stop low-cadence work for at least a week. When you restart, go back to Week 1. If it recurs, see a physiotherapist before continuing.
• Can't maintain target cadence (grinding 5+ rpm below target): The gear is too hard. Reduce power target, not cadence.
• Any sharp pain anywhere: Stop immediately. This is not normal muscular discomfort.
• Feeling overly fatigued from the sessions: This is Zone 2 power — it should not be draining. If it is, your overall training load may be too high. Reduce other sessions that week.

How it works

You are not following a periodized block. You are adding a permanent training tool to your weekly routine. The biggest improvements will come in the first months (novel stimulus), but the sessions continue to provide value indefinitely — the same way that any structured interval work does.

Frequency

1–2 sessions per week. One is the norm, two is the ceiling.

• 1 per week is the standard. Some weeks you'll fit none. That's okay — this is a long-term practice, not a protocol with a deadline.
• Maximum: never more than 2 per week. Every coaching source converges on this ceiling (Henderson, EVOQ, EF Pro Cycling). The joint stress is the limiter, not the aerobic demand. When doing 2, make one harder and one easier.[Henderson][EVOQ.BIKE][EF Pro]
• Never on back-to-back days. At least one normal-cadence day between torque sessions.
• Race weeks: If you have a target Zwift race with hard sprints that week, consider dropping to 1 torque session and make it an easier one. Don't stack a Tier 4 sprint session and a race in the same week.
• Race season maintenance: 1–2 low-cadence sessions per month is sufficient to preserve the adaptation.[EVOQ.BIKE]

How to fit this into your existing ~10 hrs/week

• Replace 1–2 of your current interval sessions with a torque session from the calendar. Do not add torque sessions on top of your existing volume.
• Keep your long endurance ride at normal cadence.
• All other sessions (endurance, recovery, non-torque intervals): normal cadence.

The workout library

All sessions are seated throughout. All require a minimum 15-minute normal-cadence warm-up. The library is organized into four tiers by intensity and knee stress:

• Tier 1 — Entry: Henderson's tempo torque work and a short EVOQ staple. First weeks of ongoing training.
• Tier 2 — Development: EVOQ's staple format (5×5 and 5×8) and a scaled Hebisz HIIT intro. The bread-and-butter sessions.
• Tier 3 — Challenging: Henderson's threshold work, Hebisz HIIT at study intensity, and the Rüegg VO2max workout. Higher knee stress.
• Tier 4 — Advanced: EVOQ TorqueMax, Hebisz SIT and full HIIT volume. Highest knee stress, monthly at most initially.

Progression guidelines

There is no fixed schedule for moving between tiers. Progression is based on how your body responds, not on a calendar.

• Starting out (first ~4 weeks): Mix Tier 1 and Tier 2 sessions. Try each Tier 2 workout at least once to find what suits you.
• Advancing to Tier 3: After 3–4 weeks of Tier 2 with no knee issues and the sessions feeling manageable (RPE ≤ 7/10). Introduce one Tier 3 as your "hard" session; keep a Tier 1 or 2 as the "easy" one. Don't do two Tier 3 sessions in the same week initially.
• Advancing to Tier 4: After at least 8 weeks of consistent ongoing training (~11 weeks total including adaptation). Start with TorqueMax or the 2-set SIT. Monthly at most initially.
• FTP retesting: Retest every 6–8 weeks. As FTP increases, all percentage-based workouts automatically scale up in absolute power, maintaining the training stimulus.

Warning signs — when to back off

Honest answer: interesting, but not conclusive. Here's why:

Strengths

• Published in a peer-reviewed journal (PLOS One)
• Controlled design: same hours, same structure, same effort levels, only cadence differed
• Effect size is large and practically meaningful (+8.7% vs +4.6% VO2max)

Weaknesses

1. Tiny sample. 12 cyclists per group. The authors themselves state the findings are "not representative for the general group of training female cyclists."[Hebisz 2024]
2. Female-only, youth cohort. Participants were 17–20 year old competitive female cyclists. Results cannot be assumed to transfer directly to older, male, or recreational cyclists.
3. No injury monitoring. The study tracked performance but not joint stress, knee pain, or any injury risk markers. Low cadence is mechanically harder on the knees, and this was not studied.
4. Mixed literature overall. The authors acknowledge that "to date, there is no data reporting a clear and unequivocal benefit of torque (low cadence) training." Earlier studies using moderate-intensity low-cadence work with amateur middle-aged cyclists showed no advantage over freely chosen cadence — suggesting that intensity matters: the benefit may only emerge when combining low cadence with high intensity.[Hebisz 2024][Muñoz 2014]
5. Comparison to male cohorts. The same polarized model applied to well-trained men in previous research produced ~14% VO2max improvements overall — larger than either group here. Gender and training history appear to significantly affect the magnitude of response.[Hebisz 2024]

Confidence table

How solid each claim is and what would change it: