Dylan Johnson's “oscillation training” was an elite 2025 self-experiment, not a proven cycling training plan. He described alternating approximately 30–35-hour weeks with 10–15-hour weeks while preparing for Unbound and the Lifetime Grand Prix. In the same conversation, he said the exact sequence was not the gold standard and was not backed by good research.
Both parts belong in the answer. Publishing the volume without the caveat turns a transparent N=1 experiment into a prescription Johnson did not make.
This page owns Johnson-specific oscillation intent. Use the Friel/Lorang/Johnson comparison to compare planning approaches, the training-plan methodology guide for a general planning process and the training-plan service page for Roadman's current commercial offer.
Source boundary
- Dylan Johnson on the Roadman Cycling Podcast, published 2 April 2025. Establishes what Johnson said about his own 2025 calendar, volume alternation, intensity, monitoring, durability and uncertainty.
- 2019 block-periodisation systematic review. Establishes that block periodisation showed promising effects in some endurance variables while the included studies were small and generally low quality. It did not test Johnson's exact sequence.
- 2023 periodisation review in trained road cyclists. Establishes that only seven studies met its criteria and no evidence favoured one model across eight-to-twelve-week interventions.
The interview is primary evidence for Johnson's statements. It is not comparative evidence that his plan works better than another plan.
What Johnson described
Start with the event demand
Johnson said he normally prefers training to become more specific as racing approaches. For Unbound, he no longer saw the demand as simply holding tempo for ten hours; he described the race as increasingly punchy, with repeated high-power rollers layered onto the long duration.
That event analysis was the rationale for exploring concentrated volume while retaining race-relevant intensity. It does not establish that every gravel event or rider shares the same limiter.
Block volume rather than only intensity
Johnson contrasted common intensity-block examples with what he had been doing: concentrating volume. His early-2025 sequence alternated a very high-volume week of roughly 30–35 hours with a lower-volume week of roughly 10–15 hours, then repeated.
The numbers are descriptive. They belong to a professional athlete whose occupation, training history and events differ from those of most Roadman readers.
Intensity changed across the preparation
Johnson did not describe two intensity sessions as a timeless constant. He said January and February contained very little intensity when he was less fit. More recently, he was completing two standard high-intensity days in a high-volume week because he entered the week fresh enough to do so.
That phase context matters. The old article converted “recently” into “two structured intensity days stay in both weeks.” The transcript does not support that universal rule.
Monitor the hypothesis, not just the hours
Johnson discussed a durability test consisting of a twenty-minute effort early in a ride and another after several hours. He also reported some of his highest raw power numbers at that point, then noted that he was heavier and that watts per kilogram were similar.
Those observations helped him monitor his own training. They cannot isolate the effect of oscillation from body mass, prior training, measurement variation or other changes. No control condition existed, and the target season had not started.
Johnson's own caveat
When Anthony Walsh asked whether the high-week/low-week gap was the gold standard, Johnson's answer was explicit:
- it was not the gold standard;
- there was not good research on his exact oscillation theory;
- he did not know a study prescribing that repeated sequence;
- the Lifetime Grand Prix season had not begun;
- he did not yet know whether it would be a good or bad approach;
- he was willing to experiment on himself.
This is the central evidence label. “Early numbers look promising” is a pre-outcome observation, not a result.
What the block-periodisation evidence says
A 2019 systematic review and meta-analysis reported small favourable effects for block periodisation on VO2max and maximal power compared with traditional training, alongside promising results for some other variables. The authors also warned that the studies were small and generally low quality (Mølmen, Øfsteng and Rønnestad, PubMed 31802956).
That literature studied several block designs. Similar terminology does not make it evidence for alternating Johnson's exact 30–35-hour and 10–15-hour weeks.
A cyclist-specific 2023 review found traditional, block and day-by-day approaches could all improve included outcomes and found no evidence favouring one model across eight-to-twelve-week studies (Galán-Rioja et al., PubMed 36640771). The evidence supports cautious option-setting, not a universal hierarchy.
Periodisation order
Johnson referred to research comparing different orders of training and said he generally preferred standard periodisation because he wanted training to become more race-specific closer to competition. The old article reduced this to “order matters less than total intensity.” That is broader than the interview can establish without identifying and appraising the exact protocol, population and outcomes.
The safe conclusion is narrower: Johnson knew several models existed, preferred increasing specificity for his context and used a volume-block experiment within that preparation.
Sprint training in fatigued conditions
Johnson described sprinting as one of his weaknesses and discussed training some sprints after accumulated work because late-race gravel performance requires producing power under fatigue. That is a plausible event-specific choice and a report of his practice.
It is not evidence that fatigued sprinting is “most useful” for every cyclist or that fresh sprint work should be removed. Neuromuscular quality, technique, injury risk, target event and fatigue all change the decision. Readers should not convert one racer's priority into a universal sprint protocol.
Why there is no amateur oscillation template here
The previous version suggested alternating 14–18-hour weeks with 6–8-hour weeks and holding two intensity sessions through both. No source on the page validated those numbers. Scaling an elite plan by percentage does not preserve training history, recovery capacity, work stress, sleep, health or event demand.
For an amateur, a sudden high-volume week can also change fuelling, musculoskeletal load, sleep needs and exposure to illness or injury. “Lower than 35 hours” does not automatically mean appropriate.
What a self-coached rider can learn
The transferable skill is experimental discipline:
- State the event demand. Name the performance problem the change is meant to address.
- Describe the starting point. Use recent completed training, not an aspirational week.
- Change one major variable where possible. If volume, intensity and frequency all change together, attribution becomes difficult.
- Define monitoring. Choose process measures, subjective response and an outcome relevant to the hypothesis.
- Write stop conditions. Persistent pain, illness, unusual fatigue, poor sleep or repeated session failure should trigger review rather than stubborn completion.
- Set a review date. Do not call the experiment successful because of one early power file.
- Keep the evidence label. An N=1 result can guide that rider without becoming a universal recommendation.
Use the Roadman training-plan methodology guide to write the rider brief and review rules. The training-plan selection guide explains how to assess evidence and modification logic before buying or copying a plan.
Corrections from the previous version
The earlier article has been corrected in these areas:
- the headline and answer now identify a 2025 experiment rather than an established plan;
- Lorang's stale Bora role and second-hand calendar framing have been removed from Johnson's page;
- two intensity days are phase-specific context, not a constant rule;
- early power data is not presented as proof of success;
- no 14–18-hour/6–8-hour amateur template is prescribed;
- “principles translate directly” has been removed;
- two or three A races is Johnson's 2025 choice, not a standard every rider needs;
- fatigued sprinting is described as Johnson's event-specific practice, not universally superior;
- the exact oscillation pattern is separated from broader block-periodisation research.
Scope and safety
This article documents one athlete's reported experiment. It does not diagnose readiness or prescribe a high-volume block. Riders with persistent pain, illness, fainting, chest symptoms, unexplained performance loss or suspected low energy availability need suitable clinical assessment rather than a copied volume sequence.
Johnson supplied the right final word in the interview: he did not yet know whether the approach would be good or bad. A trustworthy knowledge page should preserve that uncertainty, not edit it away.