Skip to content
Coaching12 min read

THE 5 MISTAKES SELF-COACHED CYCLISTS KEEP MAKING (AND HOW TO FIX EACH ONE)

By Anthony Walsh
Share

Self-coaching can work. Some of the most thoughtful cyclists I know coach themselves successfully across years and continue to improve. But the population of self-coached cyclists who stall is much larger than the population who succeed, and the reasons they stall are remarkably consistent. The same five mistakes appear in nearly every plateaued self-coached cyclist I talk to.

When I did the 5 mistakes self-coached cyclists make episode, the diagnostic was structured around these five patterns. The detail varies by cyclist; the pattern doesn't. Here's what each one is, why it keeps happening, and the specific fix.

Mistake 1: AI coaching

This is the current version of the oldest problem in self-coaching. Cyclists have always reached for substitute coaches — books, magazine articles, internet forums, training apps. AI is the newest version of the same instinct, and it's the most seductive because the answers feel personalised and authoritative.

My position is direct. Don't use AI for coaching. Use it for narrow questions — interval format suggestions, fuelling calculations, exercise substitutions, post-session questions. Those it handles competently. Don't use it as a primary coach.

The reason is contextual judgment. Coaching isn't answering training questions. Coaching is reading between the lines of what an athlete reports, knowing when "tired" means recovery needed and when it means push through, understanding the life stress that's accumulating, noticing the pattern across weeks that the athlete hasn't noticed yet. The current AI tools don't have access to most of this context, and the context is the actual coaching.

The deeper issue is that AI tools respond to what you ask. The skilled coach responds to what you should be asking but aren't. The amateur cyclist who asks ChatGPT "should I do VO2max intervals today" gets an answer based on the question. The skilled coach asks "tell me about your sleep this week, the work stress, the last hard session," then prescribes based on the actual situation. The two outputs aren't comparable.

The fix. Use AI tools for narrow questions where the context isn't critical — interval prescriptions, exercise substitutions, calculation help. Get human coaching support, structured community programmes, or a clear self-coaching framework for the higher-order decisions about periodisation, recovery, and progression. The coaching pathways cover specific options; the Not Done Yet community is the community-based accountability route.

Mistake 2: No periodisation

The cyclist who runs the same week forever. Same Tuesday VO2max session, same Thursday threshold work, same Saturday long ride, same Sunday recovery spin. Held across years. Total volume might be 8 hours a week or 12 hours a week, but the structure doesn't change with the season, the events, or the rider's fitness phase.

This produces a predictable arc. First 6–12 months: meaningful FTP gains as the body adapts to the structured training. Next 6–12 months: gains slow. Year three onward: complete plateau. The cyclist trains as hard as ever, sometimes more, and the numbers don't move.

The mechanism is that all training adaptations have a ceiling determined by the stimulus you're applying. Threshold work develops threshold up to a point. Beyond that point, threshold work just maintains threshold without lifting the ceiling. The ceiling rises when the underlying systems rise — aerobic capacity, VO2max, neuromuscular function. Those systems need their own stimulus.

The periodisation guide covers the structure in detail. The headline: base phase builds the aerobic foundation, build phase develops specific intensity, peak phase sharpens, race phase performs, transition phase recovers. The phases rotate across a year. The stimulus changes across the year.

The fix. Pick one A-race per year. Build the calendar backwards using Friel's 14-week base, 10-week build, 3-week peak, 2-week taper structure. Identify which phase you're currently in. Adjust the training to match the phase. Hold the structure for 12 months before assessing. The first year of running real periodisation often surprises self-coached cyclists with how much improvement is available once the structure changes.

Mistake 3: Ignoring recovery

The training-focused mindset that produced the progress in the first 6–12 months becomes the limiter at year two. The cyclist who trained through fatigue successfully early on continues to train through fatigue when it stops working, and doesn't notice the pattern until the deep overtraining hole arrives.

The signals are visible if you're watching for them. Morning resting heart rate trending up. HRV 7-day rolling average trending down. Power numbers 5%+ off target on hard sessions. Sleep quality deteriorating. Mood and motivation declining. Food cravings increasing. Frequent minor illness.

Two or more of these together is the signal to back off. The self-coached cyclist who keeps pushing through these signals usually crashes into a 2–4 week forced recovery (illness, deep fatigue, injury) that could have been prevented by a 1-week voluntary recovery.

The deeper issue is identity. The cyclist who identifies as "the one who works hardest" finds it psychologically difficult to take a deliberate easy week. The recovery feels like weakness, like falling behind. The reality is the opposite — the recovery is what enables the next block of meaningful work.

The fix. Build recovery weeks into the periodisation structure. Every fourth week is reduced volume (40–50% of normal) and reduced intensity (no work above tempo). Non-negotiable. Track the recovery signals weekly: morning resting HR, HRV 7-day average, sleep quality. When two or more trend wrong, take a recovery week now rather than waiting for a forced recovery. The recovery for cyclists guide covers the protocols in detail.

Mistake 4: Tracking the wrong metrics

The cyclist who obsesses over TSS. Or daily HRV readings. Or weekly weight checks. The metrics are real and quantifiable, but the decisions they drive are often wrong.

TSS obsession. Training Stress Score is a quantification of training load. Useful for general trends. Often misused as a target. The cyclist who chases 600 TSS per week loses the ability to make session-specific quality decisions — they end up doing a moderate-intensity session because the TSS math works, rather than the right intensity for their current phase. The pros mostly don't watch TSS daily; they execute the session and let TSS be a downstream measurement, not a target.

Daily HRV reactions. Single-day HRV readings are mostly noise. The cyclist who modifies training based on today's HRV often overreacts to normal day-to-day variation. The recovery guide covers the right framework — the 7-day rolling average is the signal, not the daily number.

Weekly weight checks. Daily weight has too much noise (hydration, glycogen, food timing). Weekly weight has slightly less noise but still drives short-term reactions. Monthly body composition trends are the working metric. The cyclist who weighs themselves daily and reacts to short-term changes typically eats badly.

Strava chasing. Segments and KOMs as training targets distort the structure. The threshold ride that becomes a segment hunt becomes a different physiological session. Strava as a record-keeper is fine; Strava as a coach is destructive.

The fix. Track slower-moving metrics. Weekly total training hours and key session execution (did you hit the target intensities?). Monthly rolling 7-day HRV average, body composition direction, sleep quality trend. Every 8–12 weeks FTP test on consistent protocol. The slower the metric moves, the more decision-useful it is. Daily fluctuations are mostly noise.

Mistake 5: No accountability

The self-coached cyclist who executes the plan when no one is watching is rarer than the self-coached cyclist who thinks they execute the plan. The honest assessment is hard. Did you actually do all five sessions this week, or did you skip Wednesday's easy ride because the kids needed something? Did you hold the easy days at proper Zone 2 intensity, or did the Saturday ride accidentally become a moderate-intensity 3-hour effort?

Without external accountability, the gap between intended training and actual training tends to grow. Not dramatically week by week. Slowly, across months. The accumulated drift between plan and execution is one of the dominant reasons self-coached training stalls.

The feedback loop matters here. A coach who sees your weekly numbers, asks about the sessions, notices the pattern — that's the feedback loop. Without it, the cyclist relies on self-assessment, which is biased. Most of us think we worked harder than we actually did, completed more sessions than we actually completed, and held the structure better than we actually held it.

The fix. Get external accountability somewhere in the system. Three workable options:

  • Coach. Direct, expensive, highest accountability. The coaching pathways cover specific options at different price points.

  • Community programme. The Not Done Yet community at $195 per month includes weekly coaching calls, training plan templates, and the kind of peer accountability that holds the structure even when motivation is variable.

  • Training partner. A specific person who knows your plan, asks about your sessions, holds you to the structure. Lowest cost, highest dependence on finding the right person.

The cyclist who tries to self-coach without any external accountability does so against the grain of how behaviour actually works. Some succeed; most don't. The honest assessment is whether your current approach has worked for the previous 12 months. If not, accountability is likely the missing piece.

What Dan Lorang says amateurs consistently get wrong

When Dan Lorang was on the podcast — covered in 13 years of coaching pros — his framing on amateur mistakes was clean. Three patterns repeat almost universally across the amateurs he sees.

Too many hard sessions. The amateur who runs 3–4 hard sessions a week with the recovery capacity of someone running 2 hard sessions a week. The accumulated fatigue produces stalled adaptation. The fix is reducing the number of hard sessions and protecting the recovery between them.

Insufficient easy volume. Amateurs underweight the easy riding because it doesn't feel like training. The pros run enormous volumes of easy work; amateurs at lower volumes still need the same intensity distribution. The fix is committing to genuine Zone 2 work as a meaningful training stimulus.

Plan fragmentation. Amateurs treat every interesting event as worth peaking for. The result is fragmented periodisation, multiple half-peaks across a year, and no real A-race. The fix is selecting one A-race per year and building the calendar around it.

Lorang's overall point: the structure that produces World Tour results — when scaled — is what produces serious amateur results. The amateur mistakes aren't specific to amateurs; they're the universal mistakes that pros avoid because they have coaches preventing them.

When to get a coach

The honest cost-benefit. Coaching has costs (financial, time for sessions, occasional friction over disagreements) and benefits (faster progress, less wasted training, structured accountability). The break-even depends on your situation.

Beginning cyclists (first 1–2 years). Coaching usually accelerates progress meaningfully. The structured introduction to training principles, plus the accountability, plus the avoidance of the common mistakes — all combine to compress what otherwise takes 2–3 years of self-experimentation into one year.

Plateaued intermediate cyclists. Coaching has the highest leverage here. The cyclist who's been at the same FTP for 18 months can usually identify one of the five mistakes above as the limiter, and a coach (or structured community programme) accelerates the fix.

Experienced self-coached cyclists making consistent gains. Coaching has lower leverage. If your self-coached approach is working, the marginal benefit of coaching is small. The honest test is whether you're still making gains. If yes, continue. If not, the value of external input rises.

Masters cyclists. Coaching value rises with age because the margin for error narrows. The 55-year-old who runs aggressive intervals on a fatigued body pays a higher cost than the 35-year-old running the same session. The recovery profile of masters athletes makes the prescription stakes higher, and external support pays back.

The Not Done Yet community is the lower-investment option ($195/month) that captures most of the coaching benefit through structured group programmes and weekly calls. The Roadman Method is the $297-397/month structured 12-month programme with personal coaching support. The direct coaching pathways are the higher-touch options.

The accountability alternative: community as coaching

The pattern I've watched develop in the Not Done Yet community is that community-based accountability captures much of the coaching benefit at lower cost. The structure that works:

  • Weekly live coaching call where members bring specific questions
  • Shared training plan templates the group runs together
  • Peer accountability through structured check-ins
  • Direct access to me and the coaching team for specific situations

The members who use this well execute their training more consistently than they would on their own, get their specific limiter questions answered, and benefit from watching other serious cyclists working through similar challenges. The cost is much lower than individual coaching; the accountability is real.

What to do next

Honest self-assessment first. Run through the five mistakes:

  1. Are you relying on AI for coaching decisions you should be making with human input?
  2. Do you have a real periodisation structure or are you running the same week forever?
  3. Are you taking deliberate recovery weeks, or accumulating fatigue until it breaks you?
  4. Are you tracking metrics that drive useful decisions, or daily noise that drives overreactions?
  5. Do you have any external accountability, or are you self-assessing your way to plateau?

The Plateau Diagnostic walks through the assessment systematically in four minutes. Free. Returns the one change most likely to move your numbers. If two or more of the five mistakes show up, the structure needs revision before any specific training change will help.

For external support, the Not Done Yet community at $195/month is the most common entry point for self-coached cyclists who need accountability — weekly coaching calls and the peer-led check-ins that catch drift before it becomes a plateau. The coaching pathways cover higher-touch options. The Roadman Method at $297-397/month is the structured 12-month programme for cyclists ready to commit to a full periodisation cycle with personal support.

Self-coaching can work. The cyclist who runs it well sometimes outperforms cyclists with coaches, because the self-directed motivation is real and the alignment with their own life is tight. The cyclist who runs it badly stalls for years and doesn't understand why. The five mistakes are the most common reasons. Pick the one that hits closest to home, fix it, and you'll often see more progress in the next 12 weeks than you did in the previous 12 months.

FAQ

FREQUENTLY ASKED QUESTIONS

Can cyclists really coach themselves effectively?
Yes, with structure and honest self-assessment. The cyclists who self-coach successfully are typically experienced, have strong analytical skills, and know their own bodies well. Beginning cyclists and those with deep plateaus usually benefit more from coaching support. The honest assessment is whether your current approach is producing the gains you want. If not, the question is whether you can identify and fix the specific limiter or whether external coaching support would accelerate the process.
Should I use AI to coach my cycling training?
Not as a primary coach. AI tools can answer narrow questions (interval format suggestions, recovery timing, nutrition calculations) competently. They can't read your sleep, life stress, family situation, recent illness, or the way you describe a session in nuanced terms. Coaching involves contextual judgment that the current AI tools don't have access to.
How do I periodise my training if I don't have a coach?
Pick one A-race per year and use Joe Friel's structure as a framework: 14 weeks base, 10 weeks build, 3 weeks peak, 2 weeks taper. Add a recovery week every fourth week within each phase. The detailed framework is in our periodisation guide. The structure is the framework; the execution is the work.
What metrics should I actually track?
Weekly: total training hours, key session execution (did you hit the target intensities?), morning resting heart rate trend. Monthly: rolling 7-day HRV average, body composition direction, sleep quality trend. Every 8–12 weeks: FTP test on a consistent protocol. Avoid daily HRV reactions, daily weight checks, and TSS obsession — these drive overreactions.
How do I get accountability without a coach?
Community-based accountability through structured groups like Not Done Yet provides much of the benefit at lower cost. Training partners who hold you to plans work for some cyclists. Online communities provide some accountability but less than direct coaching. The honest assessment is whether you actually execute your plan when no one is watching — if not, external accountability is the missing piece.

KEEP READING — THE SATURDAY SPIN

The week's training takeaways, pro insights, and what to do about them. 65,000+ serious cyclists open it every Saturday.

AW

ANTHONY WALSH

Host of the Roadman Cycling Podcast

Share

RELATED PODCAST EPISODES

Hear the conversations behind this article.