ClimbWatts

An open training log · 98 → 86.75 kg · 228 W and climbing

A real training plan, published as it is ridden. Built around FTP and power zones, fitted around rotating shift work. Climbing watts, one week at a time.

FTP
228 W
Weight
86.75 kg
W/kg
2.63
To 77 kg
9.8 kg

Every training plan I could find assumes the same week, every week. Mine changes three times a month. This is how I made a structured, power-based cycling plan survive a rotating roster — the five rules it runs on, what I gave up to make it work, and how to build your own version in an evening.

I work in a power plant. Early shift, afternoon shift, night shift, in runs of three to five, and the pattern is handed to me months ahead but never repeats. I am 45. In January I weighed 98 kg. In July I rode 100 km in one go, badly, and decided to do it properly. Everything below is what I actually do — you can watch it happen on the plan, day by day, including the weeks that go wrong.

Why a normal training plan breaks on shifts

Open any 16-week plan and you'll find Tuesday intervals, Thursday threshold, Saturday long ride. That structure assumes three things a shift worker doesn't have: the same free hours each week, a stable sleep schedule, and recovery that resets overnight.

On a run of nights I sleep in the day, badly, and I lose two to three hours a night against my normal. By the third night the deficit is real and measurable — my resting heart rate is up, my HRV is down, and the same wattage feels ten percent harder. If the plan says "threshold intervals Thursday" and Thursday is my third night shift, I have three bad options: ride it and dig a hole, skip it and break the plan, or ride it badly and lie to myself about it. Most people pick the third, then quit in week six and conclude they have no discipline.

The problem was never discipline. The plan was wrong.

Rule 1 — the shift decides the session, not the calendar

I stopped thinking in weekdays. Every day in my plan is labelled by shift first, and the shift determines what is even possible:

Write those windows down before you write a single interval. If a session doesn't fit inside a real window, it isn't a session — it's a wish. My plan is generated with a hard limit per shift type, and a workout that exceeds its window fails to build at all. That one constraint removed more failed weeks than anything else I did.

Rule 2 — the morning numbers get a veto

Every morning, before anything, I check four things: Body Battery, hours slept, overnight HRV against my own 7-day average, and how I honestly feel on a 1–5 scale. Any one of these trips the switch:

When one trips, the hard session doesn't happen today. And during a run of nights there is a harder rule: under 5 hours of daytime sleep means no riding at all, not even easy. On four hours of sleep a Z2 ride isn't recovery, it's another stressor on a body that's already losing.

This felt like cheating when I started. It isn't. The numbers are not an excuse generator — they're a lie detector. On a good morning they tell me to go hard when I'd have talked myself into going easy, and that happens as often as the reverse. If you don't have a watch that gives you HRV, resting heart rate first thing in the morning does most of the same job for free.

Rule 3 — move the session, don't delete it

Here's where most adaptive plans get it wrong: they skip the workout and carry on as if nothing happened. Mine moves it. A hard session that gets vetoed goes to the next day that can actually take it, within four days — which in practice usually means the next afternoon-shift morning. If no such day exists inside the window, it's gone and I don't make it up.

The distinction matters because a training block is a dose, not a schedule. Three threshold sessions in nine days is the same stimulus whether they land on Tuesday, Thursday and Saturday or on Wednesday, Sunday and the following Tuesday. What ruins a block isn't moving work around — it's quietly doing less of it every week and not noticing.

Rule 4 — rate every session, and let it change the watts

After every ride I give it a score from 1 to 5. One means very easy, three means it was about right, five means it was too hard. Takes two seconds. Then the plan does something with it:

This is the closest thing to a coach that a spreadsheet can give you. An FTP test tells you where you were on one particular morning; the ratings tell you where you are this week, on this much sleep, after this many nights. On the plan you can see the corrections being applied — every rated session shows its score.

Rate honestly or don't bother. A plan fed flattering numbers gives back flattering numbers.

Rule 5 — the volume you can repeat beats the volume that looks good

Three weeks of rising load, one week of recovery, three times over, each block ending with a ramp test. Nothing original there. What is unusual is where the ceiling sits: I'm at roughly 6–7 hours a week, not 10–12, because 6–7 is what survives a month with two runs of nights in it.

Most of those hours are Z2 — unglamorous, conversational endurance. The hard work is sweet spot (88–90% of FTP) and threshold (95–98%), because they give the most fitness per unit of recovery, which is exactly the currency a shift worker is short of. VO₂max appears rarely, only on afternoon-shift mornings, and never in a week with nights in it. If you don't recognise those terms, I keep a plain-language glossary — nothing on this site needs more than that list.

What this costs

I want to be straight about the trade, because plans that promise everything are the ones people quit.

I train fewer hours than someone on a normal schedule. My hard weeks are less hard, because I won't stack intensity next to a run of nights. I lose four to six days of useful training a month to the roster, and no amount of cleverness gets them back. Progress is slower — my own honest projection is on the fitness page, and it flattens well before it gets anywhere impressive.

What I get instead is a plan that's still running in week 16. That turns out to be the whole game. Consistency at 80% beats brilliance that collapses in a month, and on a rotating roster brilliance always collapses.

How to build your own version tonight

  1. Write down your real windows. For each shift type, the hours you could actually be on the bike — after commute, after food, before sleep. Be pessimistic. That's your maximum session length.
  2. Get an FTP number. Ride a ramp test. Everything is a percentage of it, so one number keeps the whole plan honest, and re-testing every four to six weeks updates all of it at once.
  3. Put the hard sessions where the good window is. For me that's afternoon-shift mornings. Find yours and defend it.
  4. Fill everything else with Z2 and accept that it's meant to be boring.
  5. Pick your veto rules and write them down before you need them. Deciding at 05:00 on four hours' sleep whether today counts as "too tired" is a decision you'll get wrong. Decide it now, while you're rested.

If you want a starting point rather than a blank page, my whole plan is free: all 91 sessions as Zwift files, written in percentages of FTP, so they scale to your numbers without editing. No email, no sign-up.

One honest caveat

I'm a shift worker who got obsessive about this, not a coach and not a doctor. These are the rules I built for my own roster and my own body, published so that someone in the same situation doesn't start from zero. Sleep loss from night work is a genuine health issue, not just a training inconvenience — if you're struggling with it, that's a conversation for a doctor, not a training plan.

What I can offer is the evidence. Everything here is running live: the plan as written, the morning numbers behind every decision, and the fitness curve with the real test results plotted on it. In a few months it will be obvious whether any of this worked. If you also train around shifts, tell me what you do differently — I answer everything.