How the plan is built

Three things — where you're heading, where you're starting from and how much time you have — turn into a schedule for the weeks ahead. Below is how exactly: which decisions the platform makes for you, and why the plan holds so much easy work and so much planned rest.

The top level: where you are in the season#

The annual phase answers the question of what's going on right now. There are four: recovery after the season, general preparation, preparation for the main race, and the competition period. This is the top level of periodisation, and the platform picks the phase itself — from your races, with an explanation of the choice. There's no separate phase switch in the interface: to change the phase you change your race calendar, not the phase itself.

The logic is simple. While your next main race is still outside the preparation window, you're in general preparation; once it comes inside, preparation for that race begins. From the start of the taper to the end of the mandatory recovery is the competition period, and after the last main race of the season comes recovery. B and C races don't switch the phase on their own.

Blocks: base, build, taper, transition#

Within the year, training is laid out in four types of block. How many weeks each takes depends on the distance:

• sprint — 6 weeks of base, 4 of build, 2 of taper (12 in all);
• Olympic — 8 / 6 / 2 (16);
• 70.3 — 10 / 8 / 2 (20);
• full distance — 13 / 10 / 3 (26).

If there's less time, weeks are taken first from build, then proportionally from base, and the taper is never cut; the compressed minimums are 8, 10, 12 and 16 weeks. With less than that the platform doesn't promise full preparation: it builds a safe “get to the finish” plan and says plainly that the goal has been narrowed.

Why there's so much easy work in the plan#

The share of easy work is counted by time, not by number of sessions: in base, 80–90% of the time goes easy, in a build block 75–85%, and in the first two weeks, while the platform is calibrating, at least 90%.

That trained people do mostly easy work is a robust observation; that any one distribution beats another is not proven. So what we have is a range to steer within, not a dogma: moderate intensity isn't banned, and how much of it there is depends on the phase and the goal.

How fast you can add#

There are two limiters, and the stricter of them applies.

The first is on fitness: over seven days it doesn't grow faster than +3 for a beginner, +4 for an intermediate and +5 for an experienced athlete (the targets are more modest — +2, +3 and +4).

The second is on the week itself: planned load no higher than 1.15 of the reference week, with a target of 1.10. The same multiplier separately caps running minutes — the overall ceiling can be dodged by shuffling sports, and the risk observations come specifically from running volume. The reference week is the highest of the last four loading (not deload) weeks, and it's counted from what you did, not from what was prescribed — but never above what was prescribed. Otherwise a habit of doing slightly more than the plan would keep raising its own ceiling, week after week.

These are guardrails, not a prediction of injury: the numbers were deliberately set below the one risk boundary that has actually been observed — an increase in running volume of more than 30% in a week.

This is the very ceiling that stops you “catching up” after a holiday. There's nothing to catch up: that week has gone, and trying to win its volume back in one push is a reliable way to get injured on the way back.

The week in which you actually grow#

The intensity distribution is measured rather than assumed: the Fitness page shows what share of zoned time is easy work in this week's plan and in the last four weeks of fact, and which band applies in your phase (80–90% in base, 75–85% in a build block). If the week's plan falls below the band, the generator softens one hard session itself — it never adds intensity upward.

There is also a ceiling for a single session: no session runs more than 10% longer than your longest session of that sport in the last 30 days. One big run is riskier than a big week — among recreational runners it is the jump in the length of a single run that is most strongly linked to injury — so if you cut your long run to an hour, the plan will not prescribe ninety minutes but climb back step by step.

For a beginner, intensity arrives as a ladder rather than all at once: the first quality session is a short tempo touch, threshold work opens only after two completed tempo sessions, and the rung is set by what you have actually done, not by what was on the plan. That way, after months of easy work, your first “fast” session is never a threshold one.

Loading weeks alternate with deload weeks: two to one for a beginner, three to one for everyone else. In a deload week volume drops by 25–40% (usually by 30%), at least 80% of the sessions are kept, and no new intensity is added.

When a loading week sits on a deload position, the week label says why: the position was moved to next week because this one had already started, or skipped because last week itself ran at deload level and the recovery has happened. The label's expanded view has a section “What limited the week” — lines built from the codes the generator uses to explain its decisions: how many days you marked available, where the growth ceiling applied, why last week's plan was taken as the reference. Every code has a human-readable line — a code without one never reaches the product.

Growth happens not during the load but during the recovery from it: a session is a request, and the body answers it when it gets the chance. The coaching picture of this is called supercompensation — we have no separate number by that name, but planned rest is as much a part of training as the work is.

One exception: a deload is never placed immediately before a taper — otherwise you'd get three reduced weeks in a row and arrive at the start line undertrained.

The taper and your races#

A, B and C races differ in what gets built around them. A is the main one: the taper is done for its sake. B is important, but it only gets a full taper if that doesn't clash with the main race. C is a tune-up, slotted into an ordinary week with no taper at all.

The taper lasts two weeks, and three for the full distance. Volume comes down by 40–60%, while session frequency and short pieces of familiar intensity are kept: what has to go is the fatigue, not your form. This is the best-evidenced block in the methodology: the optimum was measured for a two-week window specifically, so we neither cut the taper short nor stretch it.

After a main race comes mandatory recovery: one week after a sprint or an Olympic, two after a 70.3, three after a full distance. Readiness can extend it — up to four weeks — but not shorten it.

Session time and travelling to the race#

A session card carries a time — when you usually train on that day of the week, based on your recorded sessions over the last weeks. It is a hint and an input for matching plan to fact, not an obligation: it does not affect days, duration or load. If a weekday has too few recorded sessions, there is no time.

If your target race is in another time zone — three hours' difference or more — a “Travelling to the race” block appears on the plan four weeks before it: the difference and its direction, a guideline for how many days early to arrive, and a reminder that the first day after the flight is easy work or rest only. Key sessions of the last three weeks take the race's start time when the catalogue has it, so the body gets used to working at that hour in advance. The platform prescribes no sleep or light schedules — that is a question for a doctor.

How much time you have#

Your hours budget is the frame all of this is laid out inside, and it comes from your weekly schedule: the time you have given to the days is your weekly volume. There's no separate “how many hours a week?” question sitting beside the schedule — one source of truth instead of two. If what you actually do keeps diverging from the schedule — three weeks out of four — the platform will suggest updating it, but it won't change it itself.

Then the daily schedule takes over. By default, weekdays allow up to 90 minutes of training, Saturday and Sunday up to 240, and the long day up to 360. A brick session is only placed where there's enough time for it, and both parts count together towards the day's ceiling. Strength work is spread across the season by its own rules — see Strength training.

If you have no races, and if you're over fifty#

A race isn't compulsory. With no future main race at all, the platform doesn't invent a goal: it builds general preparation — aerobic base, technique, strength. There's no separate “off-season” switch; the mode follows from your races.

Age on its own brings no penalty to your load, doesn't lower the growth ceilings and is not a medical restriction — there is no “age” reason in the Health section and there cannot be one. From 60 onwards the target growth runs along the lower edge of your experience band: the ceiling is the same, the climb is gentler. That is a coaching rule, not a doctor's ban. The spacing changes too: at least 48 hours between hard sessions that stress the same system, at least 72 after the age of 60, and from 50 onwards those 48 hours can't be shortened by any scheduling preference. Hard work goes more often to the bike and the pool, and running progression is handled more carefully. Intensity isn't taken away — it's surrounded by rest.

What happens when a week doesn't go to plan is in How the plan adapts to life.