Periodisation without the textbook.

The textbook doesn't know you.

If you've ever read a coaching manual or taken a coaching course, you've been taught periodisation. Base, build, peak, race. The classic model. Spend weeks building your aerobic foundation, then add intensity, then sharpen, then perform. It's clean. It's logical. And it's been the organising principle of endurance training for decades.

It's also, for most of the people watching this, not how training actually works.

Here's why. That model is built on a theory from the 1930s called the General Adaptation Syndrome, GAS. It was developed by a guy called Hans Selye who was studying stress in rats. Not athletes. Not cyclists. Rats. And the idea is simple: apply stress, the body adapts, repeat at a higher level. If you get the dose right, you get fitter. If you get it wrong, you overtrain.

And on paper, it makes sense. The problem is that the science underneath it has, to put it politely, not held up. Kiely reviewed this in depth and his conclusion was blunt. The original scientific platform upon which periodisation theory was founded has, his words, disintegrated.

Now I want to be careful here, because I'm not saying periodisation is useless. It's not. Having a structure, having phases, having progression, all of that matters. What I'm saying is that the idea that there's one correct way to sequence training, one model that predicts how your body will respond, one textbook answer that works for everyone, that's the part that doesn't hold up.

Why I disagree with GAS.

I've had this argument with other coaches plenty of times, so let me lay out exactly what I think and why.

GAS assumes a predictable stress-response curve. You apply a training stress, you recover, you adapt, you come back stronger. And if that were reliable, coaching would be easy. You'd just calculate the dose, apply it, wait, and watch the numbers go up.

But that's not what happens. Not with real people, not in real life.

The response to training is individual. Wildly individual. Tucker and Collins illustrated this in 2012 in the British Journal of Sports Medicine. They sketched out six hypothetical athletes, each with a different genetic ceiling, and showed that the same training input lands very differently depending on what an individual's body can absorb. For two of those six, training alone never produces an elite outcome because the genetic potential isn't there. For the others, the same programme produces wildly different realised performance levels. I extend that framework to include recovery capacity and life load, but the point is the same. Same input, different outputs.

And it's not just genetics. It's life. It's the body, brain, life model I talked about in the durability chapter. A rider who's sleeping eight hours and has low stress will respond to the same training block completely differently than a rider who's sleeping six hours and managing a difficult project at work. The training plan can be identical. The adaptation will not be.

So I made a deliberate choice, early on in my coaching. I stopped writing plans that assume a predictable response, and I started building systems that respond to what's actually happening.

What I do instead: signal-based training.

The shift: instead of "follow this plan for eight weeks," I think in terms of signal and response. The training provides a signal. The body responds, or it doesn't. And the next decision comes from reading that response, not from looking at the calendar.

What this looks like in practice: you start a block with an intent, build aerobic capacity, develop threshold power, whatever the goal is. You apply the training. And then you watch. Is the heart rate responding? Is the power shifting? Is the RPE matching? Are the durability markers improving? If yes, keep going. If not, something needs to change.

When to stop a block: not because it's been six weeks. Because the signal says you've got what you came for, or because the signal says you're digging a hole. Both are valid reasons to change direction.

The boring block is the best example of this. When I prescribe a boring block, a period of focused Zone 2 volume, I don't say "do this for eight weeks." I say "do this until we see the adaptation we're looking for." Sometimes that's six weeks. Sometimes it's ten. The block length is an output, not an input.

The apply / observe framework.

So if I'm not following the base-build-peak model, what am I doing? Here's the framework I actually use.

I think in two phases that repeat. Apply, observe.

Apply is the training stimulus. Whatever you're trying to develop, aerobic base, threshold power, durability, sprint capacity, you apply the work. Consistently, with intent, with appropriate load.

Observe is the most underrated phase, and it's bigger than people think. This is where you watch what happens. Not just "am I tired" but "what's changing?" Is my decoupling improving? Is my power at hour three holding better? Is my RPE at the same intensity dropping? You're reading the response to the signal. And observe isn't just a beat between blocks. It includes the easy days, the rest days, the recovery weeks where the body actually consolidates the work. The adaptations don't happen during the apply. They happen during the observe. The length of that phase depends on the individual. Some riders need three days to come back, some need ten.

And then you go again. Apply the next signal. Observe the response. This is iterative, not linear. You're not marching from A to B on a predetermined path. You're navigating.

Bottom-up and top-down.

There's one more piece to how I think about this, and it ties the whole approach together.

I think about training in two directions. Bottom-up and top-down.

Bottom-up means you're building from the rider's physiology. You're looking at who this person actually is, their strengths, their limiters, their recovery capacity, their durability profile, and you're building training that matches them. Not a template. Not what worked for someone else. Training that addresses what this specific body needs right now.

Top-down means you're working backwards from the demands of the event. What does the race ask for? How long is it? Where are the decisive moments? What power, what duration, what fuelling does it require? And you're shaping the training to match those demands.

Here's the key. You start bottom-up and you finish top-down.

Early in the season, the boring block, the base work, the aerobic development, that's all bottom-up. You're building the rider. You're not thinking about the race yet. You're thinking about the engine, the durability, the platform.

As you get closer to the event, you start flipping the direction. The training becomes more specific. You're simulating race demands. You're practising the effort profile, the fuelling strategy, the pacing. That's top-down. You're no longer asking "what does this body need?" You're asking "what does this race need, and can this body deliver it?"

And the taper or peak period is pure top-down. Everything is shaped by the event. The volume drops. The specificity sharpens. You're not building anymore, you're revealing what you've built.

The mistake most people make is starting top-down too early. They jump straight to race-specific work before the platform is there. And it works for a few weeks, the numbers look good, the sessions feel sharp, but it doesn't last. Because you're shaping something that hasn't been built yet.

Bottom-up first. Top-down to finish. That's the direction of the whole year.

Why this matters in the real world.

You can't afford a bad block. If you've got 7 hours a week, an eight-week block that isn't working is a quarter of your training year wasted. Signal-based training catches that earlier. If the adaptation isn't happening by week three, you adjust. You don't grind it out because the spreadsheet says so.

Life disrupts plans. You miss a key session because your kid is sick. You travel for work. You have a week where sleep goes to hell. A textbook plan doesn't account for this. A signal-based approach does, because the next decision is always based on where you are now, not where the plan assumed you'd be.

Periodisation for the real world is responsive, not prescriptive. I still plan ahead. I still think in blocks and phases. I still have a seasonal structure. But I hold it loosely. The plan is a hypothesis. The data tells me whether the hypothesis is right.

Peter Leo said something I agree with: "I don't believe in any training method." And what he means isn't that methods don't matter, it's that no single method works for everyone. The method is secondary to the application. How you respond to the training is more important than which model you followed.

The experiment of one.

I'll leave you with this. Every single rider I coach is an experiment of one. I've never coached two people the same way, because no two people are the same. Different genetics, different lives, different stress loads, different recovery capacities, different goals, different willingness to suffer.

The textbook gives you a template. And templates are useful as a starting point. But the moment you treat the template as the answer, the moment you stop reading the signals and start following the calendar, you've stopped coaching and started administering a plan.

I'd rather be wrong and adjust than right by accident. Because the adjusting is where the learning happens. And that's what the next few chapters are about. The specific tools for applying this: how to build your engine, which workouts earn their place, and how to execute on the day.