A static training plan is a document written before you start training. It assumes your fitness on day one, predicts your fitness progression week by week, and prescribes sessions based on those assumptions.
The problem is that training doesn't work that way. Life intervenes. Fitness doesn't progress linearly. The runner who arrives at week 11 is rarely the runner the plan was written for.
Adaptive training plans don't assume. They respond.
I learned this preparing for my second marathon. Week 11 of 18, a chest infection took me out for 12 days. I came back and forced myself through the prescribed sessions anyway. I arrived at the start line with legs that never properly recovered — and finished 18 minutes off my goal. The plan had no mechanism to account for what happened. That race is directly behind why Zarkus exists.
What makes a plan "adaptive"
The term gets used loosely, so it's worth being precise.
A plan is adaptive when it changes what you do next based on what you actually did — and what that did to your body. Not what you were supposed to do. Not what an average runner at your fitness level might have done. What you specifically did, and what the data shows about how you've responded to it.
This requires three inputs that static plans don't track:
1. Current fitness (CTL — Chronic Training Load) CTL is a rolling measure of your fitness built over the past 42 days. It rises when you train consistently and falls when you back off. It represents the aerobic capacity you've actually built, not the capacity the plan assumes you should have.
2. Current fatigue (ATL — Acute Training Load) ATL is a measure of your fatigue over the past 7 days. It responds quickly — a hard week spikes ATL significantly, a rest week drops it. High ATL means your body is currently carrying significant training stress.
3. Training form (TSB — Training Stress Balance) TSB = CTL − ATL. When TSB is negative, you're in a fatigue hole — building fitness, but not ready to perform. When TSB is positive, you've absorbed your training and are primed to race or do a key session.
An adaptive plan monitors these three numbers and adjusts what it prescribes based on them. This is the mechanical difference between adaptive and static training.
A concrete example: the illness week
This is the scenario that breaks most static plans.
You're on week 9 of an 18-week marathon plan. You get sick on a Tuesday — nothing serious, a bad cold — and miss five days of training. You return to running on Sunday, feeling okay but not 100%.
Static plan response: The plan doesn't know you were sick. Week 10 prescribes the same mileage and intensity that week 9 would have built toward. You face a choice: follow the plan and risk overloading a weakened system, or skip sessions and feel perpetually behind.
Adaptive plan response: Your CTL dropped during the illness week. ATL is low but your aerobic base has declined. The plan detects this divergence and rebuilds your base over the next 10–14 days before reintroducing intensity. The total training cycle may extend slightly, or the taper may be adjusted. You don't fall behind because the plan resets its trajectory to match where you actually are.
The difference isn't philosophical. It's physiological. Training on a body that hasn't recovered from illness produces stress without adaptation. The session exists in the log, but the fitness gain doesn't happen.
The three types of "adaptation"
When runners talk about adaptive training, they usually mean one of three things — and they're not equivalent:
| Type | What changes | Who controls it | How accurate |
|---|---|---|---|
| Manual adjustment | You decide to change sessions | You | Only as good as your self-knowledge |
| Coach-based adjustment | A human coach reviews your data and modifies the plan | Human coach | High, but expensive and requires consistent communication |
| Algorithm-based adaptation | Software recalculates the plan based on training data | System | High, consistent, available to all runners |
Most runners using static plans are doing informal manual adjustment — skipping a session when they feel awful, running shorter when legs feel heavy. This is better than blindly following a plan, but it's reactive and based on subjective feel rather than objective load data.
A human coach is the gold standard. The problem is access: good running coaches are expensive, and the relationship only works with consistent communication and detailed logging.
Algorithm-based adaptation makes the logic of a good coach available to self-coached runners. It doesn't replace the nuance a human coach brings, but for the vast majority of intermediate runners who can't afford or access a coach, it's a significant improvement over a static plan.
Who benefits most from adaptive training
Adaptive training has the highest impact for runners who fit one or more of these profiles:
The inconsistent scheduler You're training for a marathon, but work, family, or travel regularly disrupts your week. You might get 4 sessions one week and 1 session the next. A static plan can't survive this. An adaptive plan rebuilds its prescription around what actually happened.
The injury-prone runner Runners who have a history of overuse injuries are typically getting injured because training load spikes beyond what their tissues can adapt to. An adaptive plan that monitors ATL and prevents excessive week-over-week load increases is directly addressing the most common cause of overuse injury.
The runner returning from a break Whether it's injury, illness, life, or a planned off-season, returning to structured training with a plan calibrated to your current fitness — not the fitness you had before the break — is the difference between a smooth return and a relapse.
The intermediate runner without a coach This is the largest group. Runners who have been training for several years, have run a few marathons, and are trying to break a meaningful time barrier. Smart enough to follow structured training, but without the resources or access to work with a coach. Adaptive training is built for this profile.
What adaptive training doesn't do
It's worth being honest about the limitations.
Adaptive training cannot compensate for insufficient base fitness. If you're trying to run a 3:15 marathon on 40 km/week of base training, no algorithm changes that constraint. The plan can optimise what it works with. It can't manufacture fitness that doesn't exist.
Adaptive training also doesn't replace periodisation. A well-designed adaptive plan still moves through structured phases — base, build, peak, taper — and adjusts within that structure rather than abandoning it. Adaptation is about execution; periodisation is about strategy.
And adaptive training is only as good as its data inputs. If you're not logging your runs accurately, the algorithm is working from noise. Garbage in, garbage out.
The static plan vs. adaptive plan over a full cycle
Here's how the two approaches compare over an 18-week marathon build for a runner with VDOT 50 who misses 10 days due to illness in week 9:
| Static plan | Adaptive plan | |
|---|---|---|
| Week 9 prescribed load | Continues as planned | Reduces on illness detection |
| Week 10 entry fitness | Plan assumes CTL = 65 | Measures CTL = 52 |
| Intensity week 10 | T + I sessions as scheduled | Rebuilds aerobic base first |
| Race goal recalibration | Never — goal fixed at start | Ongoing — adjusts if fitness diverges |
| Taper timing | Fixed at 2 weeks from race | Calculated from actual CTL/ATL |
| Race day TSB | Unknown — often negative | Targeted: +10 to +20 |
| Injury risk | Elevated after illness return | Managed via ATL threshold |
The race day TSB difference alone is significant. Arriving at a marathon with TSB of −5 (still carrying fatigue) versus +15 (fresh, fitness preserved) can account for 10–20 minutes over 42 km.
The practical question: does it require more effort?
One legitimate concern: does tracking CTL, ATL, and TSB mean spending hours on training metrics?
In practice, no. The calculation happens automatically from your Garmin, Strava, or training log data. What changes is that the plan responds to that data rather than ignoring it.
For the runner, the experience is simpler than managing a static plan: you run the sessions the plan prescribes, log them, and the plan adjusts next week based on how things went. You don't need to understand the metrics to benefit from them.
The complexity moves from the runner to the system. That's the point.
The research behind CTL/ATL/TSB modelling in endurance sport was originally developed by Banister et al. (1975) and later popularised for running by Andrew Coggan. The underlying impulse-response model has been validated across cycling, triathlon, and distance running over four decades.
Zarkus is built on this model. Your VDOT from your most recent race sets your training paces. Your CTL, ATL, and TSB are tracked week by week. When something disrupts your training — or when things go better than expected — the plan recalculates and adjusts what comes next. If you're preparing for a race and you're done with static plans, join the waitlist below.
Frequently Asked Questions
What is an adaptive running training plan?
An adaptive running training plan changes what you do next based on what you actually did — and what that did to your body. It tracks metrics like CTL (Chronic Training Load, your fitness), ATL (Acute Training Load, your current fatigue), and TSB (Training Stress Balance, the difference between the two) to recalibrate sessions week by week. Unlike static plans that ignore missed training or illness, adaptive plans rebuild their prescription around where you actually are.
Who benefits most from adaptive marathon training?
Adaptive training has the highest impact for four runner profiles: inconsistent schedulers whose weeks vary due to work or travel; injury-prone runners who need ATL-managed load increases to avoid overuse; runners returning from injury, illness, or a break; and intermediate runners without access to a coach targeting a meaningful time goal. If you've run several marathons and are trying to break a barrier time without a coach, adaptive training is built for you.
How is an adaptive training plan different from just skipping sessions?
Skipping sessions is informal manual adjustment based on subjective feel. An adaptive plan uses objective training load data (CTL, ATL, TSB) to decide precisely how to modify the following sessions — not just whether to rest, but how much volume to rebuild, how quickly to reintroduce intensity, and whether the taper window needs adjusting. The difference is the precision of the response and the fact that it's data-driven, not guesswork.