Training Zones & the Mader Model: Beyond LT1/LT2 (2026)
Open any training app and it will hand you five or six tidy zones. What it rarely tells you is where those lines came from — and that almost every one of them is drawn off a lactate curve someone fitted to a step test, using a definition the field has argued about for fifty years. A Faster You does it differently: we measure your two metabolic engines and let a physiological model draw the zones. Here is exactly how, and why it matters for every session you ride or run.
The quiet problem with LT1 and LT2
Most platforms anchor your zones on two points: LT1 (the aerobic threshold) and LT2 (the anaerobic or lactate threshold). They sound definitive. They are not.
LT1 and LT2 are points read off a fitted lactate-step-test curve — descriptive markers, not mechanisms. And there is no single agreed way to find them. Depending on whose method your lab or app uses, "your" threshold moves: the Dmax method, a fixed +0.5 or +1.0 mmol/L rise, a log-log breakpoint, a fixed 4 mmol/L line, a proprietary baseline-plus-offset method. These can disagree by several percent of intensity on the same test — and a comprehensive review of the competing concepts concluded exactly that the definitions are inconsistent and their validity is debatable (Faude, Kindermann & Meyer 2009).
So when an app tells you your threshold sits at a specific number, the honest follow-up question is: by whose definition? Build your whole training year on a line that arbitrary, and the zones inherit the arbitrariness.
Your body has two engines, not one number
Here is the deeper issue. A single threshold number tries to compress your physiology into one point. But endurance performance runs on two largely independent engines:
- VO2max — your aerobic capacity, the ceiling on how much oxygen you can use per minute.
- VLamax — your maximum glycolytic rate, how fast your muscles produce lactate.
Two athletes with the identical threshold power can have completely different engines underneath — a high-VO2max diesel versus a high-VLamax sprinter — and they need different training. A one-number model can't see that difference. We cover the second engine in depth in
, and how the two interact in VLamax vs VO2max.
This is the core of measured, not guessed: we don't reduce you to one line on a curve. We measure both engines and work from them.
What we do instead: simulate the model, don't fit the curve
Once we have your two measured numbers, we don't look for breakpoints on a graph. Instead of fitting a lactate curve, we run a dynamic simulation of the Mader model (Mader & Heck 1986; Mader 2003). Your muscles make energy in two ways. Glycolysis burns carbohydrate for fast, hard efforts. Oxidative phosphorylation is your aerobic system — it burns fuel with oxygen for sustainable efforts. The model works out how those two balance against the power you put out. Driving that balance are your VO2max and your VLamax.
From that simulation we read physiologically defined anchors directly off the model, rather than back-reading them from a fitted test curve:
- FatMax — the intensity of maximal fat oxidation. This is our base/aerobic anchor: it plays the role LT1 plays in the old framework. FatMax and LT1 correlate and serve the same purpose. They are not the identical point, and we don't pretend they're interchangeable. The widely taught principle of keeping easy work below the first threshold maps cleanly onto keep your base around or below FatMax — the foundation we build in base training.
- MLSS — the maximal lactate steady state, the highest intensity at which lactate production still equals clearance. In our model it is deterministic, computed from your two engines rather than estimated from a curve — the mechanistic equivalent of "LT2" (Billat et al. 2003).
- CP — critical power, the sustainable-power fatigue threshold that separates steady from unsustainable efforts (Poole et al. 2016).
- VO2max — the ceiling above it all, the value you can benchmark against age and sex on our VO2max chart.
So we are not "missing" LT1 and LT2. We provide the mechanistically grounded equivalents — derived from a model of why your body behaves as it does, not from where a line happened to bend on one afternoon's test.
Why a model beats a fitted curve
The difference is the difference between description and mechanism.
A fitted lactate curve describes what your blood lactate did on the day of the test, then asks a human (or an algorithm) to pick where the meaningful points are — by a definition chosen from several that disagree. A model simulation starts from the two physiological quantities that actually cause the curve — your aerobic capacity and your glycolytic rate — and computes the anchors from first principles. Same physiology, opposite direction: we reason from cause to zones, not from a single curve back to a guessed cause.
That is what lets us put real numbers on your screen without a lactate lab, and re-compute them the moment either engine changes. It is also why two athletes who would land on the same "threshold" in a step test get genuinely different zones from us — because their engines differ, and the model knows it.
What is the science, and what is our own work
We are careful to draw this line honestly, because trust depends on it.
The Mader metabolic model is the published scientific framework — Mader and Heck's theory of the metabolic origin of the anaerobic threshold (1986) and Mader's muscle-cell model of glycolysis and oxidative phosphorylation versus power output (2003). MLSS and critical power are established concepts in the literature (Billat et al. 2003; Poole et al. 2016). You can read every one of those papers.
Estimating VLamax from a short maximal sprint is published laboratory science. It has been measured and validated in peer-reviewed research for years (Adam et al. 2015; Nitzsche et al. 2018; Hauser et al. 2014). It is not our invention, and we don't claim it as one.
Our own work sits on top of that. We run the measurement without a lab and without blood draws. We calibrate it with a neural network against roughly 15,000 real Powertests. That is exactly where the published sprint method stays measurement-sensitive on its own (Harnish et al. 2023). We then combine your VO2max and your VLamax into individual zones. On top of that sits our body-reserve and protein-turnover modeling, which schedules your training.
So when someone asks "which paper says that?", the honest answer is simple. The physiology and the VLamax method belong to the field. The lab-free, calibrated, personalized application is ours.
How this compares to the alternatives
To be fair to the field:
- Other metabolic-model tools — some platforms also run a metabolic model, which is genuinely rigorous; the honest difference is the route to the anchor: a curve-derived offset versus our anchors read directly off the model from your own measured data.
- TrainingPeaks, and most %-of-threshold or %-FTP systems, teach zones as fractions of a single threshold or FTP number. Practical and popular — but they're population averages applied to your one number, with no individual VLamax in the math.
Our angle isn't that everyone else is wrong. It's that the same rigor can be run on your own measured data, explained transparently — and that a two-engine model tells you something a one-number model structurally cannot.
How you get your own model-derived zones
The on-ramp is the Powertest: a standardized effort — a ramp, a maximal 12-minute effort, and a 15-second sprint — recorded on your own bike or trainer (or, for running, your GPS watch). No lab, no mask, no needles. From it we measure your VO2max and VLamax, simulate the Mader model, and hand back FatMax, MLSS, CP and VO2max as concrete training zones. The full walkthrough is in getting going with your Powertest.
Between tests, connect your Garmin, Wahoo, or Strava and our AI keeps a running estimate of your zones from every session — accurate enough to train on, and re-anchored each time you re-test. Measured baseline, live trend.
Run a Powertest and start your free trial → — get your two engines, get model-derived zones, and stop training on a borrowed line.
Where the zones are pointing: up
One thing the model never does is ask you to shrink. Your zones exist to serve a rising VO2max: base work around FatMax builds the aerobic engine that everything else sits on; threshold and MLSS work raise the sustainable ceiling; the sharp end lifts VO2max itself. Running economy and technique complement that climb — they help you spend your engine more efficiently — but they never replace building a bigger one. The whole system is built to move your ceiling up over the long term, then redraw your zones around the fitter you. That long-term climb is the entire reason we measure the engine in the first place.
FAQ
Why doesn't A Faster You use LT1 and LT2? Because LT1 and LT2 are points read off a fitted lactate-step-test curve, and the field uses several competing, somewhat arbitrary definitions to place them (Dmax, +0.5/1.0 mmol, log-log, fixed 4 mmol/L, baseline + 0.45 mM) that can disagree on the same test (Faude et al. 2009). Instead we simulate the Mader model from your measured VO2max and VLamax and read physiologically defined anchors — FatMax and MLSS — directly. We're not missing LT1/LT2; we provide the mechanistic equivalents.
What is FatMax, and is it the same as LT1? FatMax is the intensity at which you burn the most fat per minute, and we use it as your base/aerobic anchor — the role LT1 plays. FatMax and LT1 correlate and serve the same purpose, but they are not the identical point, so we don't treat them as one measurement.
What is MLSS and why use it instead of "lactate threshold"? MLSS — maximal lactate steady state — is the highest intensity at which lactate production still equals clearance. It's a physiologically defined steady state, and in our model it's computed deterministically from your two engines rather than estimated from a curve, which makes it the mechanistic equivalent of "LT2" (Billat et al. 2003).
Do I need a lactate test or a lab to get these zones? No. The Powertest measures your VO2max and VLamax from a standardized effort on your own equipment — no lab, no mask, no blood draws — and the model derives your zones from there.
Isn't this the same as other metabolic-model tools? Both approaches run a metabolic model, which is the rigorous way to do this. The difference is how the anchors are derived: some tools frame the anchor as a curve-derived offset, whereas A Faster You runs on your own measured data, derives the anchors from the model directly, and explains the two-engine logic to you in plain terms.
Why do my zones differ from a friend with the same FTP? Because FTP is one number and you are two engines. If your friend has a higher VLamax and you have a higher VO2max, the model gives you different FatMax, MLSS and CP anchors even at the identical FTP — which is exactly the point of measuring both.
My app zones and my Powertest zones don't match — which is right? The Powertest zones are the measured baseline and the more accurate source. Between tests, AI-estimated zones update from your activities and can drift; when both exist, anchor on the most recent Powertest.
The Mader metabolic model is the scientific framework behind A Faster You's zones: Mader, A. & Heck, H. (1986) — A theory of the metabolic origin of the "anaerobic threshold". Int J Sports Med, 7 Suppl 1:45–65 (PMID 3744647). Mader, A. (2003) — Glycolysis and oxidative phosphorylation as a function of cytosolic phosphorylation state and power output of the muscle cell. Eur J Appl Physiol, 88(4–5):317–338. Maximal lactate steady state: Billat, V.L., Sirvent, P., Py, G., Koralsztein, J.P. & Mercier, J. (2003) — The concept of maximal lactate steady state: a bridge between biochemistry, physiology and sport science. Sports Med, 33(6):407–426. Critical power: Poole, D.C., Burnley, M., Vanhatalo, A., Rossiter, H.B. & Jones, A.M. (2016) — Critical Power: An Important Fatigue Threshold in Exercise Physiology. Med Sci Sports Exerc, 48(11):2320–2334. Lactate-threshold definitions and their disagreement: Faude, O., Kindermann, W. & Meyer, T. (2009) — Lactate threshold concepts: how valid are they? Sports Med, 39(6):469–490. VLamax estimation from a short maximal sprint is itself established, peer-reviewed methodology (Adam, J. et al. (2015) — Reliability of the calculated maximal lactate steady state in amateur cyclists. Biol Sport, 32(2):97-102; Nitzsche, N., Baumgärtel, L. & Schulz, H. (2018) — Comparison of Maximum Lactate Formation Rates in Ergometer Sprint and Maximum Strength Loads. Dtsch Z Sportmed, 69:13-18; Hauser, T., Adam, J. & Schulz, H. (2014) — Theor Biol Med Model, 11:25). Test–retest reliability of that sprint protocol: Harnish, C.R., Swensen, T.C. & King, D. (2023) — Reliability of the 15-s Maximal Lactate Accumulation Rate (VLamax) Test for Cycling. Physiologia, 3(4):542–551. A Faster You's application layer — running that method lab-free, calibrating it against 15,000+ standardized Powertests with a neural network, individualized zone derivation from combined VO2max + VLamax, and body-reserve/protein-turnover modeling — is our own engineering on top of that published foundation. Cohort reference: 1,202 A Faster You athletes, 15,000+ standardized Powertests, drawn from 1 million+ analysed training sessions.
Cover photo: Anna Lena Duschl