- A threshold model is a calculation rule. It uses your lactate curve to estimate where the anaerobic threshold lies.
- A rule of thumb from practice: Mader for running, Dickhuth for cycling.
- The model has to match the test protocol. Mader belongs with 5-minute stages on the treadmill.
- Stick to one model. If you switch, you can no longer compare your tests.
Mader, Dickhuth, Dmax, Stegmann: anyone opening an analysis for the first time is faced with a list of names. Behind them are different routes to the same goal. This chapter explains the most important ones and tells you when to use which.
Why models are needed
The anaerobic threshold is the highest workload at which your lactate still just remains stable. It can only be measured directly with several constant-load tests of about 30 minutes each on different days. That takes a lot of time and effort.
The step test is a shortcut. It gives you a curve in half an hour. From this curve, a threshold model estimates where the threshold lies. Each model does this according to its own rule, which is why the same test produces different results. That is not an error; it lies in the nature of the method.
Values as a table
| Speed | Lactate |
|---|---|
| 8 km/h | 1.4 mmol/L |
| 10 km/h | 1.3 mmol/L |
| 12 km/h | 1.6 mmol/L |
| 14 km/h | 2.4 mmol/L |
| 16 km/h | 4.3 mmol/L |
| 18 km/h | 8.2 mmol/L |
| Baseline lactate (Dickhuth) | 1.5 mmol/L at 10.7 km/h |
| Dickhuth threshold | 3.0 mmol/L at 14.8 km/h |
| Mader threshold | 4.0 mmol/L at 15.8 km/h |
Mader: the fixed 4
Alois Mader’s model from 1976 is the simplest. The threshold lies where the lactate curve reaches 4 mmol/L. The value comes from treadmill tests with 5-minute stages.
- Strength: simple, unambiguous and well reproducible from test to test.
- Weakness: 4 mmol/L is an average. For some people the actual threshold lies below it, for others above it. In athletes with a very high level of endurance training it is often lower.
- Requirement: The model presupposes its protocol. With shorter stages the lactate values come out lower, and the fixed 4 no longer fits.
Dickhuth: baseline plus a fixed amount
Hans-Hermann Dickhuth’s model starts from the individual. First it looks for the baseline lactate. This is the point at which the ratio of lactate to power output is at its lowest. It corresponds to the aerobic threshold. A fixed amount is added to this, 1.5 mmol/L by default. Where the curve reaches this value is the anaerobic threshold.
- Strength: The model takes your own baseline into account. The absolute level of the lactate values does not matter. That is why it works equally well for running and cycling tests.
- Weakness: It needs clean values in the lower part of the curve. An outlier in the first stages shifts the baseline lactate and with it the threshold.
- Requirement: enough easy stages at the start; 3-minute stages are sufficient.
Further models at a glance
| Model | How it calculates | What it needs |
|---|---|---|
| LT1 | First rise of the curve above the baseline. Gives the aerobic threshold. | Several easy stages at the start |
| Dmax | Joins the first and the last data point with a straight line. The threshold lies where the curve is furthest from this line. | A test to exhaustion and a suitable starting stage |
| Modified Dmax | Like Dmax, but the straight line starts at the baseline lactate. | A test to exhaustion |
| Stegmann | Draws a tangent to the curve, starting from the lactate values during recovery. | Exhaustion and readings after stopping |
| Fixed values | Power output at a freely chosen lactate value, for example 2 mmol/L. | A protocol that stays the same |
| Conconi | Looks for the deflection point in the heart rate curve, without any blood at all. | A clear deflection point, which not everyone shows |
Which one, and when?
The reason is the protocol. Mader found his 4 mmol/L on the treadmill with 5-minute stages. That is where the model belongs. On the bike, tests are usually run with 3-minute stages, and the lactate values turn out differently at the same level of strain. Here, a model that takes its bearings from your own baseline rather than from a fixed number has the advantage.
| Your test | Natural choice of model | Why |
|---|---|---|
| Treadmill, 5-minute stages | Mader | The protocol to which the fixed 4 applies |
| Running with 3-minute stages or field test | Dickhuth | Independent of the absolute level of the values |
| Bike, 3-minute stages | Dickhuth | The fixed 4 does not match the protocol |
| Test not taken to exhaustion | Mader or Dickhuth | Dmax and Stegmann need the final point |
| You are looking for the aerobic threshold | LT1 or baseline lactate | Both describe the first rise |
| Comparison over months and years | Always the same one | Only then do you see your progress |
Three rules for working with models
- Protocol, model and device stay the same. Then every change in your threshold is a change in you and not in the method.
- Look at the curve before you look at the number. An outlier distorts every model. If a point does not fit the trend, the cause is usually the sampling.
- The threshold is a range. No model pins it down to a tenth of a km/h. A model is a tool; it does not replace the judgement of an experienced diagnostician.
Models in the app and software
The Lactate Express App comes with Mader, Dickhuth, LT1, Dmax, modified Dmax and freely selectable fixed values. You define the model together with your training zones in a concept. How this works is explained in Chapter 6.
WinLactat goes further: it adds, among others, Stegmann, Conconi, the free Freiburg model, angle models and the analysis of the lactate minimum test. Several models can be applied to one test at the same time and compared. The training zones are always based on exactly one model, which you define as the leading one.
The matching of models to protocols is a guide drawn from diagnostic practice. It does not replace an assessment of the individual case.