Critical Speed for Runners: The Field Test That Sets Your Run Zones

Most age-group triathletes have a power-based anchor for the bike, whether it's FTP or critical power, and an increasing number now use critical swim speed in the pool. On the run, many still train by feel or by heart rate alone. Heart rate is useful, but it drifts with heat, fatigue, caffeine and sleep. Pace is more direct, but only if you know which paces mean what for you.

Critical speed fills that gap. It's the running equivalent of critical power, the concept explored in TriForward's "Is FTP 'Dead'?" and "The W′ Bank," and of critical swim speed, covered in "Critical Swim Speed: Setting Swim Intensity Without a Coach on Deck." This article explains what critical speed is, how to estimate it, how reliable the estimates are, and how to use it in training and racing.

Why triathletes need a run anchor

Triathletes have a particular problem with run intensity. Much of your running happens tired: the day after a long ride, off the bike in a brick, or squeezed in before work after a hard swim. On those days, the same heart rate can correspond to very different paces, and the same pace can feel very different. Without a fixed reference, it's easy for "easy" runs to drift into a moderate middle zone and for "hard" sessions to be held back by fatigue nobody accounted for.

A test-derived pace doesn't solve fatigue, but it gives you a stable yardstick. If your threshold session is set as a percentage of your critical speed, you know whether you hit the intended intensity, and you can see when fatigue is pushing you off it. Over a season, repeated CS tests also show whether your running is improving, independently of how your heart rate behaves on a particular day.

There's also a structural reason. Run training carries more injury risk than swimming or cycling, so the intensity of each session matters more. Knowing exactly which sessions are hard, and keeping the rest genuinely easy, is one of the simplest ways to manage that risk. TriForward's "Training Load & Injury Risk" articles cover the load side in more detail.

What critical speed is

If you run several maximal efforts of different durations and plot the distance covered against the time taken, the points fall close to a straight line. The slope of that line is your critical speed (CS). The intercept, the extra distance you can cover above CS before you run out, is called D′ (D-prime). It's the running equivalent of W′ in cycling.

In physiological terms, critical speed marks the boundary between two exercise intensity domains. Below it, blood lactate and oxygen uptake settle at a steady level. Above it, they keep rising until you have to stop. Jones and Vanhatalo's 2017 review in Sports Medicine describes this boundary, and explains how work above it draws on a finite reserve that recovers when you drop below it again. That's why surges above CS feel sustainable for a while, then suddenly don't.

CS isn't your lactate threshold

It's tempting to treat CS as another name for "threshold pace," but the two aren't identical. Jones and colleagues argued in 2019 that maximal lactate steady state (MLSS), a common lab-based threshold, typically sits at a slightly lower power or speed than critical power, and that the two shouldn't be used interchangeably. They described critical power as sustainable for only about 20–30 minutes.

For runners, that means:

• CS is hard. It's roughly the pace you could hold for a 5 km to 10 km race, depending on your speed, not an hour.
• Your "threshold" sessions may be slightly below CS, if your coach defines threshold using lactate.
• Most endurance training sits well below CS.

How to estimate it

There are three practical options.

1. Two or three time trials

The classic approach uses maximal efforts of different durations, typically between about 2 and 15 minutes. Common field versions are:

• Time-based: run as far as you can in 3 minutes, and as far as you can in 12 minutes, on separate days or with full recovery.
• Distance-based: run a set distance as fast as you can, for example 1,200 m and 3,600 m.

In a study by Galbraith and colleagues (2014), 10 male distance runners completed three time trials (3,600 m, 2,400 m and 1,200 m) on a single visit to an outdoor track, with either 30 or 60 minutes of recovery between runs. Critical speed came out the same with either recovery time, correlated well with a traditional treadmill protocol, and was highly repeatable. The coefficient of variation for CS was 0.4%. D′ was much less reliable, at 13%, and didn't agree with the treadmill estimate.

2. A 3-minute all-out test

An alternative is a single 3-minute all-out run. The speed you can hold over the final part of the test approximates CS. It's convenient, but pacing a genuine all-out effort is difficult, and a 2019 methods review by Muniz-Pumares and colleagues noted concerns about the validity of some time-trial and all-out approaches compared with traditional lab protocols.

3. Your training data

If you regularly run hard, your watch already contains much of the information needed. Smyth and Muniz-Pumares (2020) estimated critical speed from the raw training records of more than 25,000 recreational marathoners in the weeks before big-city marathons. The distance-time relationship was close to linear, and CS estimates correlated with marathon performance.

Hunter, Ledger and Muniz-Pumares (2023) tested this more directly in 23 recreational runners (4 women) wearing foot pods for 8 weeks. They compared CS estimated from each runner's best 3-, 7- and 12-minute efforts in normal training with CS from formal time trials and from 3-minute all-out tests. Estimates from training (3.44 m/s on average) and time trials (3.42 m/s) were very similar, and the authors concluded CS can be estimated remotely. D′ again agreed poorly between methods.

The catch is that training-data estimates only work if your training includes near-maximal efforts at the relevant durations. If you never run hard for 3 or 12 minutes, the estimate will be too low.

A 2025 systematic review by Lipková and colleagues, covering 19 studies, concluded that field-based CS tests are a practical alternative to the lab, but that reliable results depend on following recommended settings, particularly for time trials.

Running the test well

Whichever version you choose, the quality of the estimate depends on the quality of the efforts. A simple protocol for the two-trial version, based on common practice rather than a single tested method:

1. Pick a flat, measured course. A 400 m track is ideal. GPS on a road is acceptable if you use the same route each time, but it's less precise over short efforts.
2. Warm up properly. Around 15–20 minutes of easy running, a few strides, and some drills.
3. Run the shorter effort first. For the 3-minute trial, aim for an even, hard pace you can just about hold. Starting in a sprint usually leads to a big fade.
4. Recover fully. At least 30 minutes of easy jogging and walking if both trials are on the same day, based on Galbraith's finding that 30 or 60 minutes gave the same CS. Many athletes prefer separate days.
5. Run the longer effort evenly. For the 12-minute trial, treat it like a short race and try to finish with nothing left.
6. Record distance precisely. On a track, note laps and the final partial lap.
7. Cool down. Then calculate.

Treat the test as a hard training session in your week, not an extra. Put it after an easy day and don't schedule another hard run within 48 hours.

How CS compares across the three disciplines

Triathletes who already know their critical power and critical swim speed will recognise the pattern. The same two-parameter model underpins all three:

• Bike: critical power (CP) and W′, in watts and kilojoules.
• Swim: critical swim speed (CSS) and D′, in metres per second and metres.
• Run: critical speed (CS) and D′, in metres per second and metres.

In all three, the critical value is far more reliable than the reserve above it. In TriForward's swim article, a study of 32 national-standard swimmers found critical speed varied by less than about 3.4% between tests, while D′ varied by 13–45%. Galbraith's running data show the same pattern. The practical lesson carries across: trust the critical value, and treat the reserve as a rough guide at best.

A worked example

The following numbers are illustrative, not from a study.

You run as far as you can in 3 minutes on a track and cover 900 m. On another day, fresh, you run as far as you can in 12 minutes and cover 3,000 m.

• Difference in distance: 3,000 − 900 = 2,100 m
• Difference in time: 720 − 180 = 540 seconds
• Critical speed: 2,100 ÷ 540 = 3.89 m/s
• As a pace: 1,000 ÷ 3.89 = 257 seconds per km, or about 4:17/km
• D′: 900 − (3.89 × 180) ≈ 200 m

Based on Galbraith's data, the CS figure is likely to be repeatable within a fraction of a percent if you test the same way. The D′ figure is much less trustworthy, so don't read much into small changes in it.

Using CS to set training paces

There's no universally agreed set of running zones based on CS. The bands below are common coaching conventions, labelled as such, using the 4:17/km example:

• Easy and long runs: well below CS, often around 70–80% of CS speed (roughly 5:20–6:10/km in the example). This is where most training time goes.
• Steady or tempo: about 85–90% of CS speed (roughly 4:45–5:00/km).
• Threshold-type intervals: about 95–100% of CS speed (roughly 4:17–4:30/km), for example 4–6 × 5 minutes with short recoveries.
• Intervals above CS: faster than 4:17/km, in short repetitions. These use D′ and need longer recovery.

Remember that these are percentages of speed, not pace. Running at 80% of CS speed means a pace about 25% slower, not 20%.

What CS says about pacing

The 25,000-runner study offers one of the few large-scale links between CS and race pacing. On average, the marathoners completed the race at about 85% of their critical speed. Faster runners raced closer to CS: around 93% for those finishing in about 150 minutes, compared with around 79% for those finishing in about 360 minutes. Runners who started faster than 94% of their CS were more likely to slow by more than 25% in the second half.

A separate analysis of 82,303 marathoners by the same research group, covered in TriForward's "Why the Ironman Marathon Falls Apart After 25 km," found that runners whose heart rate drifted most relative to pace finished slower, despite racing at a lower average percentage of CS.

For triathletes, inference only: an Ironman marathon follows hours of swimming and cycling, so your sustainable pace will be lower than in a stand-alone marathon. Your standalone marathon percentage is a ceiling, not a target. No study has yet measured where Ironman age-groupers typically run relative to their CS.

For sprint and Olympic-distance races, the run is short enough that CS is a more direct guide, but running off the bike still costs some speed. See TriForward's "The Bike-to-Run 'Brick' Problem."

When to retest

Fitness changes, so CS should be updated. A practical rhythm is every 6–8 weeks during a training block, or after a change in training emphasis. Retest under similar conditions: the same track or route, similar weather, and after an easy day. Look at trends across several tests rather than a single number.

Common mistakes

• Going out too fast in the longer effort, then fading. This underestimates CS.
• Not recovering fully between efforts. A tired second trial distorts the slope.
• Using durations that are too close together. Efforts of 3 and 5 minutes give a less stable estimate than 3 and 12.
• Treating CS as an easy pace. It's hard. Most training should be well below it.
• Reading meaning into D′ changes. They're unreliable.
• Comparing tests in different conditions, such as heat versus cool, or track versus a hilly road.
• Using CS from training data when you never run hard. The estimate will be too low.

Who the studies tested

The single-visit validation study involved 10 trained men. The training-data study included 23 recreational runners, only 4 of them women. The large marathon analysis included thousands of women but used consumer devices and estimated CS indirectly. None of the studies tested triathletes running off the bike. The underlying model is robust, but the exact reliability figures may not transfer perfectly to every age-grouper.

References

• Jones AM, Vanhatalo A. The 'critical power' concept: applications to sports performance with a focus on intermittent high-intensity exercise. Sports Medicine, 2017;47(Suppl 1):65–78. doi:10.1007/s40279-017-0688-0
• Jones AM, Burnley M, Black MI, Poole DC, Vanhatalo A. The maximal metabolic steady state: redefining the 'gold standard'. Physiological Reports, 2019;7(10):e14098. doi:10.14814/phy2.14098
• Galbraith A, Hopker J, Lelliott S, Diddams L, Passfield L. A single-visit field test of critical speed. International Journal of Sports Physiology and Performance, 2014;9(6):931–935. doi:10.1123/ijspp.2013-0507
• Muniz-Pumares D, Karsten B, Triska C, Glaister M. Methodological approaches and related challenges associated with the determination of critical power and curvature constant. Journal of Strength and Conditioning Research, 2019;33(2):584–596. doi:10.1519/JSC.0000000000002977
• Smyth B, Muniz-Pumares D. Calculation of critical speed from raw training data in recreational marathon runners. Medicine & Science in Sports & Exercise, 2020;52(12):2637–2645. doi:10.1249/MSS.0000000000002412
• Hunter B, Ledger A, Muniz-Pumares D. Remote determination of critical speed and critical power in recreational runners. International Journal of Sports Physiology and Performance, 2023;18(12):1449–1456. doi:10.1123/ijspp.2023-0276
• Lipková L, Struhár I, Krajňák J, Puda D, Kumstát M. Field-based tests for determining critical speed among runners and its practical application: a systematic review. Frontiers in Sports and Active Living, 2025;7:1520914. doi:10.3389/fspor.2025.1520914
• Smyth B, Maunder E, Meyler S, Hunter B, Muniz-Pumares D. Decoupling of internal and external workload during a marathon: an analysis of durability in 82,303 recreational runners. Sports Medicine, 2022;52(9):2283–2295. doi:10.1007/s40279-022-01680-5