Your lab test says 4:45/km threshold. You step onto the track and that pace feels like a tempo run gone wrong. Your heart rate spikes early. You’re blowing up where you should be cruising.
The problem isn’t your fitness. It’s your test environment.
Blood lactate runs systematically lower on a treadmill than on a track at the same pace. Suzuki and colleagues (2026) tested 9 elite male runners (5000 m PB: 13:35) on both surfaces and found the lactate gap grew from 1.0 mmol/L at easy pace all the way to 3.7 mmol/L at near-race speed. That widening gap shifted threshold velocity by 7–8 seconds per kilometer. Not a rounding error. A full training zone.
Why Treadmill Lactate Is Lower Than Track Lactate at the Same Pace
The treadmill does some of your work for you.
The moving belt passively assists your trailing leg on every stride. That reduces demand on your hamstrings and hip extensors, which are the big muscles that generate lactate under load. Van Hooren et al.’s 2019 meta-analysis of 33 studies and 494 participants found treadmill runners showed 0.04 body weights less propulsive force per step, 14–16 ms less ground contact time, and a 9.8° flatter foot-ground angle at landing. Each of those changes means less muscular work per stride.
Less muscular work means less lactate produced.
On a track, you push against still air. At 5000 m race pace, air resistance consumes roughly 7.5% of your total energy budget, according to Pugh (1970). The treadmill eliminates that entirely. Your muscles aren’t working as hard to maintain the same pace number on the display.
Think of it like running in a tailwind that never stops. Your pace looks identical. Your body is doing less.
The result: a treadmill lactate profile reads like a fitter, fresher version of you. Zones derived from it will be set too fast for real-world running.
The 1% Incline Rule: What It Fixes and What It Doesn’t
Jones and Doust (1996) showed that setting the treadmill to 1% incline roughly matches the energy cost of outdoor running at moderate speeds — specifically between 3.75 and 5.0 m/s (4:27 to 3:20/km). That became the standard advice everywhere.
But the rule has edges, and they matter.
Below ~7:30/mile (4:40/km), the 1% incline actually over-corrects. You’re adding more grade penalty than the air resistance you’re trying to simulate. Above ~6:30/mile (4:02/km, where many threshold workouts sit), the 1% grade under-corrects — air resistance at faster paces is higher, and 1.5–2% would be more accurate.
Even when 1% is the right fix for air resistance, it doesn’t close the full lactate gap. Miller et al.’s 2019 meta-analysis of 34 crossover studies found that at 0% grade, submaximal treadmill blood lactate averaged 1.26 mmol/L lower than overground. At 1% grade, the gap narrowed to 0.52 mmol/L lower. Better, but not gone.
Pind and Mooses (2019) compared 17 male endurance athletes running on an indoor track versus a 1%-incline treadmill. Running economy was still 7.9% better on the track at 11 km/h, narrowing to 2.8% better at 15 km/h. The 1% fix helped. It didn’t solve the problem.
The 1% rule is necessary. It isn’t enough.
Track vs Treadmill Lactate Testing: Quantifying the Zone Error
Here’s what the gap looks like in practical terms across the three surfaces you might use.
| Surface | Submaximal Lactate vs Track | Running Economy vs Track | Zone Prescription Error |
|---|---|---|---|
| Track / road | Baseline | Baseline | 0 s/km |
| Treadmill 0% grade | −1.26 ± 0.71 mmol/L | −8.8% | ~7–8 s/km too fast |
| Treadmill 1% grade | −0.52 ± 0.50 mmol/L | −2.8 to −7.9% | ~3–4 s/km too fast |
Sources: Miller et al. 2019, Pind/Mooses 2019, Suzuki et al. 2026. Economy figures at 11–16 km/h.
Seven seconds per kilometer sounds small. Over a 10 km threshold block, that’s 70 seconds of cumulative over-effort. You’ll finish it. But you’ll be digging into a deeper lactate debt than your training plan intended, week after week.
A Case Where the Zones Were Wrong for Six Months
Daniel is 34, training for a 3:10 marathon with about 70 km/week. He did his lactate threshold test in January on a gym treadmill at 0% grade. His sport scientist set his threshold at 4:52/km, giving him Zone 4 intervals at 4:40–4:50/km pace.
Every spring track session felt brutal. He blamed fitness. He backed off the intervals, worried about overtraining.
In May, he retested on an outdoor track. His threshold came back at 5:01/km — nearly 10 seconds per kilometer slower than the treadmill had suggested. His “threshold intervals” all winter had been running well above threshold. He wasn’t undertrained. He was chronically over-stressed.
Six weeks of correctly prescribed Zone 3 and Zone 4 work later, he ran a 3:07 marathon. Not because he trained harder. Because he trained at the right intensity.
Environmental Confounds That Bias Outdoor Tests
Hot outdoor tests carry their own error, in the opposite direction.
James et al. (2017) compared lactate tests at 32°C versus 13°C. Threshold velocity dropped by 0.6–0.7 km/h in the heat. Zones set from a summer track test will be too conservative for cool-weather or indoor training.
The table below shows how environment stacks onto surface to compound the error.
| Test Condition | Direction of Error | Pace Error (approx.) |
|---|---|---|
| Treadmill 0% (cool lab) | Zones too fast outdoors | +7–8 s/km |
| Treadmill 1% (cool lab) | Zones slightly too fast outdoors | +3–4 s/km |
| Track, 32°C summer | Zones too slow for cool/indoor training | −4–5 s/km |
| Track, ~13°C (neutral) | Best outdoor reference | ~0 error |
If you test on a treadmill in a cool lab and race on a road in April, you’re likely overtrained in the months leading up to it.
What Your Smartwatch Gets Wrong
Your watch probably makes this worse.
Lu et al. (2025) validated Garmin, Coros, and Huawei LT estimates against real blood lactate testing. Garmin overestimated threshold pace by 2.17 km/h (25.78% MAPE). Coros overestimated by 1.93 km/h (22.63%). The researchers specifically named the treadmill-vs-outdoor lactate discrepancy as a contributing mechanism — these algorithms are calibrated on lab treadmill data, then applied to outdoor running.
That means your Garmin’s threshold pace is probably faster than your actual threshold. Not because the watch is bad. Because it learned from a different surface than where you’re running.
This connects to what you see in aerobic decoupling data: if your drift ratio climbs on every “Zone 3” run, your zone might be wrong, not your fitness.
Practical Correction Strategies
Match surface to goal. If you race on roads or a track, test there. A road test at 13–18°C with low wind is the cleanest reference most runners can get without a lab.
If you must test on a treadmill, use 1% grade and subtract 3–4 s/km from your threshold pace when applying those zones outdoors.
Re-test when you change environments. Winter treadmill training followed by spring track work needs a re-anchor. Your fitness carries over. Your surface-specific zones don’t.
Don’t rely on a watch alone for zone prescription. Wearable LT estimates carry 12–26% error (Lu et al. 2025) before surface effects compound the problem further. Use heart rate as a check, not a primary zone anchor, unless you’ve verified the device against a blood test.
AthleteOS detects whether your workout GPS signature matches indoor (treadmill) or outdoor running, then applies a surface-specific correction to your prescribed threshold and Zone 3 paces. If your threshold test was done outdoors, your indoor prescribed paces shift 3–7 s/km faster to hit the same metabolic target. You don’t recalibrate manually — the correction happens automatically each time you sync.
Understanding why this correction matters connects back to how training zones are built in the first place. Zone 2 training and your CTL/ATL/TSB trends only mean what they’re supposed to mean if the underlying pace anchors are accurate.
Get the surface wrong and everything built on top of it drifts.
Test on the surface that matters. Or know the correction factor if you can’t. Seven seconds per kilometer is the difference between the right zone and the wrong one — and six months of training you can’t get back.
Set up surface-aware zones in AthleteOS and let the correction run in the background while you focus on training.