Guides · Treadmill time trial and race prediction
Can a treadmill time trial predict your race time?
Saturday is 5K time trial day on the belt. You set 7.5 and hold it to the end. You stop the clock at 24:51 and feed that time to a race predictor, and it hands back a half marathon goal. That goal is only as real as the pace you fed it, and nothing on the machine told you whether 7.5 was true.
Yes, a treadmill time trial can predict your race time, as long as the pace it is built on is real. A time trial measures how long you can hold a speed setting, and nothing on the machine checks that setting against the belt.
The pace it's built on isn't a measurement
Think of a ceiling fan with a pull chain. One click sets it to low, another to high, and the chain's click tells you which setting you chose, nothing more. Nothing inside the fan counts how fast the blades are actually turning, so a tired motor still gets called high.
Your treadmill's display works the same way. The number you see is the setting you chose, echoed back to you. The motor is driven toward a stored value filed under that number, and as far as I have been able to find out, nothing watches the belt's own moving surface to check it. Press 7.5 today and press 7.5 next month, and the display reports the same number both times, whether or not the belt did the same thing.
A race predictor adds a second layer on top of that. It is a multiplier, not a measurement. It takes the time you give it and scales it up to your race distance, and any error in that time scales up with it. It has no way to tell a real pace from an optimistic one, so whatever doubt was in the number you typed comes back out the other side, stretched across a longer distance.
What a wrong pace costs the prediction
In a lab study of four treadmills, the belts averaged between 98 and 103 percent of the speed that was set while people walked and ran on them. That was a controlled lab rather than a gym floor, and even there a belt landed a few percent off its number.
Say you hold 7.5 mph, an 8:00 mile pace, for 24:51, which is the time a 5K takes at that pace. Feed 24:51 for a 5K into a predictor and it hands back 1:54:20 for a half marathon, an 8:43 pace.
If the belt runs 3 percent slow, the pace you held was really 8:15, and the same predictor would have said 1:57:52, an 8:59 pace. So you line up planning 8:43 when an honest 5K time would have told you 8:59, sixteen seconds a mile faster than the prediction you actually earned. Your pacing plan and your fueling plan were built on that number, and the first mile of the race is where you find out.
Same 24:51, two belts
The clock reads the same either way. Only the belt underneath it changed.
| Belt runs 7.5 as set | Belt runs 3 percent slow | |
|---|---|---|
| Pace held | 8:00 | 8:15 |
| Distance in 24:51 | 3.1 miles | 3.0 miles |
| Honest 5K time | 24:51 | 25:37 |
| Half marathon prediction | 1:54:20 | 1:57:52 |
| Predicted race pace | 8:43 | 8:59 |
This is worked arithmetic on an example belt running 3 percent slow, not a reading I took.
The predictor passes the belt's error straight through, so a belt a few percent slow hands you a race pace about that much faster than you have earned. A belt can run fast instead, and then the prediction comes out too timid rather than too bold. Neither direction is the safe one, and both hand you a goal your legs have never actually produced.
One of the machines I run on said 2.3 miles when the belt had moved 2.0. That is one machine on one day, and it is not a rule about yours in either direction.
How to trust the pace you feed it
Run the trial at 1 percent incline, but only if your goal pace sits inside the band the rule was tested on, about 5:22 to 9:11 per mile. An 8:00 mile qualifies. The incline corrects for the missing wind on the effort side. It takes the speed setting itself on trust, so it does nothing for the number that feeds your predictor. Read treadmill pace vs road pace for the effort side in full.
A watch cannot referee this either. Indoors it has no GPS, so it guesses your pace from your arm motion instead, a gap the watch vs treadmill guide covers. Effort and pace are two different problems on a treadmill, and fixing one does nothing for the other. The only thing that actually settles whether your speed setting is real is a measurement of the belt itself.
Know the pace before you trust the prediction
Before your next time trial, set your trial speed with the belt empty and hold an iPhone Pro with Treadwell open over the moving belt. The app reads the pace straight off the moving surface with its camera and LiDAR sensor, so the number you hand the predictor is the pace the belt really ran. The predictor still does the multiplying. What changes is whether the input it starts from is real.
Know it before the trial
Find out what your treadmill's belt really does at your trial speed, in seconds, on the machine you run the trial on. Treadwell needs an iPhone Pro with LiDAR.
Quick answers
Can a treadmill time trial predict my race time? Yes, as long as the pace it is built on is real. A treadmill time trial measures how long you can hold a speed setting, and nothing on the machine checks that setting against the belt.
Should I run a treadmill time trial at 1 percent incline? Only if your goal pace sits inside the band the rule was tested on, about 5:22 to 9:11 per mile in nine trained male runners. Outside that range the 1 percent is a habit, not a proven fix, and it still says nothing about whether your speed setting is real.
Sources.
- The incline rule is Jones and Doust, "A 1% treadmill grade most accurately reflects the energetic cost of outdoor running," Journal of Sports Sciences, volume 14, 1996, pages 321 to 327, which measured nine trained male runners at six speeds between about 5:22 and 9:11 per mile; paces slower than that were not part of the study.
- The predictor formula traces to Peter Riegel, "Time Predicting," Runner's World, August 1977, and "Athletic Records and Human Endurance," American Scientist, volume 69, number 3, 1981, pages 285 to 290, which covers efforts between about 3.5 and 230 minutes.
- Four lab treadmills, one of them documented as substantially worn, ran between 98 and 103 percent of their set speed on average while people walked and ran on them, per Willwacher et al., Scientific Reports, 2021.
- Indoors, a running watch has no GPS signal and estimates pace from arm motion and its accelerometer, per Apple Support article 105002 and the Garmin Forerunner owner's manuals.
- The 2.3-vs-2.0 result is from one run on June 13, 2026, checked against that treadmill's standing tachometer table, with no handheld tachometer used on the same day.