Most beginners test themselves one of two ways. Either they never time anything and go by feel, insisting they feel faster this month than last, or they sprint all out on the first rep of every session to see what the watch says. Both feel like discipline. Neither gives you something to point to and say: this program is working, here is the proof.
"I feel faster" does not survive a bad week of sleep, a stressful stretch at work, or the fact that motivation lies to you in both directions. Some days you feel like a different athlete and run your usual time; other days you feel sluggish and quietly set a personal best. Without a real benchmark, you are steering by the least reliable instrument you own: your own mood.
Why you need a number at all
A training program is a bet: do this work, in this order, and you will be measurably faster by a specific date. Without a benchmark you can only check whether you completed the workouts, not whether the bet paid off, and only one of those tells you the program is working.
Sprinting punishes vague self-assessment more than most sports, because real gains are small numbers moving slowly. A beginner who trains well for eight weeks might drop their 40-meter time by two or three tenths of a second, a genuine improvement invisible to how the sprint felt while running it. You cannot feel a tenth of a second, you can only measure one. Without a number, a program that is working and one that has stalled feel identical from the inside.
Why not the 100
Running an actual 100 meters looks like the obvious top-speed test. For a beginner it is the wrong one, for two reasons.
The first: a 100 does not measure one thing. The front 40 to 50 meters is acceleration into top speed, but the back half is a different event, holding form while decelerating, a speed-endurance and technique test wearing a top-speed costume. Your finishing time gets contaminated by fatigue tolerance, not the quality you meant to measure.
The second reason is risk. Eccentric load on the hamstring climbs steeply near top speed, and a 100 holds you near that peak load, on tissue that has not adapted to it yet, for an extended stretch well past where a 40 would already be over. The risk per rep of a full 100 is not modestly higher than a 40, it is a different category of exposure.
Why not your watch
GPS watches report a top speed number, tempting to trust instead of running a real test. It feels objective because a device produced it, but how it gets produced makes it a poor tool for this job.
Consumer GPS units sample your position a few times a second, then smooth the data to cancel signal noise, the right trade-off for a two-hour run and the wrong one for a two-second burst. A sprint's true top speed appears for a fraction of a second, right around the window the smoothing is built to average away. The same sprint can report different peak numbers from one day to the next, depending on satellite lock and the averaging window in use. You are measuring your speed filtered through an algorithm built for a different problem.
Why 40 meters works
A beginner reaches something close to top speed around 30 to 40 meters into a sprint, not at the gun. Before that point you are still accelerating, so a 40-meter time trial captures two things in one number: how well you get up to speed, and your speed as you arrive near your ceiling, a good proxy for what a beginner program is trying to build.
It is also short enough to be safe, repeatable, and practical. A 40 asks for a hard, honest maximal effort, but it ends before the extended near-maximal hold that makes a 100 risky, and before fatigue contaminates the number with something other than speed. It also fits the spaces most beginners have, a park straightaway or a stretch of empty road, without needing a full track. A test you cannot repeat under the same conditions every time is not really a test, and 40 meters is close to the shortest distance that still qualifies.
How to run one honestly
A time trial only means something if conditions are controlled enough that a change in the number reflects a change in you, not the weather or a rushed warm-up. Treat it with the seriousness of any max-effort session: warm up fully, the full 15 minutes, ending in build-up strides the way you would before any hard sprint day. A trial off a rushed warm-up tests how fast you can run underprepared, not your fitness, and raises injury risk on the rep you are asking to be maximal.
Run two attempts with full recovery between them, at least 4 to 5 minutes, and record the better one. The first rep carries a small tax from nerves that a rested second usually does not, and two attempts protect against one mistimed start ruining the test. Keep everything else identical every time: same location, surface, and shoes, and note the wind if outdoors. A tailwind can shave real time off a 40, and a number you cannot honestly compare to your last one is not tracking anything.
Timing it without kidding yourself
How you start the watch matters as much as how you run. A friend trackside with a stopwatch, reacting to your first movement by eye, carries built-in reaction-time error, usually a couple of tenths of a second either way. On a test where a real improvement might be one tenth, that is not a rounding error, it can be the whole result.
Phone video is more reliable and needs no special equipment, with one setup rule: the camera has to see both ends of the run. Prop the phone side-on near the midpoint, far enough back that the start line and the finish line both fit in the frame, then step through the footage frame by frame, counting from the first frame your body moves to the frame your torso crosses the line. A camera at the finish line cannot see your first movement 40 meters away, and a frame count missing either endpoint is not a time. Most phones shoot at 30 or 60 frames per second, tighter resolution than human reaction time, with no guesswork about when a thumb landed on a stopwatch.
Pick one method and never switch it mid-program. A stopwatch time and a frame-counted time are not directly comparable, even for the same run, so switching methods can look like progress or regression that never happened. The trend matters more than the precision of any single number, and it only means something if the measurement behind it stayed constant.
What progress looks like
The first time trials after starting a program tend to show a fast drop. Some of that is genuine speed, but a good share is neural, the nervous system learning the pattern before tissue has actually gotten stronger. Do not extrapolate that early rate forward, it will not hold.
After the early weeks, progress slows into a grind where a tenth of a second between trials is a hard-earned win. A flat result between two tests is not automatically failure, it is information: maybe you need more recovery, or the current phase is quietly building something that shows up as speed later. Reading it correctly, instead of panicking into more maximal effort, is what a periodized program is doing in the background.
None of this argues for testing constantly. A time trial is a maximal effort that a good program budgets carefully, spending it every few weeks at phase boundaries rather than every session, once enough training has accumulated for a new number to mean something. sprntr schedules time trials at set points for this reason, and tracks your 40-meter PR as a trend on the Progress screen, so you see its shape instead of reacting to any single run.
Where to go next
For the submaximal end of the effort scale that fills most of your training, read what 90 percent effort actually feels like. For the mechanism behind why a maximal effort has to be spent carefully, read the hamstring problem.