Updated v1.06 min read
Race Time Prediction
Your quick, science-backed guide to race time prediction — how VDOT turns one race result into finish times at every distance, and when to trust the number.
TL;DR
One honest race result predicts your time at every other distance — the input matters more than the formula.
- Use a recent all-out race or time trial as the input
- Predict between nearby distances (5K → 10K) for the tightest accuracy
- Prefer a watch's running VO₂max over a lab number
- Treat marathon predictions as a range, not a promise
- Re-test every 6–8 weeks as fitness shifts
- Don't extrapolate a 5K straight to marathon distance
- Don't feed in a hot, hilly, or badly-paced effort
- Don't trust a lab VO₂max reading over a real race result
- Don't skip distance-specific training and expect the number to hold
Garbage in, garbage out: a clean input race beats any amount of formula-tweaking.
What race time prediction actually is
A predictor doesn't measure your fitness directly — it reverse-engineers it from a performance you've already produced. The standard method, the Daniels-Gilbert formula, converts a distance-and-time pair into VDOT: a number that folds together both your aerobic ceiling and your running economy (how efficiently you convert oxygen into speed). Two runners can share an identical lab VO₂max and still post very different race times — the more economical runner has the higher VDOT, and the formula captures that automatically because it works from a real performance, not a treadmill reading.
That's also why the source of your VO₂max matters. A running watch estimate is distance-specific and tracks VDOT closely. A lab-measured VO₂max ignores economy entirely, so it typically reads 5–10% higher than your true VDOT — plug it in and every predicted time comes out slightly optimistic.
What sharpens the prediction
Where accuracy actually comes from. Sorted roughly by how much each one moves the number.
| Factor | Why it helps | Impact | Best for |
|---|---|---|---|
| A recent all-out race result | Captures fitness and economy in one number — no separate formula guesses needed. | ●●●● | Any prediction |
| Predicting between nearby distances | 5K → 10K keeps the extrapolation short; the %VO₂max gap between them is small. | ●●●● | 3K–15K range |
| A flat, cool, honestly-paced input effort | The formula reads pace as pure effort — heat, hills, and uneven pacing all distort it. | ●●●● | Any prediction |
| Training specifically for the target distance | The table assumes race-ready fitness at that distance; specific endurance work closes the gap. | ●●●● | Half & marathon |
| A running watch's VO₂max estimate | Distance-specific and close to VDOT — a solid stand-in with no race on the calendar. | ●●●● | Quick estimates |
| Re-testing every 6–8 weeks | VDOT moves with training; a stale input race quietly drifts the whole table. | ●●●● | Ongoing training |
What throws it off
The leaks. The cost column shows how much each one distorts the table, with the fix beside it.
| Factor | What goes wrong | Cost | Fix |
|---|---|---|---|
| Big distance gaps (5K → marathon) | Sustainable %VO₂max drops from ~95–97% at 5K to ~75–80% at marathon — a short race barely constrains the far end of that curve. | ●●●● | Anchor with a race or long tempo closer to your target distance. |
| Undertrained for the target distance | The table assumes matched training; a 5K specialist's marathon prediction ignores glycogen, fueling, and durability they haven't built. | ●●●● | Build distance-specific long runs and tempo work first. |
| A lab VO₂max as the input | Runs 5–10% above true VDOT because it ignores running economy — every prediction skews optimistic. | ●●●● | Use a watch estimate or a real race result instead. |
| A hilly, hot, or poorly-paced input race | Distorts the pace-to-effort mapping the whole formula depends on. | ●●●● | Re-test on a flat course in cool conditions, run evenly. |
| Treating the output as a guarantee | Pacing discipline and fueling execution on race day still decide the actual result. | ●●●● | Use the prediction as a target range, not a promise. |
Three ways to get your starting number
You don't need a race on the calendar to start. Pick whichever is available.
Recent race result
●●●●A race you ran hard in recent weeks, close to full effort. The gold-standard input — it captures real fitness and economy together, with no estimation step.
All-out time trial
●●●●Nothing on the calendar? A genuine all-out effort over a known distance, ideally 3K–10K on a flat, honest course, works just as well.
Watch VO₂max
●●●●Running-specific watch estimates track VDOT closely. Skip a lab VO₂max here — it reads 5–10% high and will over-predict every distance.
How far you can trust it
Every prediction ships with a margin, and it widens with distance — pacing, fueling, and glycogen management add variability a formula alone can't capture.
| Distance | Prediction margin |
|---|---|
| 1K | ± 3% |
| 3K – 10K | ± 2.5% |
| 15K | ± 3.5% |
| Half Marathon | ± 4% |
| Marathon | ± 5.5% |
Which tool answers which question
Four tools cover this territory, and it's easy to reach for the wrong one.
| Question | Tool |
|---|---|
| What time could I run at another distance? | Race Time Predictor |
| What pace should today's easy, threshold, or interval run be? | Training Pace Calculator |
| How do I manually plan my own pacing splits for a race? | Race Pace Calculator |
| How should my splits look based on my actual fitness? | Race Strategy Calculator |
How to track it
A single prediction is a snapshot, not a trend. What actually matters is whether your VDOT is climbing.
Common mistakes
The traps that quietly wreck an otherwise sound prediction.
Go deeper
There's no single deep-dive article for this one yet — the full methodology lives on the Race Time Predictor itself.