Athletes ask this a lot. How long do you think my race will take.
There are loads of ways to model a finish time using pace, heart rate, power, VO2, threshold, long run history, terrain, heat, and so on. Those can be useful, but they can also mislead people, especially when the course is unfamiliar or the distance is new.
One method I keep coming back to is simpler.
It is a percentile method. You work out where you usually place in a field, then you apply that same placing percentile to the target event’s results. It sounds almost too basic, but it is surprisingly consistent for most recreational runners, especially across similar event types.
The core principle
Most athletes have a fairly stable “competitive fingerprint”.
Not their pace. Not their heart rate. Not their wattage.
Their finishing position relative to everyone else.
If you regularly finish around the top third of a field, you tend to keep doing that. If you usually land mid pack, you tend to keep doing that. Same for back third. It is not perfect, but it is often more stable than people think.
Step by step
✅ Step 1: Collect a few past results
Grab 3 to 6 events where the athlete raced with reasonable intent. Skip novelty days where they jogged with a mate, got injured, treated it as a long walk, or stopped for 40 minutes at the bakery.
✅ Step 2: Convert each result to a percentile
You can do this roughly or properly.
Rough version:
If you finished about 100th in a field of 300, you are about one third of the way through.
More exact version:
Percentile position ≈ placing ÷ field size
100 ÷ 300 = 0.333, so about the 33rd percentile from the front.
Do this for each past event, then look for a repeating pattern.
Note that you can often use your percentage finish position in the total field of all finishes, and/or your percentage finish position by gender and by age group. You will notice that they often all align more or less the same.
✅ Step 3: Apply that percentile to the target event
Once you have a stable pattern, you assume they will land around the same percentile in the upcoming race, then you map that percentile onto the likely finish time distribution for that event.
This is easiest if you have last year’s results for the target event.
✅ Step 4: Adjust slightly for distance and experience
If it is a longer distance than they have done before, assume they will be a little more conservative or less efficient. That usually means slipping a few percent, not a total collapse.
Think in terms of 2 to 6 percent worse, not 30 percent worse, assuming training has been sensible and there is no major blow up risk.
A worked example
Let’s say an athlete has these recent results:
🏁 Race A: 210 finishers, placed 72nd
72 ÷ 210 = 0.343
🏁 Race B: 480 finishers, placed 155th
155 ÷ 480 = 0.323
🏁 Race C: 160 finishers, placed 58th
58 ÷ 160 = 0.363
These cluster around 0.34 to 0.36. Call it 0.35, meaning they typically finish around 35 percent of the way through the field from the front, or roughly top third.
Now they are doing a new event, same vibe, but a bit longer. Last year the target event had 600 finishers.
Expected placing ≈ 0.35 × 600 = 210th.
So we now go to last year’s results and find the 210th finisher time. If the 210th finisher time was 6:18, that is our base estimate.
Because it is a longer distance than they have raced before, we adjust slightly more conservative. Add maybe 3 percent.
6:18 is 378 minutes.
378 × 0.03 = 11.34 minutes.
Estimated finish time becomes about 6:29.
That gives the athlete a realistic finish window and it is gold for pacing plans, aid station timing, and support crew logistics.
Why this works more often than it should
It automatically bakes in a bunch of stuff that single metrics miss:
Decision making under fatigue
How well they handle hills and trail rhythm
How well they fuel and drink in real life
How they cope when things go wrong
Their ability to race, not just train
Percentile is a proxy for the whole performance package.
When it does not work well
There are a few big caveats.
⚠️ The course or conditions change a lot
If last year was cool and dry and this year is hot, humid, or muddy, everyone slows down and spreads out. The athlete might still finish around the same percentile, but the time attached to that percentile can shift a lot. In those years, time prediction gets shaky.
⚠️ The event attracts a very different field
A small local trail race and a championship style event can have very different depth. Your percentile might drift.
⚠️ The athlete has changed materially
Big fitness gains, a major injury, large weight changes, or a new endurance base can shift the fingerprint. This method assumes the athlete is broadly similar to the athlete who produced those previous results.
⚠️ DNF risk is meaningful
Percentile assumes a finish. If the athlete is undertrained for the distance, has major nutrition issues, or is carrying an injury, you need a different conversation first.
A note on pacing behaviour
One practical use of this method is aid station planning.
And yes, there is a common pattern where many men go out too hot and then pay for it later. Many women tend to pace more evenly across the day. That does not mean women always pace perfectly or men always race like idiots, but it is a pattern you see often enough that it matters for crew timing.
So if you are using a percentile based predicted finish time, it can help to also give a split range:
⏱️ Early race: could be a bit faster than predicted
⏱️ Late race: could be a bit slower than predicted
Especially for athletes who love an enthusiastic start.
How I use this in coaching
I do not use this method to replace training. It is not a magic calculator.
I use it to set expectations and reduce anxiety, especially for first timers at a distance. It gives athletes something grounded, and it gives crews a usable plan.
If you want to try it, pull up your last few results, work out your typical percentile, then go and find that percentile in last year’s results for your target race. You will usually end up with a finish time estimate that passes the sniff test.
This can be particularly effective when combined with a tool like https://ultrapacer.com/
