Champagne Running / Race-time predictor

Your next race.

A range of outcomes.

Start with a race you’ve run. Explore what it suggests about another distance.

Estimated marathon finish

—

Loading the outcomes…

Middle 80% of outcomes

—

A range from real runners’ personal bests.

01 / The bigger picture

A range of outcomes

Real runners with a similar personal best at your starting distance.

Middle 80%Outside 80%Observed middleYour goal

These are recorded best times, not a promise for race day. Preparation for your target distance matters.

02 / Different approaches

Compare the predictions

Choose a method to see its estimate above.

MethodFinish time

The range comes from similar runners. Differences between formulas are not a range of uncertainty.

03 / Your finish line

A time in mind?

See where your goal sits among similar runners.

Goal time optional

04 / Keep exploring

At other distances

A little context

Behind the numbers

Useful estimates, with room for race day.

Where does the range come from?

The middle estimate is the median personal best among runners with a similar PB at your starting distance. The range covers the middle 80% of their target-distance best times. Each runner counts once.

We match PBs within ±2.5% of the nearest supported lookup time. The comparisons do not account for training, course, weather or when the races were run. Goal percentages describe this group, rather than a calibrated personal probability.

What do the other methods mean?

Original research uses the archived median and smoothed (LOWESS) relationships from the Poultry project. Riegel uses the standard distance-scaling exponent of 1.06.

World Athletics finds equal-point performances in the original men’s and women’s scoring tables. The tables’ edition is unverified and is not the current 2025 edition. Scoring equivalence is not an individual forecast.

A selected formula changes the headline estimate. The observed range and goal share always remain based on similar runners.

Why are some estimates unavailable?

Some distances and finish times have too few comparable runners, or fall outside the original tables. We leave those results blank rather than extrapolate. A formula may still provide a single estimate when there is no supported range.

Data and source notes